DotTalk++ / x64base Developer Manual

Assembled build — candidate artifact

2026-07-20

DotTalk++ / x64base Developer Manual

Assembled artifact — produced by the manifest-driven assembler from the declared bill of materials. This is a candidate build; acceptance is gated.

Manual id developer_manual_publication_v1
Public source snapshot 8ee746dee
Accepted reader baseline EA2E12A9D3E1 (accepted 2026-07-18T05:06:03Z)
Command-reference pages 183
Build date (UTC) 2026-07-20T19:17:26Z
Machine (maintainer-attested) Alienware m16 R2 / Intel Core Ultra 9 185H

Provenance & attestation

This manual is assembled from source, HELP/metadata, SelfDoc reports, and reviewed manualgen sections. Proof labels travel with each part.

Preface — how to read this manual

To be authored.

Table of Contents

Command Reference

183 command pages bound from the reference set.

ALLTRIM

ALLTRIM

Summary

Remove leading and trailing spaces from <cExpression>.

Status
Syntax
Provenance

APPEND

APPEND

Summary

Append a new record using current table defaults and active buffering rules.

Status
Syntax
Usage
Note
Provenance

AREA

AREA

Summary

Report the current DotTalk++ work-area state.

Status
Syntax
Usage
Example
Note
Provenance

AREA51

AREA51

Summary

Developer sandbox / experimental command.

Status
Syntax
Provenance

ASC

ASC

Summary

Return the ASCII code (0-255) of the first character of <cExpression>.

Status
Syntax
Provenance

ASCEND

ASCEND

Summary

Set ascending sort direction for the active order/tag.

Status
Syntax
Usage
Note
Provenance

ASCII

ASCII

Summary

Display ASCII/character-code reference information for teaching and diagnostics.

Status
Syntax
Provenance

AT

AT

Summary

Return the 1-based position of <cSearch> within <cExpression> (case-sensitive).

Status
Syntax
Provenance

ATC

ATC

Summary

Return the 1-based position of <cSearch> within <cExpression> (case-insensitive).

Status
Syntax
Provenance

AUTODBF

AUTODBF

Summary

Create or infer a DBF table from an input source using DotTalk++ automatic DBF-generation rules.

Status
Syntax
Usage
Argument
Note
Provenance

BIBLETALK

BIBLETALK

Summary

Use the BibleTalk educational seed database for quote, lookup, and teaching/demo workflows.

Status
Syntax
Provenance

BOTTOM

BOTTOM

Summary

Move to the last record in the current table/order.

Status
Syntax
Usage
Example
Note
Provenance

BROWSE

BROWSE

Summary

Open the classic browse surface for the current table/work-area context.

Status
Syntax
Usage
Note
Provenance

BROWSER

BROWSER

Summary

Developer browser command (experimental).

Status
Syntax
Usage
Note
Provenance

BUFFERING

BUFFERING

Summary

Buffering means changes are staged before permanent commit.

Commands
TABLE ON
REPLACE …
COMMIT
ROLLBACK (planned/deferred in some contexts)

Observed model
With TABLE ON:

Status
Syntax
Provenance

BUILDLMDB

BUILDLMDB

Summary

Build or rebuild the LMDB backing store for the current CDX container; may mutate LMDB/index files but not table records.

Status
Syntax
Usage
Argument
Note
Provenance

CANARY

CANARY

Status notice: PARTIAL; supported=F. Treat this page as limited or review-required evidence.

Summary

CANARY is a registered DotTalk++ command; curated DOTREF support status and help summary are pending.

Status
Provenance

CDX

CDX

Summary

Inspect or manage CDX container metadata and tag directories.

Status
Syntax
Usage
Note
Provenance

CHR

CHR

Summary

Convert an integer ASCII code (0-255) to a 1-character string.

Status
Syntax
Provenance

CHRISTMAS

CHRISTMAS

Summary

Display the DotTalk++ Christmas/holiday console splash screen.

Status
Syntax
Provenance

CLOSE

CLOSE

Summary

Close the current table, a selected area, or all open work areas.

Status
Syntax
Usage
Note
Provenance

CMDREL

CMDREL

Status notice: PENDING; supported=F. Treat this page as limited or review-required evidence.

Summary

Print the recipe for relating HELP COMMANDS to CMD_ARGS.

Usage
Note
Provenance

CNX

CNX

Summary

Index container command (CNX multi-tag support).

Status
Syntax
Usage
Argument
Note
Provenance

COBOL

COBOL

Summary

Display COBOL-oriented educational/demo material for historical data-processing context.

Status
Syntax
Provenance

CODASYL

CODASYL

Summary

Display CODASYL/network-database educational/demo material for historical database context.

Status
Syntax
Usage
Note
Provenance

COMMANDSHELP

COMMANDSHELP

Summary

Command help (alias of CMDHELP).

Status
Syntax
Provenance

COMMIT

COMMIT

Summary

Commit buffered TABLE updates to disk.

Notes (current shakedown observations):
- Clears TABLE stale state on success.
- May currently trigger full INX rebuild work as

Status
Syntax
Usage
Example
Note
Warning
Provenance

CONCAT

CONCAT

Summary

Concatenate string arguments. (Available as CALC function; also usable as a command where wired.)

Status
Syntax
Usage
Example
Note
Alias
Provenance

CONTINUE

CONTINUE

Summary

Continue a running loop/scan block (scripting control flow).

Status
Syntax
Usage
Note
Provenance

COPY

COPY

Summary

Copy table or data content into another target using DotTalk++ copy semantics.

Status
Syntax
Usage
Example
Note
Provenance

CREATE

CREATE

Summary

Create a new table, structure, or local project artifact through the implemented DotTalk++ creation surface.

Status
Syntax
Usage
Example
Note
Provenance

CTOD

CTOD

Summary

Convert a character date to a Date value using current SET DATE/SET CENTURY.

Status
Syntax
Provenance

DATE

DATE

Summary

Return the current system date.

Status
Syntax
Provenance

DECISION

DECISION

Summary

Programming Construct 2: Decision Flow

Definition
Decision flow chooses between alternatives based on a condition.

Current DotTalk++ form
IF <expr>

ELSE

ENDIF

Status
Syntax
Provenance

DELETE

DELETE

Summary

Mark the current record deleted using current table semantics.

Status
Syntax
Usage
Note
Provenance

DESCEND

DESCEND

Summary

Set descending sort direction for the active order/tag.

Status
Syntax
Usage
Note
Provenance

DIR

DIR

Summary

List directory or file entries through the DotTalk++ shell surface.

Status
Syntax
Usage
Note
Provenance

DO

DO

Status notice: PARTIAL; supported=F. Treat this page as limited or review-required evidence.

Summary

Execute a program. (Use DOTSCRIPT instead.)

Status
Syntax
Provenance

DOTHELP

DOTHELP

Summary

Show project-native DotTalk++ reference entries from the DOTREF catalog.

Status
Syntax
Usage
Note
Provenance

DOTSCRIPT

DOTSCRIPT

Summary

Run a DotTalk script file (test harness / automation).

Example:
DOTSCRIPT run_all_sql_tests.dts

Status
Syntax
Usage
Note
Provenance

DRAWIO

DRAWIO

Summary

Generate or manage Draw.io-oriented diagram artifacts for documentation and teaching workflows.

Status
Syntax
Usage
Note
Provenance

DTOC

DTOC

Summary

Convert a Date value to a character string using current SET DATE/SET CENTURY.

Status
Syntax
Provenance

DUMP

DUMP

Summary

Debug: print a structured dump of an internal value/state.

Status
Syntax
Usage
Note
Provenance

EDIT

EDIT

Summary

Edit a field/value interactively (where supported).

Status
Syntax
Provenance

EDUCATIONAL_USE

EDUCATIONAL_USE

Summary

A practical study order

1. Open a table
USE STUDENTS
STRUCT
FIELDS
LIST 5

2. Learn navigation
TOP
SKIP 1
BOTTOM
RECNO

3. Learn ordering
SET INDEX TO students.cdx

Status
Syntax
Provenance

ELSE

ELSE

Summary

Begin the alternate branch of an IF block when the IF condition is false.

Status
Syntax
Usage
Note
Provenance

ENDIF

ENDIF

Summary

End an IF/ELSE conditional block.

Status
Syntax
Usage
Note
Provenance

ENDLOOP

ENDLOOP

Summary

End a LOOP block (scripting).

Status
Syntax
Usage
Note
Provenance

ENDSCAN

ENDSCAN

Summary

End a SCAN block.

Status
Syntax
Usage
Note
Provenance

ENDUNTIL

ENDUNTIL

Summary

End an UNTIL block (scripting).

Status
Syntax
Usage
Note
Provenance

ENDWHILE

ENDWHILE

Summary

End a WHILE block (scripting control flow).

Status
Syntax
Usage
Note
Provenance

ERASE

ERASE

Summary

Erase a file or supported target through the DotTalk++ shell surface.

Status
Syntax
Usage
Example
Note
Provenance

ERP

ERP

Summary

Display ERP-oriented educational/demo material connecting database concepts to business systems.

Status
Syntax
Usage
Note
Provenance

ERSATZ

ERSATZ

Summary

Launch or drive the Ersatz relational browser/demo surface for workspace and relation-graph inspection.

Status
Syntax
Usage
Note
Provenance

EVAL

EVAL

Summary

Evaluate an expression and print its value.

Status
Syntax
Provenance

EVALUATE

EVALUATE

Summary

Evaluate an expression through the DotTalk++ expression layer and display the result.

Status
Syntax
Provenance

EXAMPLE

EXAMPLE

Summary

Show or run small DotTalk++ example/demo material for teaching and verification workflows.

Status
Syntax
Usage
Note
Provenance

EXPFUNCS

EXPFUNCS

Summary

Export the expression/function catalog to Markdown documentation.

Status
Syntax
Provenance

EXPORT

EXPORT

Summary

Export to CSV.

Status
Syntax
Usage
Note
Provenance

EXPORTSQL

EXPORTSQL

Summary

Export DotTalk++ table or workspace data through the SQL export helper surface.

Status
Syntax
Usage
Example
Note
Provenance

EXPRESSION

EXPRESSION

Summary

An expression is something the engine evaluates to produce a value.

Kinds of expressions
numeric
character
date
logical

Examples
1 + 2
UPPER(“abc”)
GPA + 0.01
DATE()

Status
Syntax
Provenance

FIND

FIND

Summary

Find text or values using the active order when possible, with scan fallback when needed.

Status
Syntax
Usage
Example
Note
Warning
Provenance

FOXHELP

FOXHELP

Summary

List or search command help topics.

Examples:
FOXHELP
FOXHELP WORKSPACE
FOXHELP REL

Notes:
Uses the help catalogs and any HELP DBFs.

Status
Syntax
Usage
Note
Provenance

FOXPRO

FOXPRO

Summary

DotTalk++ UI / browser command.

Status
Syntax
Provenance

FOXSTANDARD

FOXSTANDARD

Summary

Display FoxPro/xBase standard-reference or compatibility teaching material.

Status
Syntax
Usage
Note
Provenance

FOXTALK

FOXTALK

Summary

Legacy alias for the ArcticTalk Turbo Vision TUI shell.

Example:
FOXTALK

Notes:
Intended for keyboard-driven browsing and diagnostics.
Exits back

Status
Syntax
Provenance

GLOSSARY

GLOSSARY

Summary

Quick glossary

Alias
Name used to refer to a table/work area.

Buffer
Temporary staged change area before persistence.

Child
Related table reached from a parent.

Current area
The ac

Status
Syntax
Provenance

GO

GO

Summary

Execute a DotTalk++ command line (dispatch to command handlers).

Status
Syntax
Usage
Example
Note
Warning
Provenance

GOTO

GOTO

Summary

Move the current work area to a specific record or boundary position.

Status
Syntax
Usage
Example
Note
Warning
Provenance

GPS

GPS

Summary

Report current session/navigation position and work-area orientation diagnostics.

Status
Syntax
Usage
Note
Provenance

IDX

IDX

Summary

Memory-only educational index lab for teaching sort and index concepts without writing persistent index files.

Status
Syntax
Usage
Example
Note
Provenance

IF

IF

Summary

Begin a conditional DotScript block; execute the following block when the logical expression is true.

Status
Syntax
Usage
Example
Note
Provenance

IMAGE

IMAGE

Summary

Inspect image file metadata or open a supported image file in the operating-system viewer.

Status
Syntax
Usage
Note
Provenance

IMPORT

IMPORT

Summary

Import from CSV.

Status
Syntax
Usage
Note
Provenance

IMPORTSQL

IMPORTSQL

Summary

Import or bridge SQL data into DotTalk++ through the SQL import helper surface.

Status
Syntax
Usage
Example
Note
Provenance

INDEX

INDEX

Summary

General index-management command surface for the current table.

Status
Syntax
Usage
Example
Note
Provenance

INIT

INIT

Summary

Initialize default paths, perform best-effort stale-lock cleanup, and process startup ini scripts.

Status
Syntax
Usage
Note
Provenance

INTRO

INTRO

Summary

DotTalk++ Educational Reference

Purpose
EDREF is the instructional help catalog for learning DotTalk++ as a system.

What belongs here
- programming concepts shown by DotScript and command fl

Status
Syntax
Provenance

LEFT

LEFT

Summary

Return the left-most <nChars> characters of <cExpression>.

Status
Syntax
Provenance

LEN

LEN

Summary

Return the length (character count) of <cExpression>.

Status
Syntax
Provenance

LIST

LIST

Summary

List records from the current table using DotTalk++ command semantics.

Status
Syntax
Usage
Example
Note
Provenance

LMDB

LMDB

Summary

Inspect or manage per-area LMDB-backed index/storage wiring where supported.

Status
Syntax
Usage
Note
Provenance

LMDB_UTIL

LMDB_UTIL

Summary

Developer/diagnostic LMDB utility surface for low-level backend inspection and maintenance workflows.

Status
Syntax
Usage
Note
Provenance

LMDBDUMP

LMDBDUMP

Summary

Developer/diagnostic surface for dumping or inspecting LMDB-backed index/storage state.

Status
Syntax
Usage
Example
Note
Provenance

LOCATE

LOCATE

Summary

Position on the first record matching a predicate or simple field comparison.

Status
Syntax
Usage
Example
Note
Provenance

LOCK

LOCK

Summary

Lock the current record (or table) for editing.

Status
Syntax
Usage
Note
Provenance

LOOP

LOOP

Summary

Begin a LOOP block (scripting).

Status
Syntax
Usage
Note
Provenance

LOOP_BUFFER

LOOP_BUFFER

Summary

Developer/diagnostic helper surface for inspecting or testing buffered LOOP control-flow behavior.

Status
Syntax
Provenance

LOOPS

LOOPS

Summary

Programming Construct 3: Iteration / Looping

Definition
A loop repeats work.

DotTalk++ loop families
LOOP … ENDLOOP
WHILE … ENDWHILE
UNTIL … ENDUNTIL
SCAN … ENDSCAN

1. L

Status
Syntax
Provenance

LOWER

LOWER

Summary

Convert <cExpression> to lower-case.

Status
Syntax
Provenance

LTRIM

LTRIM

Summary

Remove leading spaces from <cExpression>.

Status
Syntax
Provenance

MCC

MCC

Summary

Load the MCC v32 demo workspace by running DotScript x32 and WORKSPACE LOAD mcc.dtschemas.

Status
Syntax
Usage
Note
Provenance

MODEL

MODEL

Summary

How to think about DotTalk++

DotTalk++ is best understood as four layers working together:

1. Command Layer
The CLI accepts commands such as:
USE
SELECT
LIST
SET

Status
Syntax
Provenance
Summary

Navigation means moving the current record pointer.

Common navigation commands
TOP
BOTTOM
SKIP
GOTO
SEEK
FIND
LOCATE

Important idea
Many commands are pointer-sensitiv

Status
Syntax
Provenance

NORMALIZE

NORMALIZE

Summary

Normalize/clean an input expression or text (developer utility).

Status
Syntax
Provenance

ORDER

ORDER

Summary

Order means the currently active navigation sequence.

Examples
physical order
LNAME order
FNAME order

Commands
SET ORDER TO TAG LNAME
SET ORDER TO 0
ASCEND
DESCEND
TO

Status
Syntax
Provenance

PACK

PACK

Summary

Permanently remove deleted records from the current table.

Status
Syntax
Usage
Example
Note
Provenance

PADC

PADC

Summary

Center-pad to length <nLen>.

Status
Syntax
Provenance

PADL

PADL

Summary

Left-pad to length <nLen>.

Status
Syntax
Provenance

PADR

PADR

Summary

Right-pad to length <nLen>.

Status
Syntax
Provenance

PREDHELP

PREDHELP

Summary

Help for predicates/expressions and filtering.

Status
Syntax
Usage
Note
Alias
Provenance

PREDICATE

PREDICATE

Summary

A predicate is a true/false expression.

Examples
GPA > 3.5
LNAME = “TAYLOR”
GPA >= 2 AND GPA <= 4

Used by
IF
LOCATE
COUNT
WHERE
SMARTLIST
SET FILTER

Boolean logi

Status
Syntax
Provenance

PREDICATES

PREDICATES

Summary

List supported predicates/operators for filtering.

Status
Syntax
Provenance

PROJECTION

PROJECTION

Summary

Projection means choosing which columns/fields to show.

Examples
TUPLE LNAME,FNAME
LIST FIELDS LNAME,FNAME
REL ENUM … TUPLE STUDENTS.SID,TEACHERS.LNAME

Educational point
Projection

Status
Syntax
Provenance

PROJECTS

PROJECTS

Summary

Inspect or manage DotTalk++ project/workflow entries and project-oriented local state.

Status
Syntax
Usage
Example
Note
Provenance

PROPER

PROPER

Summary

Convert <cExpression> to Proper Case (title case).

Status
Syntax
Provenance

PSHELL

PSHELL

Summary

Invoke or document the PowerShell/platform-shell helper surface where enabled by the runtime policy.

Status
Syntax
Usage
Example
Note
Provenance

REBUILD

REBUILD

Summary

Rebuild table/index/help-related derived state where supported by the current target context.

Status
Syntax
Usage
Note
Provenance

RECALL

RECALL

Summary

Unmark the current deleted record when supported by the current table state.

Status
Syntax
Usage
Example
Note
Alias
Provenance

REINDEX

REINDEX

Summary

Rebuild index files for the current table (or all open tables).

Status
Syntax
Usage
Note
Provenance

REL

REL

Summary

Relations engine commands (native DotTalk++ relation graph).

Subcommands:
REL LIST
REL LIST ALL
REL REFRESH
REL ENUM [LIMIT <n>] <child1> [<child2> …] TUPLE <tuple-expr>
REL SAV

Status
Syntax
Usage
Example
Note
Provenance

REL_ENUM

REL ENUM

Summary

Traverse a relation chain and emit tuple rows.

Syntax:
REL ENUM [LIMIT <n>] <child1> [<child2> …] TUPLE <tuple-expr>

Example:
REL ENUM LIMIT 5 ENROLL CLASSES TASSIGN TEACHERS
TUPLE

Status
Syntax
Provenance

REL_REFRESH

REL_REFRESH

Summary

Refresh relation state for the current workspace through the native REL backend.

Status
Syntax
Usage
Note
Provenance

RELATION

RELATION

Summary

A relation connects a parent table to a child table.

Example
STUDENTS -> ENROLL ON SID
ENROLL -> CLASSES ON CLS_ID

Meaning
If the current STUDENTS row has SID = 50000020,
the child

Status
Syntax
Example
Note
Provenance

RELATIONS

RELATIONS

Summary

Compatibility-facing relation listing surface backed by the native REL engine.

Examples:
RELATIONS
RELATIONS ALL

Notes:
RELATIONS and REL LIST point at the same relation-state reporting

Status
Syntax
Usage
Example
Note
Alias
Provenance

REPLACE

REPLACE

Summary

Replace one field in the current record using table-buffer and memo-aware semantics.

Status
Syntax
Usage
Example
Note
Provenance

REPLICATE

REPLICATE

Summary

Repeat <cExpression> <nTimes> times.

Status
Syntax
Provenance

RETRO

RETRO

Summary

Display ASCII-safe retro computer and system splash screens.

Status
Syntax
Usage
Example
Note
Provenance

RIGHT

Summary

Return the right-most <nChars> characters of <cExpression>.

Status
Syntax
Provenance

ROLLBACK

ROLLBACK

Summary

Discard staged TABLE/buffered changes without committing them; operational behavior depends on the current buffering state.

Status
Syntax
Usage
Example
Note
Provenance

RTRIM

RTRIM

Summary

Remove trailing spaces from <cExpression>.

Status
Syntax
Provenance

RULE

RULE

Summary

Inspect or manage validation-rule metadata and rule-oriented helper workflows.

Status
Syntax
Usage
Example
Note
Provenance

RUN

RUN

Status notice: PARTIAL; supported=F. Treat this page as limited or review-required evidence.

Summary

Run an OS command (FoxPro).

Status
Syntax
Provenance

SCAN

SCAN

Summary

Iterate records using a SCAN block (record loop).

Status
Syntax
Usage
Note
Provenance

SCAN_BUFFER

SCAN_BUFFER

Summary

Developer/diagnostic helper surface for inspecting or testing buffered SCAN/table-iteration behavior.

Status
Syntax
Provenance

SCHEMAS

SCHEMAS

Summary

Deprecated compatibility shim for WORKSPACE.

Current mapping:
SCHEMAS -> WORKSPACE
SCHEMAS OPEN <arg> -> WORKSPACE OPEN <arg>
SCHEMAS CLOSE

Status
Syntax
Usage
Example
Note
Provenance

SCRIPT

SCRIPT

Summary

DotScript is the scripting surface of DotTalk++.

Purpose
- automate repetitive commands
- build regression tests
- teach execution flow
- preserve reproducible scenarios

Why it matte

Status
Syntax
Provenance

SCX

SCX

Summary

Student/local SCX index-file lab command for creating, tagging, building, listing, and inspecting SCX index files.

Status
Syntax
Usage
Note
Provenance

SECHO

SECHO

Summary

Scripted/student echo helper for emitting text from teaching, demo, or test scripts.

Status
Syntax
Provenance

SECURITY

SECURITY

Summary

Inspect DotTalk++ security policy/runtime rules and manage the current shell-session role identity and assignment view.

Status
Syntax
Usage
Note
Provenance

SEEK

SEEK

Summary

Seek through the active order or by scanning a requested field.

Status
Syntax
Usage
Example
Note
Warning
Provenance

SELECT

SELECT

Summary

Select the active DotTalk++ work area by number or logical name.

Status
Syntax
Usage
Note
Provenance

SFTP

SFTP

Summary

File-transfer helper surface for SFTP-oriented workflows where enabled by local policy.

Status
Syntax
Usage
Note
Provenance

SHELLO

SHELLO

Summary

Small shell/student hello demonstration command for teaching command wiring and output.

Status
Syntax
Provenance

SHOWINI

SHOWINI

Summary

Display DotTalk++ initialization/configuration files and resolved startup settings.

Status
Syntax
Usage
Example
Note
Provenance

SHUTDOWN

SHUTDOWN

Summary

Run DotTalk++ shutdown processing, including shutdown.ini when present.

Status
Syntax
Usage
Note
Provenance

SIX

SIX

Summary

Experimental or compatibility index-related surface for SIX-style indexing concepts and diagnostics.

Status
Syntax
Provenance

SKIP

SKIP

Summary

Move relative to the current record in the active work area.

Status
Syntax
Usage
Example
Note
Provenance

SMARTLIST

SMARTLIST

Summary

LIST-style output with SQL-grade filtering (order-aware).

Examples:
SMARTLIST 20 FOR gpa >= 3.5
SMARTLIST ALL FOR lname LIKE “SMI%”

Notes:
Evaluat

Status
Syntax
Usage
Example
Note
Provenance

SPACE

SPACE

Summary

Return a character string containing <nSpaces> spaces.

Status
Syntax
Provenance

SQL

SQL

Summary

Execute an SQL statement using the configured SQL engine.

Status
Syntax
Usage
Example
Note
Provenance

SQLERASE

SQLERASE

Summary

SQL helper surface for erase/drop/delete-style SQL maintenance workflows where enabled.

Status
Syntax
Usage
Example
Note
Provenance

SQLHELP

SQLHELP

Summary

Show SQL-oriented help and guidance for DotTalk++ SQL bridge workflows.

Status
Syntax
Usage
Example
Note
Provenance

SQLITE

SQLITE

Summary

SQLite integration.

Examples:
SQLITE DB data_regression.sqlite
SQLITE EXEC CREATE TABLE t(x INT)
SQLITE EXEC INSERT INTO t(x) VALUES (1)
SQ

Status
Syntax
Usage
Note
Provenance

SQLSEL

SQLSEL

Summary

Execute an SQL SELECT and display results.

Status
Syntax
Usage
Example
Note
Provenance

SQLVER

SQLVER

Summary

Report SQLite availability and version.

Status
Syntax
Usage
Note
Provenance

STATE

STATE

Summary

State is one of the most important DotTalk++ concepts.

Definition
State means the current internal condition of the session.

Examples of session state
- current work area
- current table

Status
Syntax
Provenance

STR

STR

Summary

Convert numeric <nExpr> to a character string (optionally controlling width/decimals).

Status
Syntax
Provenance

STRUCT

STRUCT

Summary

Display table structure.

Status
Syntax
Usage
Example
Note
Provenance

STU_REPEAT

STU_REPEAT

Summary

Student/demo repeat command used to teach argument handling, loops, and command output.

Status
Syntax
Provenance

STU_UPPER

STU_UPPER

Summary

Student/demo uppercase command used to teach string handling and command output.

Status
Syntax
Provenance

STUDENTECHO

STUDENTECHO

Summary

Student/demo echo command used to teach command registration, argument handling, and output.

Status
Syntax
Provenance

STUDENTHELLO

STUDENTHELLO

Summary

Student/demo hello command used to teach the simplest command registration and output path.

Status
Syntax
Provenance

STUFF

STUFF

Summary

Replace <nLen> characters of <cExpr> starting at <nStart> with <cRepl>.

Status
Syntax
Provenance

SUBSTR

SUBSTR

Summary

Return substring of <cExpr> starting at <nStart> for <nLen> characters (or to end).

Status
Syntax
Provenance

TABLE_BUFFER

TABLE_BUFFER

Summary

Developer/diagnostic helper surface for inspecting or testing TABLE buffering and stale-field behavior.

Status
Syntax
Usage
Note
Provenance

TEXT

TEXT

Summary

Text-oriented helper command surface for demonstration, teaching, or local text workflows.

Status
Syntax
Provenance

TIME

TIME

Summary

Return the current system time as a character string (HH:MM:SS).

Status
Syntax
Provenance

TOP

TOP

Summary

Move to the first record in the current table/order.

Status
Syntax
Usage
Example
Note
Provenance

TRIM

TRIM

Summary

Remove trailing spaces from <cExpression> (alias/compat; see RTRIM).

Status
Syntax
Provenance

TUPEXPORT

TUPEXPORT

Summary

Export tuple/projection output through the tuple export helper surface.

Status
Syntax
Usage
Example
Note
Provenance

TUPLE

TUPLE

Summary

Build one tuple row from fields across work areas.

Examples:
TUPLE * (all fields from current area)
TUPLE LNAME,FNAME (selected fields from

Status
Syntax
Usage
Example
Note
Provenance

TUPLEDELTA

TUPLEDELTA

Summary

Tuple diagnostic helper surface for comparing projected tuple output or tuple-state deltas.

Status
Syntax
Usage
Note
Provenance

TUPTALK

TUPTALK

Summary

DotTalk++ tuple/logical-row command.

Status
Syntax
Usage
Example
Note
Provenance

TUPVALIDATE

TUPVALIDATE

Summary

Tuple validation helper surface for checking tuple projection, relation-walk, or logical-row consistency.

Status
Syntax
Usage
Example
Note
Provenance

TURBOPACK

TURBOPACK

Summary

Turbo Vision / pack-related utility.

Status
Syntax
Usage
Example
Note
Provenance

TVISION

TVISION

Summary

Turbo Vision diagnostics / demos.

Status
Syntax
Usage
Example
Note
Provenance

UNDELETE

UNDELETE

Summary

Compatibility alias for RECALL to unmark the current deleted record.

Status
Syntax
Provenance

UNLOCK

UNLOCK

Summary

Release a previously acquired lock (optionally all locks).

Status
Syntax
Usage
Example
Note
Provenance

UNTIL

UNTIL

Summary

Begin an UNTIL block (scripting).

Status
Syntax
Usage
Example
Note
Provenance

UNTIL_BUFFER

UNTIL_BUFFER

Summary

Developer/diagnostic helper surface for inspecting or testing buffered UNTIL control-flow behavior.

Status
Syntax
Provenance

UPDATE

UPDATE

Summary

Update data rows (scripting/SQL helper; see command usage).

Status
Syntax
Usage
Example
Note
Provenance

UPPER

UPPER

Summary

Convert <cExpression> to upper-case.

Status
Syntax
Provenance

USE

USE

Summary

Open a DBF table in the active DotTalk++ work area.

Status
Syntax
Usage
Note
Provenance

VAL

VAL

Summary

Convert a numeric-looking character expression to a number.

Status
Syntax
Provenance

VALIDATE

VALIDATE

Summary

Schema/sidecar validation command.

Status
Syntax
Usage
Example
Note
Provenance

WEB

WEB

Summary

Web-oriented helper command surface for local documentation, preview, or integration workflows where enabled.

Status
Syntax
Usage
Note
Provenance

WHILE

WHILE

Summary

Begin a WHILE block (scripting).

Status
Syntax
Usage
Example
Note
Provenance

WHILE_BUFFER

WHILE_BUFFER

Summary

Developer/diagnostic helper surface for inspecting or testing buffered WHILE control-flow behavior.

Status
Syntax
Provenance

WORKSPACE

WORKSPACE

Summary

Canonical live work-area/session command.

Usage:
WORKSPACE
List current open work areas.

WORKSPACE CLOSE
Close all work areas and clea

Status
Syntax
Usage
Example
Note
Provenance

WSREPORT

WSREPORT

Summary

Workspace report / diagnostics.

Status
Syntax
Usage
Note
Provenance

ZAP

ZAP

Summary

Delete all records from the current table.

Status
Syntax
Usage
Example
Note
Provenance

ZIP

ZIP

Summary

Archive helper command surface for ZIP-oriented local packaging or inspection workflows.

Status
Syntax
Usage
Example
Note
Provenance

Function Reference

63 core functions harvested from function_catalog.cpp (FunctionDoc table). Student/extension functions self-register from src/ext/fn and are not enumerated here.

Function Category Args Summary
ABS Numeric 1..1 Return the absolute value of a numeric expression.
ACOS Numeric 1..1 Return the arc-cosine of a numeric value.
ALLTRIM String 1..1 Trim leading and trailing spaces from a string value.
ASC String 1..1 Return the character code of the first character in a string.
ASIN Numeric 1..1 Return the arc-sine of a numeric value.
AT Search 2..3 Return the 1-based position of a substring occurrence.
ATAN Numeric 1..1 Return the arc-tangent of a numeric value.
ATC Search 2..3 Return the 1-based position of a substring occurrence, case-insensitive.
BETWEEN Numeric 3..3 Return whether a numeric value lies within an inclusive lower and upper bound.
CDOW Date 1..1 Return the day-of-week name for a date.
CEILING Numeric 1..1 Return the smallest integer greater than or equal to the numeric value.
CHR String 1..1 Return the single-character string for a numeric character code.
CHRTRAN String 3..3 Replace characters in a string by position-mapped translation.
CMONTH Date 1..1 Return the month name for a date.
CONCAT Construction 1..32 Concatenate all arguments into a single string.
COS Numeric 1..1 Return the cosine of an angle expressed in radians.
CTOD Date 1..1 Convert a character string to a date value.
DATE Date 0..0 Return the current system date.
DATEADD Date 2..2 Return a new date shifted by a specified interval.
DATEDIFF Date 2..2 Return the difference between two date values.
DATETIME Date 0..0 Return the current date/time value.
DAY Date 1..1 Return the day-of-month component of a date.
DOW Date 1..1 Return the day-of-week number for a date.
DTOC Date 1..2 Convert a date value to character form.
DTOS Date 1..1 Convert a date value to sortable character form.
EMPTY Logical 1..1 Return whether a value is empty according to expression-engine rules.
EXP Numeric 1..1 Return e raised to the power of the supplied numeric value.
FLOOR Numeric 1..1 Return the largest integer less than or equal to the numeric value.
GOMONTH Date 2..2 Return a date shifted by a specified number of months.
INT Numeric 1..1 Return the integer portion of a numeric value.
LEFT String 2..2 Return the leftmost characters of a string.
LEN Numeric 1..1 Return the length of a string value.
LIKE Logical 2..2 Match a value against a wildcard pattern.
LOG Numeric 1..1 Return the natural logarithm of a numeric value.
LOG10 Numeric 1..1 Return the base-10 logarithm of a numeric value.
LOWER String 1..1 Convert text to lowercase.
LTRIM String 1..1 Trim leading spaces from a string value.
MAX Numeric 2..32 Return the largest value from two or more numeric expressions.
MIN Numeric 2..32 Return the smallest value from two or more numeric expressions.
MOD Numeric 2..2 Return the remainder from dividing one numeric value by another.
MONTH Date 1..1 Return the month component of a date.
NOW Date 0..0 Return the current date/time value.
RAND Numeric 0..0 Return a pseudo-random numeric value.
RAT Search 2..2 Return the 1-based position of the last occurrence of a substring.
REPLICATE Construction 2..2 Repeat a string value a specified number of times.
RIGHT String 2..2 Return the rightmost characters of a string.
ROUND Numeric 1..2 Round a numeric value to a specified number of decimal places.
RTRIM String 1..1 Trim trailing spaces from a string value.
SECONDS Date 0..0 Return the current time expressed as seconds-of-day or equivalent engine-specific seconds value.
SIN Numeric 1..1 Return the sine of an angle expressed in radians.
SOUNDEX Search 1..1 Return a four-character phonetic Soundex code for a character expression.
SPACE Construction 1..1 Return a string containing the requested number of spaces.
SQRT Numeric 1..1 Return the square root of a numeric value.
STR Conversion 1..3 Convert a numeric value to a formatted string.
STRTRAN String 3..3 Replace occurrences of a substring within a string.
SUBSTR String 2..3 Return a substring starting at a 1-based position.
TAN Numeric 1..1 Return the tangent of an angle expressed in radians.
TIME Date 0..0 Return the current system time.
TODAY Date 0..0 Return the current system date.
TRANSFORM Conversion 1..2 Convert a value to display-oriented character form.
UPPER String 1..1 Convert text to uppercase.
VAL Conversion 1..1 Extract the leading numeric portion of a string.
YEAR Date 1..1 Return the year component of a date.

Appendix: Review and Deferred: SET-family

Status: SECTION_SKELETON_DRAFT_REPAIRED / REVIEW_REQUIRED

Purpose: Skeleton section generated from the revised manual TOC draft.

Promotion boundary: - This section is a structural draft. - Linked command pages are evidence-backed drafts, not final prose. - Review before promotion into the finished Developer Manual.

Commands in this section

Notes for future prose pass

Error / Message Catalog

Canonical error codes use an HRESULT-style 32-bit packing (severity · facility · code) defined in src/cli/xbase_error_codes.cpp.

Severities: success=0, warning=1, error=2.

Facilities (subsystems):

Facility Value
general 0x0001
dbf64 0x0002
fpt64 0x0003
security 0x0004
cli 0x0005
io 0x0006
runtime 0x0007

Review candidate. The full code→message enumeration is sourced from src/cli/message_catalog.cpp, src/cli/help_errors.cpp, and src/help/helpdata_messages.cpp; harvesting the complete message table is a later refinement of assembler:message-catalog. The severity/facility taxonomy above is source-derived and final.

Tables, Records, and Data Model

Pippets used: - PIP-004 Reviewed Candidate Pippet - PIP-003 Evidence Review Pippet - PIP-011 Source Reference Inventory Pippet

Evidence lineage: - MDO-120 original draft - MDO-128 source-lane repaired draft - MDO-133 field-note drift repaired v4 draft - MDO-134 evidence review gate

Repair reason: - MDO-121 detected evidence/prose drift in the first Tables, Records, and Data Model draft. - MDO-128 recovers or rebuilds the repair plan, then generates the source-lane repaired draft.

Evidence class: - Repaired draft assembled from MDO-120 draft intent, MDO-121 evidence gate where available, source-lane crosswalk or recovered repair plan, and registry v4. - Runtime behavior remains the source of truth. - This draft does not mutate HELP, META, CMDHELPCHK, catalogs, source files, or production SelfDoc metadata.

Purpose of this section

This section explains the table, record, field, schema, memo, and record-view vocabulary that later manual sections depend on. It follows the Getting Started and Workspaces sections: the reader should already know how the system starts, how areas frame context, and why the manual carries evidence boundaries.

The repaired draft is deliberately conservative. It does not assume every linked evidence page is the same kind of source. Some pages support current DotTalk command prose, some are educational concepts, some are compatibility references, and some need additional review before being described as runtime behavior.

Source-lane rule for this section

A DOTREF/current evidence item may support command-facing prose after evidence review. An EDREF concept item explains vocabulary or teaching structure. A FOXREF compatibility item must be gated before being described as current runtime behavior. SQL or shell reference lanes should be routed to bridge or appendix material.

This rule prevents concept pages such as TABLE_RECORD_FIELD from being documented as if they were ordinary runtime commands, and prevents compatibility-only material from being promoted as current DotTalk behavior without proof.

Tables, records, and fields

The core data model starts with tables, records, and fields. TABLE supports table-level vocabulary. RECORD and RECNO support record identity and record-position vocabulary. FIELDS and FIELDMGR support field-level vocabulary and field-management evidence.

TABLE_RECORD_FIELD is concept evidence for the relationship among tables, records, and fields. It should explain the model, not be presented as a user command unless a separate runtime command page proves that behavior.

Schema and DDL

SCHEMA and DDL belong together in this section because they describe structure. SCHEMA supports table-structure vocabulary. DDL supports definition-level or schema-language vocabulary. This draft keeps DDL wording structural and avoids claiming specific DDL behavior until command evidence and runtime examples are reviewed.

Record views and memo fields

RECORDVIEW supports record-oriented presentation or view vocabulary. MEMO supports memo-field discussion. Memo behavior is a sensitive storage and persistence topic, so final prose should distinguish visible memo fields from backend memo storage and should not overclaim persistence details without runtime evidence.

Work areas as context, not the main topic

WORKAREA may appear in the evidence set, but the main explanation of work areas belongs in the already promoted Workspaces section. Here, WORKAREA should be used only to remind the reader that table and record commands operate in a current area/session context.

Compatibility and bridge material

Compatibility evidence appears in this section and must be handled cautiously:

Compatibility evidence can support historical or compatibility notes, but it should not be promoted as current DotTalk behavior without runtime proof.

Command and concept map

What this section should not do yet

Review notes before candidate generation

Boundary

Command Surface, Dispatch, and Entry Variants

Pippets used: - PIP-001 Target Selection - MDO-166 Target Selection - MDO-167 Draft Fill

Evidence boundary: - Runtime proves behavior. - Source defines implementation and subsystem ownership. - HELP explains. - Metadata organizes. - CMDHELPCHK validates. - SelfDoc preserves provenance. - Manualgen assembles.

Slow-lane warning: - This section touches parser dispatch, public command surface, internal command-family ownership, aliases, variants, subcommands, function bridge behavior, and generated command-reference canaries. - Do not send this directly to generic PIP-003. - Run a slow-lane command-surface review first.

Evidence tokens under review: - COMMANDS - SYSCMD - SYSSUBCMD - SYSENTVAR - CMDKEY - CAN_NAME - QUAL_NAME - TOKEN - HANDLER - VIS - PUB_SURF - DISP_REACH - HELP - CMDHELPCHK - parser - dispatch - handler - command surface - entry variant - alias - subcommand - canonical command - AGGS - SET family - FUNCTION bridge

Draft notes: - This is conservative manual prose for slow-lane evidence review. - Generated command draft pages remain draft evidence, not final command reference prose. - Public command surface must be separated from internal owner/family scaffolding. - AGGS is treated as internal/family-owner evidence unless explicitly accepted as a public command surface. - SET-family canonicalization remains deferred unless separately repaired or accepted. - Function bridge behavior must preserve scalar/function entry while respecting command ownership. - SYSCMD, SYSSUBCMD, SYSENTVAR, SYSARGS, SYSHELP, SYSMSG, and SYSFUNC remain future/current feeders even when sparse.

Purpose of this section

This section explains how the Developer Manual should talk about command surface, parser dispatch, command handlers, canonical command identity, aliases, entry variants, subcommands, and function-bridge entry.

It follows Navigation, Indexing, Expressions, and Messages because those sections repeatedly depended on command/function routing and ownership boundaries. The manual now needs a section that explains how commands should be described without collapsing public command surface, internal command-family owners, generated command evidence, parser dispatch, and future metadata feeders into one authority.

The goal is not to publish the final command reference. The goal is to preserve a safe model for developer prose: command surface is what the user can intentionally invoke; dispatch is how input reaches implementation; metadata organizes identity and variants; HELP explains; CMDHELPCHK validates; runtime/source evidence still decides behavior and ownership.

Evidence lanes

This draft uses several evidence lanes.

Current command evidence lane: - COMMANDS - generated command pages - HELP command topics - CMDHELPCHK reports - HELP_COMMANDS exports where available

Current metadata evidence lane: - SYSCMD - SYSSUBCMD - SYSENTVAR - SYSARGS - SYSHELP - SYSMSG - SYSFUNC

Dispatch concept lane: - parser - dispatch - handler - command surface - public command - internal family owner - entry variant - alias - subcommand - canonical command - function bridge

Canary lane: - AGGS internal owner exposure; - SET family canonicalization; - command/function bridge behavior; - generated command duplicate and slug-collision evidence; - sparse metadata feeder coverage; - aliases and variants that should not be collapsed into canonical commands without review.

Authority boundaries

The same doctrine applies here:

The command-surface section must not let generated command pages or metadata rows become stronger proof than they are. A generated command page is draft evidence. A metadata row organizes identity. A HELP row explains. A handler in source defines implementation. A runtime test proves observed behavior.

Command surface

Command surface means the command vocabulary that a user can intentionally invoke.

Safe wording: - A public command surface should be documented as public only when HELP, metadata, source, and/or runtime evidence support that status. - A word that appears in generated command pages is not automatically a final public command. - A word that appears as an internal owner or command family should not be promoted as user-facing without explicit acceptance. - Visibility and public-surface flags should eventually be checked against metadata such as VIS and PUB_SURF.

This matters for command families, scaffolding, debug commands, and internal owner commands.

GENERIC remains a developer-utility canary

GENERIC is present in the supported DOTREF surface as a developer utility placeholder. Keep it visible for dispatch and metadata reconciliation, but do not present it as a normal user workflow until runtime behavior and intended audience are separately established. Its current value is as a command-surface canary: registration, HELP identity, handler ownership, and runtime availability can be compared without inventing stronger behavior prose.

Canonical commands

A canonical command is the main command identity used for documentation, metadata, and command-reference organization.

Safe wording: - A canonical command may have aliases or entry variants. - A canonical command may own subcommands. - A generated page slug is not automatically the canonical identity. - A handler name is implementation evidence, not necessarily the public name. - CMDKEY, CAN_NAME, and QUAL_NAME should eventually help separate command identity levels.

The manual should preserve a difference between: - token typed by a user; - canonical command identity; - qualified command/subcommand identity; - generated page path; - source handler; - metadata row; - HELP topic.

Subcommands and command families

Some command surfaces are naturally scoped through families or subcommands.

Examples needing careful treatment: - SET family - SET ORDER - SET INDEX - REL variants - WORKSPACE LOAD - CODASYL LOAD - aggregate-family ownership - AGGS as a possible internal owner

Safe wording: - A family or owner command may organize related verbs or subcommands. - A subcommand should be documented under its scoped owner when evidence supports that structure. - A family name should not be treated as public executable command unless accepted and evidenced. - SET-family canonicalization remains deferred until repaired or explicitly accepted.

Aliases and entry variants

Aliases and entry variants are not automatically separate canonical commands.

Safe wording: - An alias may route to a canonical command. - An entry variant may preserve compatibility vocabulary, shortcut spelling, or a user-facing convenience form. - A generated command page may represent an alias or variant rather than a canonical command. - SYSENTVAR should eventually organize aliases, variants, shortcut spellings, compatibility forms, and app-style entry points.

Manual prose should avoid treating every distinct token as a separate command until dedupe and variant review are complete.

Application-style UI entry points

ARCTICTALK and FOXPRO are application launch entries rather than ordinary data commands. When Turbo Vision support is present, ARCTICTALK opens the ArcticTalk shell and FOXPRO opens the FoxPro-style workbench. FOXTALK is the legacy alias for ARCTICTALK.

These names may be registered or documented even when a particular build does not contain the corresponding Turbo Vision surface. Describe availability from the selected build and runtime evidence; do not infer universal availability from registration alone. Neither entry currently exposes a separate usage branch, so the bare command form is the documented launch surface.

Parser dispatch and handlers

Parser dispatch is the route from user input to implementation. A handler is the implementation endpoint or command function that receives the routed input.

Safe wording: - Parser dispatch and handler ownership require source evidence. - Runtime evidence proves observed routing behavior. - HELP and metadata can explain or organize dispatch but do not prove runtime routing. - Handler names may be internal and should not be exposed as public command names unless explicitly intended.

This section should avoid claiming the exact parser algorithm unless source/runtime evidence is attached.

Function bridge behavior

DotTalk++ allows some function/app forms to be used directly from the command line. Prior runtime notes include examples such as UPPER and LEFT.

This is a command-surface canary because function names can look like command verbs.

Safe wording: - FUNCTION bridge behavior should preserve scalar/function entry where supported. - A function-style command-line entry does not necessarily make the function a command. - A command verb may shadow a scalar function name. - MIN/MAX ambiguity remains a separate canary from the expression section. - SYSFUNC and SYSENTVAR should eventually help organize function-command bridge forms.

AGGS boundary

AGGS is a known command-surface canary.

Current doctrine: - AGGS is intended as an internal owner or aggregate-family grouping. - Direct aggregate verbs such as SUM, AVG, MIN, and MAX are the intended user-facing aggregate command surface where evidenced. - AGGS appearing as executable or printing usage may be scaffold/debug leakage unless explicitly accepted. - Generated command pages for AGGS are draft evidence, not final public-surface proof.

This section should preserve AGGS internal family owner exposure as a visible canary.

SET family boundary

SET-family canonicalization remains deferred.

Safe wording: - SET ORDER and SET INDEX are scoped command surfaces that need canonicalization review. - SET-family pages may exist as generated draft evidence. - Final command reference organization should not be settled in this section without dedicated SET-family review. - Indexing owns order/tag semantics; this section owns command-surface identity and dispatch cautions.

Generated command pages

Generated command pages are useful evidence, but they are not final command reference.

Known generated-page issues: - duplicates; - aliases; - variants; - slug collisions; - SET-family canonicalization; - internal owner exposure; - command/function ambiguity.

Safe wording: - Generated pages identify draft evidence. - They should be deduped and reviewed before publication. - Command reference generation must not delete or rewrite generated command pages during manual draft assembly. - Generated pages should feed review, not replace review.

HELP and CMDHELPCHK

HELP explains command usage, concepts, and warnings. CMDHELPCHK validates HELP/catalog consistency.

Safe wording: - HELP can explain command surface intent. - CMDHELPCHK can detect gaps and drift. - Neither HELP nor CMDHELPCHK replaces runtime proof. - Neither HELP nor CMDHELPCHK replaces source ownership.

Manual assembly may use HELP/META/CMDHELPCHK-first workflow, but truth authority remains role-separated.

Future META alignment

This section should eventually align with the metadata system.

Expected future feeders: - SYSCMD for canonical command identity, command key, canonical name, qualified name, token, handler, visibility, public surface, display reach, owner, source authority, source file, help topic, active flag, and notes. - SYSSUBCMD for command-family subcommands such as SET ORDER, SET INDEX, REL variants, WORKSPACE variants, and other scoped command surfaces. - SYSENTVAR for aliases, entry variants, shortcut spellings, compatibility forms, app-style function entries, and reviewed variants. - SYSARGS for command argument shapes, expression arguments, predicates, tag names, scopes, deleted filters, required arguments, and repeatable arguments. - SYSHELP for help text connected to command owners, canonical commands, subcommands, variants, warnings, examples, and reference material. - SYSMSG for parser diagnostics, unknown command messages, syntax errors, invalid arguments, and ambiguity warnings. - SYSFUNC for function-command bridge cases where scalar function entry overlaps command-like syntax. - HELP_COMMANDS and generated command pages as draft evidence lanes that require dedupe, alias, and slug-collision review.

Temporary evidence is acceptable only when marked as temporary and crosswalked to future META feeders.

Slow-lane canary tracking names

The slow-lane review tracks these canaries by exact name. These names are review anchors, not final user-facing prose.

These anchors preserve the canaries that the prose discusses in ordinary language. They should remain until the section is promoted through evidence review.

Review notes before PIP-003

This is a slow-lane section. Before generic PIP-003 is allowed to create a reviewed-candidate path, an MDO slow-lane evidence review should check:

Recommended required tokens for later PIP-003: - COMMANDS - SYSCMD - SYSSUBCMD - SYSENTVAR - CMDKEY - CAN_NAME - QUAL_NAME - TOKEN - HANDLER - VIS - PUB_SURF - DISP_REACH - HELP - CMDHELPCHK - parser - dispatch - handler - command surface - entry variant - alias - subcommand - canonical command - AGGS - SET family - FUNCTION bridge

Boundary

Expressions, Querying, and Aggregates

Pippets used: - PIP-001 Target Selection - MDO-151 Target Selection - MDO-152 Draft Fill

Evidence boundary: - Runtime proves behavior. - Source defines implementation and subsystem ownership. - HELP explains. - Metadata organizes. - CMDHELPCHK validates. - SelfDoc preserves provenance. - Manualgen assembles.

Slow-lane warning: - This section touches known command/function and parser canaries. - Do not send this directly to generic PIP-003. - Run a slow-lane expression/function/aggregate review first.

Evidence tokens under review: - CALC - CALCWRITE - WHERE - FOR - LOCATE - CONTINUE - SCAN - COUNT - SUM - AVG - MIN - MAX - AGGS - FUNCTION - FUNCTIONS - PREDICATES - MIN() - MAX() - xexpr - DELETED - NOT DELETED - !DELETED

Draft notes: - This is conservative manual prose for slow-lane evidence review. - Generated command draft pages remain draft evidence, not final command prose. - xexpr owns expression evaluation surfaces unless current source proves a narrower owner for a specific path. - SYSFUNC is a future semantic feeder even if current seeding is sparse. - Scalar function forms and aggregate command forms must remain separate until runtime/source evidence closes the ambiguity. - AGGS is treated as internal/family-owner evidence unless explicitly accepted as a user-facing command surface.

Purpose of this section

This section explains expression evaluation, predicates, query filters, aggregate commands, and function surfaces in DotTalk++.

It follows Navigation and Indexing because those sections establish record movement, search context, active order, tags, projection, and relation boundaries. Expressions sit beneath many of those surfaces. Predicates decide which records are considered. Aggregate commands summarize values across records. Function calls support scalar calculation, command-line function applications, and calculated command behavior.

The goal is not to publish a final function reference. The goal is to establish safe developer-manual prose that preserves the boundaries between expression evaluation, command parsing, predicate filtering, aggregate traversal, scalar functions, and future metadata feeders.

Evidence lanes

This draft uses several evidence lanes.

Current DotTalk evidence lane: - CALC - CALCWRITE - WHERE - FOR - LOCATE - CONTINUE - SCAN - COUNT - SUM - AVG - MIN - MAX - AGGS - FUNCTION - FUNCTIONS - PREDICATES - MIN() - MAX() - xexpr - DELETED - NOT DELETED - !DELETED

Generated command-reference lane: - Generated command pages may identify draft command evidence. - They are not final prose and should not be quoted as final command authority. - AGGS and MIN/MAX require extra review because generated pages can expose implementation or family-owner concepts that may not be intended as public command surface.

Runtime evidence lane: - Runtime examples have shown direct aggregate verbs SUM, AVG, MIN, and MAX working against an open table. - Runtime examples have shown WHERE and FOR producing matching aggregate results in tested cases. - Runtime examples have shown deleted-record filters affecting aggregate outputs in tested cases. - Runtime examples have shown command-line function application behavior for functions such as UPPER and LEFT. - Runtime examples have shown MIN/MAX parser ambiguity that must remain canary-sensitive.

Concept lane: - An expression computes a value. - A predicate is an expression used as a true/false condition. - A query filter selects records. - An aggregate command computes a result across records. - A scalar function computes a result from arguments. - A command parser may route command-like input differently from expression-function input.

Compatibility lane: - xBase/FoxPro lineage can explain vocabulary, but compatibility evidence must not be promoted as current DotTalk behavior without runtime proof. - Function names that overlap with command verbs are especially compatibility-sensitive.

Future META feeder lane: - SYSFUNC should eventually carry canonical function identity, display name, category, argument range, handler, CALC_CALL, PUB_SURF, SELF_REG, MSG_CAT, and active status. - SYSARGS should eventually carry function and command argument shapes, required/repeatable flags, predicate shapes, deleted filters, and expression values. - SYSCMD should eventually carry command identity for COUNT, SUM, AVG, MIN, MAX, CALC, CALCWRITE, LOCATE, CONTINUE, SCAN, and related command surfaces. - SYSSUBCMD should eventually carry aggregate-family or predicate-related subcommands if the command model keeps subcommand ownership. - SYSENTVAR should eventually carry command/function variants, aliases, and app-style function entry points after seed hygiene review. - SYSHELP should eventually carry curated/generated help text for functions, predicates, aggregate commands, and expression concepts. - SYSMSG should eventually carry expression errors, nonnumeric aggregate values, no-active-table messages, not-found messages, deleted-filter outcomes, and parser ambiguity warnings.

Expression evaluation surfaces

Expression evaluation is shared infrastructure. The manual should treat xexpr as the expression engine unless current source evidence proves otherwise for a specific path.

Expression surfaces include: - CALC - CALCWRITE - command arguments that accept value expressions - predicates in WHERE or FOR style clauses - function calls - calculated values used by aggregate commands

The developer manual should make a distinction between expression syntax and command syntax. A command may accept an expression, but that does not make the command itself the expression engine.

CALC and CALCWRITE

CALC and CALCWRITE are expression-oriented command surfaces. They are useful places to explain evaluated expressions without requiring table traversal.

Safe wording: - CALC evaluates an expression and displays or returns the result according to its command contract. - CALCWRITE evaluates and writes or displays according to its command contract. - Exact output behavior should be verified with HELP and runtime evidence before final wording.

These commands are natural cross-references for function help and the xexpr engine.

Predicates, WHERE, and FOR

A predicate is an expression used as a condition. WHERE and FOR are predicate-bearing clauses or surfaces.

Safe wording: - WHERE and FOR can restrict which records participate in a command where supported. - Runtime evidence has shown matching aggregate results for equivalent WHERE and FOR predicates in tested aggregate cases. - This equivalence should not be generalized across all commands without evidence.

The manual should not imply that WHERE and FOR are always identical. It should say that they can serve related predicate-filter roles and that each command surface must be checked.

LOCATE, CONTINUE, and SCAN

LOCATE, CONTINUE, and SCAN connect predicates with traversal.

Safe distinctions: - LOCATE searches for records matching a condition. - CONTINUE resumes a prior locate-style search where supported. - SCAN iterates over records and may use predicate or scope rules depending on command syntax.

These commands connect this section with Navigation and Indexing: - Navigation owns movement and current record context. - Indexing owns active order and tag-sensitive traversal context. - Expressions own predicate evaluation. - Commands own their own syntax and side effects.

COUNT and aggregate commands

COUNT, SUM, AVG, MIN, and MAX belong to aggregate command discussion when they operate across records.

Safe aggregate wording: - COUNT counts records or matched records according to command syntax and scope. - SUM computes a total for a numeric expression across records. - AVG computes an average for a numeric expression across records. - MIN and MAX compute minimum and maximum aggregate results where command syntax and runtime support them.

The final manual should attach runtime proof for each command surface before claiming exact syntax, deleted-record behavior, or null/empty-set behavior.

Direct aggregate verbs and AGGS

Direct aggregate verbs should be treated as the user-facing command surface where runtime and HELP evidence support them: - SUM - AVG - MIN - MAX

AGGS should be treated carefully. Current doctrine is: - AGGS is intended as an owner/internal grouping for aggregate verbs. - AGGS being executable or printing usage may be debug/scaffold leakage unless explicitly accepted. - Generated command pages for AGGS are draft evidence, not final public-surface proof.

The manual should explain aggregate family ownership without accidentally publishing an internal owner as a user command.

Scalar functions versus aggregate commands

MIN and MAX are the important canary.

There are two concepts: - scalar function form: MIN() or MAX() as a function over supplied arguments; - aggregate command form: MIN or MAX over records.

These must not be collapsed.

Known runtime canary: - command-style input has shown MIN/MAX aggregate behavior. - function-app and command-line function bridging exists for functions such as UPPER and LEFT. - MIN(2,1) versus MIN 2,1 parser behavior must remain canary-sensitive until current runtime/source evidence closes it.

The manual should preserve the ambiguity: - MIN/MAX aggregate commands are command surfaces. - MIN()/MAX() scalar functions are function surfaces if current function registry/runtime evidence confirms them. - Parser dispatch and command/function shadowing must be documented carefully.

Function command-line bridge

DotTalk++ allows some function/app forms to be used directly from the command line.

Examples in prior runtime notes include: - UPPER with a string-like value - LEFT with a value and length argument

This is important because function names may appear command-like when typed at the prompt. The manual should explain that some command-line input can bridge into scalar/function-app evaluation.

However, the bridge must not override command ownership: - A direct command verb may own a word such as MIN or MAX. - A scalar function may also exist with the same name. - The parser/dispatcher decides which surface receives the input.

Deleted-record filters

Aggregate commands may support deleted-record filters.

Evidence-sensitive filter vocabulary: - DELETED - NOT DELETED - !DELETED

Safe wording: - Deleted-record filters may affect which records participate in aggregate results where supported. - Runtime evidence has shown DELETED returning empty/null-like aggregate behavior in tested cases and NOT DELETED matching full nondeleted results in tested cases. - Exact behavior must be verified per command, expression type, and data state.

The manual should not generalize deleted-record behavior beyond tested aggregate surfaces without proof.

Proof boundary: - deleted-record aggregate filters are proof-aware. - DELETED, NOT DELETED, and !DELETED should not be generalized beyond tested aggregate surfaces. - Final wording must remain tied to command, expression type, and data-state evidence.

Error and null behavior

Aggregate and expression commands can produce errors or null-like results.

Examples needing proof-aware wording: - nonnumeric expression supplied to a numeric aggregate; - character expression supplied to AVG; - unknown field or expression; - empty or deleted-only record sets; - no active table; - invalid function argument count.

SYSMSG is the intended future feeder for error symbols, severity, short text, suggested actions, and implementation status.

HELP FUNCTIONS and FUNCTION help

HELP FUNCTIONS and HELP FUNCTION should be included as user-facing discovery surfaces for expression functions.

The manual should distinguish: - command help for command verbs; - function help for expression functions; - generated command draft pages; - future SYSFUNC metadata.

SYSFUNC is important even when sparsely seeded because it is the future semantic feeder for canonical function identity, argument ranges, handlers, public surface status, and message-catalog alignment.

Slow-lane canary tracking names

The slow-lane review tracks these canaries by exact name. These names are review anchors, not final user-facing prose.

These anchors preserve the canaries that the prose discusses in ordinary language. They should remain until the section is promoted through evidence review.

Future META alignment

This section should eventually align with the metadata system.

Expected future feeders: - SYSFUNC for canonical function identity, display name, category, argument range, implementation status, visibility tier, owner, source authority, source file, handler, CALC_CALL, PUB_SURF, SELF_REG, MSG_CAT, active status, and notes. - SYSARGS for function and command argument shapes, predicates, deleted filters, repeatable arguments, and required values. - SYSCMD for command identity and handler alignment for CALC, CALCWRITE, COUNT, SUM, AVG, MIN, MAX, LOCATE, CONTINUE, and SCAN. - SYSSUBCMD for aggregate-family or predicate-related subcommands if those are modeled as subcommands. - SYSENTVAR for variants, aliases, and command-line function-app entry points after seed hygiene review. - SYSHELP for generated and curated help text connected to command and function owners. - SYSMSG for expression, aggregate, predicate, and parser diagnostics.

Temporary evidence is acceptable only when marked as temporary and crosswalked to future META feeders.

Review notes before PIP-003

This is a slow-lane section. Before generic PIP-003 is allowed to create a reviewed-candidate path, an MDO slow-lane evidence review should check:

Recommended required tokens for later PIP-003: - CALC - CALCWRITE - WHERE - FOR - LOCATE - CONTINUE - SCAN - COUNT - SUM - AVG - MIN - MAX - AGGS - FUNCTION - FUNCTIONS - PREDICATES - MIN() - MAX() - xexpr - DELETED - NOT DELETED - !DELETED

Boundary

Indexing, Tags, Relations, and Views

Pippets used: - PIP-001 Target Selection - MDO-143 Target Selection - MDO-144 Draft Fill

Evidence boundary: - Runtime proves behavior. - Source defines implementation and subsystem ownership. - HELP explains. - Metadata organizes. - CMDHELPCHK validates. - SelfDoc preserves provenance. - Manualgen assembles.

Slow-lane warning: - This section touches known canaries. - Do not send this directly to promotion review. - Run a canary-aware evidence review before PIP-003 is allowed to produce a reviewed candidate.

Evidence tokens under review: - INDEX - REINDEX - REBUILD - SET ORDER - SET INDEX - ASCEND - DESCEND - SEEK - FIND - CNX - CDX - LMDB - BUILDLMDB - REL - RELATIONS - ERSATZ - VIEW

Draft notes: - This is conservative manual prose for slow-lane evidence review. - Generated command draft pages remain draft evidence, not final command reference prose. - The section preserves the distinction between logical/user-facing abstractions and physical backend details. - CDX/CNX language should be reviewed against current implementation and HELP/META evidence before final wording. - SET-family canonicalization remains deferred unless separately repaired or accepted. - Relations, tuple traversal, browser rendering, and views must not be collapsed into one ownership model.

Purpose of this section

This section explains the concepts that control ordered traversal, indexed lookup, relation-aware traversal, and view/projection terminology in DotTalk++.

It follows Navigation, Browsing, and Search because that section deliberately deferred index-specific and relation-specific behavior. Navigation can say that traversal and search may depend on context. This section explains the important context: active order, tags, logical index surfaces, physical backend boundaries, relations, and views.

The goal is not to publish a final index command reference. The goal is to establish safe developer-manual prose that keeps ownership boundaries visible and prevents logical abstractions from being confused with physical backends.

Evidence lanes

This draft uses several evidence lanes.

Current DotTalk evidence lane: - INDEX - REINDEX - REBUILD - SET ORDER - SET INDEX - ASCEND - DESCEND - SEEK - FIND - CNX - CDX - LMDB - BUILDLMDB - REL - RELATIONS - ERSATZ - VIEW

Generated command-reference lane: - Generated command pages may identify available draft command evidence. - They are not final prose and should not be quoted as final manual authority. - Duplicate commands, aliases, slug collisions, and SET-family canonicalization still require command-reference review.

Concept lane: - Indexing provides ordered or keyed access paths. - Tags name or select logical orders. - Active order affects traversal and key-style search where supported. - Relations connect areas or tables through traversal rules. - Views and browser output are projection surfaces unless proven otherwise.

Compatibility lane: - xBase/FoxPro lineage can explain vocabulary, but compatibility material must not be promoted as current DotTalk behavior without runtime proof. - SET-family commands are especially compatibility-sensitive.

Future META feeder lane: - SYSCMD should eventually carry command identity and handler alignment. - SYSSUBCMD should eventually carry SET ORDER, SET INDEX, REL, and related subcommand identity. - SYSENTVAR should eventually carry aliases, variants, and shortcut spellings after seed hygiene review. - SYSARGS should eventually carry tag names, key expressions, relation/view arguments, and rebuild options. - SYSMSG should eventually carry missing tag, order not found, backend build, relation warning, and rebuild diagnostics. - SYSHELP should eventually carry curated/generated concept help for indexes, tags, relations, views, and traversal.

Indexing vocabulary

Indexing vocabulary needs careful separation.

A user may think about indexes as a way to find records faster or view records in a useful order. A developer manual should be more precise: an index or order is a traversal/access structure that can influence how commands move through or locate records.

Important terms: - index: a structure or command surface associated with ordered/keyed access. - order: the currently selected traversal order where supported. - tag: a named logical order within a multi-tag abstraction. - rebuild/reindex: operations that refresh or rebuild index structures. - key expression: the value or expression used to derive ordered/keyed lookup behavior.

The section should avoid claiming that every navigation or search command always uses an index. That behavior must be proven per command and context.

Open index architecture note

The indexing lane is intentionally open enough to teach and inspect.

That means the manual should preserve the distinction between:

The important manual rule is not to flatten those families into one generic word. They overlap conceptually, but they do not all serve the same role.

Logical order and active order

The active order is a current traversal context. It may affect display, movement, and search behavior depending on the command path.

A safe explanation is: - table state has a current record position; - ordered traversal may change the sequence in which records are visited; - commands such as LIST can expose the actual traversal order; - key-style commands such as SEEK or FIND may depend on active order or key context where supported.

Known canary: - Reported active order must agree with actual traversal order before an order path is marked proven. - A command saying an order is active is not enough. LIST, SEEK, rebuild, and runtime smoke evidence must agree.

Ownership reminder

Index systems own:

Projection layers such as LIST, BROWSE, ERSATZ, GUI, and TUI may reveal ordered traversal, but they do not own index truth. They are evidence surfaces, not index authorities.

Tags and tag availability

A tag names a logical order. In a multi-tag model, selecting a tag should select a user-facing logical order.

This section should preserve this rule: - tag availability and reported active tag must be verified against actual traversal behavior.

Known proof-sensitive case: - A tag may be reported as selected before it is actually available or rebuilt. - After rebuild, traversal may become consistent. - This should remain a canary until runtime proof closes it.

Do not hide this from developer documentation. The manual may keep user-facing prose simple, but the Developer Manual must preserve the proof boundary.

CDX, CNX, and LMDB boundary

The preferred doctrine is:

This does not mean LMDB is unimportant. It means ordinary command documentation should avoid making users think they are operating directly on the physical backend when they are really selecting orders, tags, or logical index structures.

Developer-facing prose may explain: - the logical abstraction seen by commands; - the physical backend used internally; - where the backend boundary must not leak into public command vocabulary.

This boundary should be reviewed against current source and HELP/META evidence before final manual promotion.

SET-family boundary and canonicalization canary

SET-family commands are known to require careful canonicalization.

The section may mention: - SET ORDER - SET INDEX - SET-family command surfaces - the need to distinguish aliases, variants, and canonical commands

But it should not resolve canonicalization casually.

Known canary: - SET-family canonicalization remains deferred. - Generated command-reference pages for SET-family items should remain draft evidence until canonicalization is repaired or explicitly accepted.

Manual rule: - Keep SET-family wording conservative. - Do not treat duplicate or variant generated command pages as final canonical command identity.

SEEK and FIND active-order boundary

Navigation, Browsing, and Search introduced SEEK and FIND as search vocabulary but deferred index-specific behavior. This section owns the active-order boundary for those commands.

A safe statement is: - SEEK and FIND are key-style search commands whose exact behavior may depend on active order, tag, or index context. - The final manual should attach runtime proof before claiming exact behavior. - If a command falls back to physical order, reports active order incorrectly, or requires rebuild first, that must remain visible as a canary.

This section should not collapse SEEK/FIND with LOCATE. LOCATE is more naturally predicate-oriented and belongs with expression/predicate search behavior, even if it appears in navigation prose.

Reindexing and rebuild behavior

REINDEX, REBUILD, and BUILDLMDB touch refresh/build behavior.

A conservative explanation is: - rebuild/reindex commands update index or backend structures; - their exact scope and backend effects must be verified by command evidence; - public-facing wording should focus on refreshing logical orders or tags; - backend-specific wording belongs in developer/backend notes.

Known risk: - BUILDLMDB and LMDB terminology can pull physical storage details into user-facing prose. - Keep public/user wording logical unless a developer/backend section explicitly opens the physical layer.

Relations and relation traversal

Relations connect work areas or tables so that a parent context can lead to related child records.

This section should explain relation traversal without making browser output the owner of relation semantics.

Ownership rule: - relation subsystem owns relation definitions and traversal intent; - tuple infrastructure owns relation-aware row projection; - browser and ERSATZ surfaces render or navigate projected relation state; - workspace/session systems own restored area and relation context.

This section can cross-reference the Workspaces and Tuple sections rather than fully restating them.

Known proof: - MCC/x32 relation paths and ERSATZ browser output have provided useful runtime evidence. - x64 workspace/ERSATZ load reporting remains canary-sensitive.

Views and projection boundary

The word view is dangerous because it can mean a saved query, a projection, a browser surface, or a conceptual display. The manual should not assume one meaning until HELP/META/source evidence classifies the actual command surface.

Safe language: - a view/projection presents selected or arranged data; - projection is not storage ownership; - relation-aware projection is not the same as relation definition; - browser output is not the same as table or relation ownership.

This section should preserve that ambiguity until the command-reference and source evidence clarify VIEW and related surfaces.

ERSATZ and browser caution

ERSATZ and relation-aware browser output are valuable evidence surfaces, but they are path-specific.

Known cautions: - plain ERSATZ/no-arg paths and MCC/x32 relation browser paths have useful proof value; - ERSATZ GRID was previously deferred because its snapshot branch did not preserve the same complete BrowserSnapshot; - browser output may be usable even when workspace load reporting is noisy; - projection output should not be promoted to semantic ownership.

Use ERSATZ evidence to explain what can be seen. Do not use it alone to prove every underlying workspace/relation/load behavior.

Known canaries

This section must keep the following canaries visible:

Slow-lane canary tracking names

The slow-lane review tracks these canaries by exact name. These names are review anchors, not final user-facing prose.

These anchors preserve the canaries that the prose already discusses in ordinary language. They should remain until the section is promoted through evidence review. ### Future META alignment

This section should eventually align with the metadata system.

Expected future feeders: - SYSCMD for command identity and handler mapping. - SYSSUBCMD for SET ORDER, SET INDEX, REL, and other subcommand identity. - SYSENTVAR for aliases, variants, and shortcuts after seed hygiene review. - SYSARGS for tag names, key expressions, relation names, view arguments, rebuild flags, and backend options. - SYSMSG for diagnostics about missing tags, unavailable orders, not-found search results, relation warnings, backend build failures, and rebuild results. - SYSHELP for curated/generated concept help about indexing, tags, relations, views, and traversal.

Temporary evidence is acceptable only when marked as temporary and crosswalked to future META feeders.

Review notes before PIP-003

This is a slow-lane section. Before generic PIP-003 is allowed to create a reviewed-candidate path, an MDO slow-lane evidence review should check:

Recommended required tokens for later PIP-003: - INDEX - REINDEX - REBUILD - SET ORDER - SET INDEX - ASCEND - DESCEND - SEEK - FIND - CNX - CDX - LMDB - BUILDLMDB - REL - RELATIONS - ERSATZ - VIEW

Boundary

Messages, Errors, and Diagnostics

Pippets used: - PIP-001 Target Selection - MDO-159 Target Selection - MDO-160 Draft Fill

Evidence boundary: - Runtime proves behavior. - Source defines implementation and subsystem ownership. - HELP explains. - Metadata organizes. - CMDHELPCHK validates. - SelfDoc preserves provenance. - Manualgen assembles.

Slow-lane warning: - This section touches message-catalog, HELP, metadata, validation, and runtime-reporting boundaries. - Do not send this directly to generic PIP-003. - Run a slow-lane message/diagnostic review first.

Evidence tokens under review: - SYSMSG - SYSTEM_MESSAGES - MSG_ID - SYMBOL - ENUM_NAME - SEVERITY - FACILITY - SHORT_TXT - SUG_ACT - HELP - HELP GIANT - CMDHELPCHK - WARNING - ERROR - STATUS - SHARED_MSG - diagnostic - message catalog - typed message - parser warning - nonnumeric aggregate - no active table - not found

Draft notes: - This is conservative manual prose for slow-lane evidence review. - Generated command draft pages remain draft evidence, not final command prose. - SYSMSG is preserved as a future typed message feeder even if current metadata seeding is sparse. - SYSTEM_MESSAGES may represent an older or alternate long-form metadata schema and must be crosswalked carefully. - HELP explains and CMDHELPCHK validates; neither should be treated as runtime proof. - Message catalog doctrine must not mutate HELP, META, CMDHELPCHK, catalogs, source, or runtime data during manual draft assembly.

Purpose of this section

This section explains how DotTalk++ developer documentation should treat messages, errors, warnings, statuses, diagnostics, and future typed message catalog work.

It follows the expression and aggregate section because that section introduced several diagnostic-heavy cases: nonnumeric aggregate values, parser ambiguity, empty or deleted-only aggregate inputs, no-active-table conditions, not-found outcomes, and expression/function argument errors. Those cases need a shared diagnostic vocabulary before the manual expands deeper into command families and subsystem behavior.

The goal is not to claim that the full typed messaging system is already complete. The goal is to preserve the intended direction while keeping current proof boundaries visible.

Evidence lanes

This draft uses several evidence lanes.

Current HELP evidence lane: - HELP - HELP GIANT - HELP_LINE - HELP_ARTIFACTS - WARNING rows - ERROR rows - STATUS rows - SHARED_MSG rows

Current metadata evidence lane: - SYSMSG - SYSTEM_MESSAGES - SYSHELP - SYSCMD - SYSSUBCMD - SYSFUNC - SYSARGS - SYSENTVAR

Diagnostic concept lane: - message - diagnostic - warning - error - status - trace - log - test output - parser warning - message catalog - typed message

Runtime/source proof lane: - parser warnings require runtime/source evidence; - expression errors require runtime/source evidence; - aggregate errors require runtime/source evidence; - no-active-table messages require runtime/source evidence; - not-found messages require runtime/source evidence.

Future catalog lane: - the messaging layer should become the single typed, catalog-backed reporting path; - current coverage must be verified before final manual claims; - sparse metadata seeding is a future feeder, not a reason to ignore the lane.

Authority boundaries

The same doctrine applies here:

This section must be careful because HELP and diagnostics are tempting to treat as proof. HELP output can explain intended or cataloged behavior, but it does not by itself prove runtime execution. CMDHELPCHK can validate consistency, but it does not by itself prove runtime behavior either.

Message vocabulary

The manual should not collapse every output line into the same kind of message.

Useful distinctions: - A message is a reported unit of text or structured information. - A diagnostic explains a condition that may require interpretation or action. - A warning reports a risk or questionable condition. - An error reports a failed operation or invalid condition. - A status reports state or completion information. - A trace reports internal execution evidence. - A log records events for later inspection. - A test result reports expected versus observed behavior. - HELP text explains usage, concepts, warnings, examples, and reference material. - A catalog row organizes message identity and metadata.

These categories may overlap in implementation, but the manual should not collapse them until source and metadata evidence prove the ownership model.

Typed message catalog direction

The project direction is that the messaging layer should become the single typed, catalog-backed reporting path for: - commands; - errors; - help text; - syntax issues; - warnings; - traces; - UI/status messages; - logs; - tests; - HELP validation; - upper-layer metadata reporting and collection.

This is directionally important, but not automatically complete. The manual should say: - intended direction: typed, catalog-backed reporting; - current proof: must be verified by source, runtime, HELP, and metadata evidence; - draft boundary: no HELP/META/CMDHELPCHK/catalog/source/runtime mutation during manual assembly.

SYSMSG and SYSTEM_MESSAGES

SYSMSG is the compact/current metadata feeder identified for message catalog work. Its fields include message identity and message metadata such as: - MSG_ID - SYMBOL - ENUM_NAME - SEVERITY - FACILITY - SHORT_TXT - IMPL_STAT - VIS_TIER - OWNER - SRC_AUTH - SRC_FILE - PUB_SURF - USED_RUN - ACTIVE - VER_AT - SUG_ACT - NOTES

SYSTEM_MESSAGES appears in earlier or alternate metadata material as a long-form schema name. It should not be assumed identical to SYSMSG without a crosswalk.

Safe wording: - SYSMSG is the current compact future feeder for message metadata where seeded and verified. - SYSTEM_MESSAGES may be legacy or alternate long-form metadata evidence. - The manual should crosswalk them carefully before treating them as one authority.

HELP rows as diagnostic evidence

HELP GIANT and HELP tables provide useful evidence for diagnostic text and categories.

Evidence examples: - WARNING rows; - ERROR rows; - STATUS rows; - SHARED_MSG rows; - SOURCE_FACT rows; - HELP_LINE records; - HELP_ARTIFACTS records.

But HELP rows are explanation/catalog evidence, not runtime proof. A WARNING row in HELP means the help system has a warning artifact; it does not prove the runtime command currently emits that warning.

SHARED_MSG caution

SHARED_MSG rows are useful because they suggest text or messages shared across HELP or command surfaces.

Safe wording: - SHARED_MSG rows are evidence. - They may be sparse. - They should not be treated as complete message catalog coverage. - They should be crosswalked to SYSMSG/SYSTEM_MESSAGES and source/runtime evidence before final implementation claims.

CMDHELPCHK role

CMDHELPCHK validates HELP/catalog consistency. It is a validator and system contract checker, not just documentation.

Safe wording: - CMDHELPCHK can identify gaps, inconsistencies, or contract drift. - CMDHELPCHK supports manual assembly by checking HELP/catalog alignment. - CMDHELPCHK does not replace runtime proof. - CMDHELPCHK does not replace source ownership.

This matters because manual assembly can use HELP/META/CMDHELPCHK-first workflow, but truth authority remains runtime/source/HELP/metadata/CMDHELPCHK according to their roles.

Runtime diagnostic examples

Expressions and aggregates introduced diagnostic examples that should eventually map into message evidence: - nonnumeric aggregate expression; - character expression used with AVG; - empty or deleted-only aggregate input; - invalid field or expression; - parser ambiguity between command and function form; - no active table; - not found; - unsupported command syntax; - missing argument; - invalid argument count.

This draft does not claim all of those are cataloged today. It says they are natural future candidates for SYSMSG and source/runtime review.

Severity vocabulary

Severity vocabulary must not be invented.

Possible severity words include: - ERROR - WARNING - STATUS - INFO - TRACE - DEBUG

But final wording should only claim severity categories when evidence supports them through SYSMSG, HELP, source, or runtime behavior. The draft should avoid inventing a complete severity taxonomy.

Parser warnings and syntax diagnostics

Parser warnings and syntax diagnostics are important because command-line behavior can bridge commands and functions. Examples include: - command/function ambiguity; - unknown command; - missing argument; - unsupported syntax; - scalar function form versus aggregate command form.

The expressions section preserved MIN/MAX parser ambiguity. This diagnostics section should explain that such cases need typed messages or at least consistent diagnostic reporting, but it should not claim the final routing is complete without source/runtime proof.

No-active-table and not-found messages

No-active-table and not-found messages are common diagnostic candidates.

They should be handled conservatively: - no-active-table behavior depends on command context; - not-found behavior depends on search command, index/order context, and runtime state; - exact wording and severity need runtime/source evidence; - future SYSMSG rows should eventually organize the message identities.

Message catalog and HELP alignment

Message catalog work should eventually align: - SYSMSG message identity; - HELP text; - command/function ownership; - argument validation; - runtime emission; - CMDHELPCHK validation; - SelfDoc provenance.

This section should preserve that alignment goal but avoid claiming it is finished.

Slow-lane canary tracking names

The slow-lane review tracks these canaries by exact name. These names are review anchors, not final user-facing prose.

These anchors preserve the canaries that the prose discusses in ordinary language. They should remain until the section is promoted through evidence review.

Future META alignment

This section should eventually align with the metadata system.

Expected future feeders: - SYSMSG for message identity, symbol, enum name, severity, facility, short text, implementation status, visibility tier, owner, source authority, source file, public surface flag, used-at-runtime flag, suggested action, notes, and active status. - SYSTEM_MESSAGES for legacy or alternate long-form message schema evidence that must be crosswalked carefully. - SYSHELP for help text connected to message owners, diagnostic concepts, generated text, and curated text. - SYSCMD for command owners that emit diagnostics. - SYSSUBCMD for subcommand diagnostic surfaces when command families own messages. - SYSFUNC for function-related diagnostics such as argument count, nonnumeric value, parser ambiguity, and calculation errors. - SYSARGS for arguments involved in diagnostic validation and error reporting. - SYSENTVAR for aliases or variants that may affect diagnostic routing. - HELP_LINE and HELP_ARTIFACTS as current HELP evidence lanes for WARNING, ERROR, STATUS, SHARED_MSG, and related text.

Temporary evidence is acceptable only when marked as temporary and crosswalked to future META feeders.

Review notes before PIP-003

This is a slow-lane section. Before generic PIP-003 is allowed to create a reviewed-candidate path, an MDO slow-lane evidence review should check:

Recommended required tokens for later PIP-003: - SYSMSG - SYSTEM_MESSAGES - MSG_ID - SYMBOL - ENUM_NAME - SEVERITY - FACILITY - SHORT_TXT - SUG_ACT - HELP - HELP GIANT - CMDHELPCHK - WARNING - ERROR - STATUS - SHARED_MSG - diagnostic - message catalog - typed message - parser warning - nonnumeric aggregate - no active table - not found

Boundary

HELP, Metadata, CMDHELPCHK, and Manualgen Alignment

Pippets used: - PIP-001 Target Selection - MDO-173 Target Selection - MDO-174 Draft Fill

Evidence boundary: - Runtime proves behavior. - Source defines implementation and subsystem ownership. - HELP explains. - Metadata organizes. - CMDHELPCHK validates. - SelfDoc preserves provenance. - Manualgen assembles.

Slow-lane warning: - This section touches truth authority doctrine, manual assembly workflow, HELP evidence, metadata evidence, CMDHELPCHK validation, SelfDoc provenance, source authority, and future feeder alignment. - Do not send this directly to generic PIP-003. - Run a slow-lane authority/alignment review first.

Evidence tokens under review: - HELP - HELP GIANT - HELP_LINE - HELP_ARTIFACTS - CMDHELP - CMDHELPCHK - META - metadata - SYSCMD - SYSSUBCMD - SYSENTVAR - SYSARGS - SYSHELP - SYSMSG - SYSFUNC - SOURCE - SOURCE_FACT - SelfDoc - manualgen - PIP - validator - crosswalk - report-only - truth authority - assembly workflow

Draft notes: - This is conservative manual prose for slow-lane evidence review. - HELP/META/CMDHELPCHK-first is a manual assembly workflow, not a replacement for truth authority. - Source remains implementation authority even when manualgen starts from HELP, metadata, and CMDHELPCHK. - Sparse metadata feeders are future alignment lanes, not dead ends. - Temporary evidence lanes must be labeled and crosswalked to future META feeders. - Manualgen and SelfDoc work remain report-only unless explicitly authorized to mutate production artifacts.

Purpose of this section

This section explains how the Developer Manual should align HELP, metadata, CMDHELPCHK, SelfDoc, and manualgen without confusing their roles.

It follows the command-surface section because recent sections repeatedly relied on the same evidence pattern. Manualgen has been reading HELP broadly, reading metadata semantically, validating with CMDHELPCHK, verifying with source, proving with runtime, and assembling manual sections. That is a useful assembly workflow. It is not a replacement for truth authority.

The goal of this section is to prevent drift. Readers should understand which evidence lane explains, which organizes, which validates, which proves, which defines implementation, and which assembles manual output.

Core doctrine

The manual should use this doctrine consistently:

The doctrine matters because each lane is useful but limited. HELP can be excellent explanatory evidence without proving runtime behavior. Metadata can organize a future target even when currently sparse. CMDHELPCHK can catch alignment drift without replacing runtime tests or source review.

Assembly workflow versus truth authority

Manual assembly can use a HELP/META/CMDHELPCHK-first workflow:

That workflow is practical because HELP, metadata, and CMDHELPCHK give manualgen a wide map of the system. But it must not be described as the truth authority doctrine.

Truth authority remains role-separated:

Safe wording: - HELP/META/CMDHELPCHK-first is assembly order. - It is not authority order. - Source is not demoted to a sidecar outside implementation truth. - Runtime proof remains required for behavior claims.

HELP lane

HELP is one of the strongest manual assembly feeders.

Current HELP evidence may include: - HELP topics; - HELP GIANT output; - HELP_LINE rows; - HELP_ARTIFACTS rows; - CMDHELP artifacts; - generated HELP reports; - usage text; - warnings; - examples; - concept notes.

HELP explains the command surface and conceptual model. It can also expose vocabulary that manualgen should review.

Safe wording: - HELP explains. - HELP may reveal intended behavior or documented behavior. - HELP should not be treated as runtime proof. - HELP should be crosswalked to source, runtime, metadata, and CMDHELPCHK where claims matter.

Metadata lane

Metadata organizes the system.

Current and future metadata feeders include: - SYSCMD - SYSSUBCMD - SYSENTVAR - SYSARGS - SYSHELP - SYSMSG - SYSFUNC - SOURCE_FACT - HELP_LINE - HELP_ARTIFACTS

Metadata may be sparse, partially seeded, or split between older and newer schema forms. Sparse metadata is not a reason to ignore the lane. Sparse metadata should be labeled as a future alignment feeder until seeded and verified.

Safe wording: - Metadata organizes identity and relationships. - Metadata does not automatically prove runtime behavior. - Sparse metadata tables are future feeders, not dead ends. - Alternate metadata schemas should be crosswalked rather than collapsed casually.

CMDHELPCHK lane

CMDHELPCHK is a validator and system contract checker.

It can: - identify missing or inconsistent HELP rows; - compare command surfaces against HELP/catalog evidence; - report drift; - support manual assembly gates; - validate generated HELP and metadata alignment.

Safe wording: - CMDHELPCHK validates. - CMDHELPCHK can identify drift and gaps. - CMDHELPCHK does not prove runtime execution. - CMDHELPCHK does not replace source ownership.

SelfDoc lane

SelfDoc preserves provenance.

SelfDoc evidence includes: - source-comment contracts; - harvested source evidence; - report-only metadata staging; - provenance reports; - canary ledgers; - source/miner evidence; - manualgen run records; - savepoint journals.

SelfDoc should keep evidence visible and traceable. It should not silently mutate production HELP, metadata, CMDHELPCHK, catalogs, source, or runtime data during draft assembly.

Safe wording: - SelfDoc preserves provenance. - SelfDoc defaults to report-only. - SelfDoc can identify evidence and drift. - SelfDoc changes to production artifacts require explicit authorization.

Manualgen lane

Manualgen assembles.

Manualgen creates: - draft prose; - reviewed candidates; - promoted draft workspaces; - pippet run records; - summary reports; - gate reports; - savepoint records; - package bundles.

Manualgen does not publish final manuals by itself. Promoted draft workspaces are still draft workspaces unless final publication is explicitly authorized.

Safe wording: - Manualgen assembles. - Manualgen gates and records decisions. - Manualgen does not mutate HELP, META, CMDHELPCHK, catalogs, source, or runtime data during manual draft assembly. - Manualgen promoted drafts are not final publication.

Source lane

Source defines implementation and subsystem ownership.

Source evidence is required when the manual makes claims about: - implementation ownership; - parser routing; - command handlers; - backend ownership; - relation traversal semantics; - expression evaluation ownership; - memo payload lifecycle; - message emission; - runtime behavior that is not proven by a direct test.

Source may be consulted later in the workflow, but later does not mean weaker. In manual assembly order, source may be a verification lane. In truth authority, source remains implementation authority.

Safe wording: - Source defines implementation. - Source verifies ownership and routing. - Source is not merely a provenance sidecar. - Source claims should be tied to files, comments, contracts, or source-miner evidence.

Runtime lane

Runtime proves behavior.

Runtime evidence includes: - observed command runs; - smoke tests; - shakedown transcripts; - exact command output; - pass/fail test logs; - canary reproduction; - before/after behavior.

Runtime evidence is especially important for: - command execution; - parser ambiguity; - error output; - no-active-table behavior; - not-found behavior; - deleted-record behavior; - relation traversal; - memo backend attachment; - index/order behavior.

Safe wording: - Runtime proves observed behavior. - Runtime proof should include concrete commands and outputs where possible. - Runtime proof should be dated or tied to a build/session when possible. - Runtime proof does not by itself describe implementation ownership without source.

Temporary evidence lanes

Manualgen may need to use temporary evidence when future metadata feeders are sparse.

Temporary evidence examples: - generated command pages; - HELP GIANT exports; - current HELP rows; - manually curated canary notes; - runtime shakedown notes; - older metadata schema files; - seed scripts; - user/MDO handoff notes; - source-contract reports; - manualgen pippet reports.

Temporary evidence is allowed when labeled.

Safe wording: - This source is temporary evidence for the current manual pass. - Future feeder should be SYSFUNC, SYSMSG, SYSCMD, SYSSUBCMD, SYSENTVAR, SYSARGS, SYSHELP, or related metadata. - Temporary evidence should be crosswalked to future META feeders when the metadata system matures. - Temporary evidence must not be promoted as final authority without verification.

Future META alignment

This section should explicitly preserve future feeder alignment.

Expected future feeders: - SYSCMD for command identity, handler alignment, visibility, public surface, display reach, owner, source authority, source file, and help topic. - SYSSUBCMD for subcommand and command-family identity. - SYSENTVAR for aliases, variants, compatibility spellings, shortcut forms, and entry points. - SYSARGS for argument shapes, predicates, filters, scopes, validation surfaces, and repeatable flags. - SYSHELP for curated and generated help text. - SYSMSG for diagnostics, warnings, statuses, parser messages, and typed message catalog alignment. - SYSFUNC for function identity, categories, argument ranges, handler links, CALC/CALCWRITE reach, public surface, self-registration, and function-command bridge surfaces. - HELP_LINE and HELP_ARTIFACTS for current HELP evidence lanes. - SOURCE_FACT and source-contract evidence for source/comment provenance. - manualgen reports and PIP records for assembly provenance and gate evidence.

Sparse feeders should be kept visible. The manual should not ignore SYSFUNC, SYSMSG, SYSCMD, SYSSUBCMD, SYSENTVAR, SYSARGS, or SYSHELP because they are empty or only partially seeded today.

Crosswalk discipline

Crosswalks prevent drift.

Useful crosswalks include: - HELP topic to command identity; - command identity to SYSCMD; - subcommand to SYSSUBCMD; - alias or entry variant to SYSENTVAR; - argument shape to SYSARGS; - help text to SYSHELP; - diagnostic text to SYSMSG; - function reference to SYSFUNC; - source comment to SOURCE_FACT; - generated command page to canonical command and variant review; - manualgen section to evidence tokens and pippet reports.

Crosswalks should preserve uncertainty. A crosswalk can say “candidate match” or “future feeder” without claiming final authority.

Safety boundaries

This alignment section should restate the safety boundaries.

Default manualgen/SelfDoc boundary: - no generated command page deletion; - no HELP mutation; - no META mutation; - no CMDHELPCHK mutation; - no catalog apply; - no source edits; - no production SelfDoc metadata promotion; - no final publication without explicit authorization.

Report-only work is the default.

Slow-lane canary tracking names

The slow-lane review tracks these canaries by exact name. These names are review anchors, not final user-facing prose.

These anchors preserve the canaries that the prose discusses in ordinary language. They should remain until the section is promoted through evidence review.

Review notes before PIP-003

This is a slow-lane section. Before generic PIP-003 is allowed to create a reviewed-candidate path, an MDO slow-lane evidence review should check:

Recommended required tokens for later PIP-003: - HELP - HELP GIANT - HELP_LINE - HELP_ARTIFACTS - CMDHELP - CMDHELPCHK - META - metadata - SYSCMD - SYSSUBCMD - SYSENTVAR - SYSARGS - SYSHELP - SYSMSG - SYSFUNC - SOURCE - SOURCE_FACT - SelfDoc - manualgen - PIP - validator - crosswalk - report-only - truth authority - assembly workflow

Boundary

Command Reference Assembly, Aliases, and Generated Page Hygiene

Pippets used: - PIP-001 Target Selection - MDO-188 Target Selection - MDO-189 Draft Fill

Evidence boundary: - Runtime proves behavior. - Source defines implementation and subsystem ownership. - HELP explains. - Metadata organizes. - CMDHELPCHK validates. - SelfDoc preserves provenance. - Manualgen assembles.

Slow-lane warning: - This section touches generated command pages, duplicate command rows, aliases, entry variants, canonical command identity, slug collisions, LOAD guard behavior, SET-family canonicalization, AGGS exposure, and command-reference publication readiness. - Do not send this directly to generic PIP-003. - Run a slow-lane command-reference hygiene review first.

Evidence tokens under review: - command reference - generated command pages - HELP_COMMANDS - HELP - HELP GIANT - CMDHELP - CMDHELPCHK - SYSCMD - SYSSUBCMD - SYSENTVAR - SYSARGS - SYSHELP - alias - variant - canonical command - slug collision - duplicate command - APPEND BLANK - APPEND_BLANK - LOAD guard - SET-family - AGGS - internal owner - public surface - manualgen - PIP - crosswalk - report-only - no mutation

Draft notes: - This is conservative manual prose for slow-lane evidence review. - Generated command pages are draft evidence, not final command reference prose. - Duplicate command rows and slug collisions must be reviewed before publication. - Aliases and entry variants must not be treated as canonical commands without review. - APPEND BLANK and APPEND_BLANK may map to the same slug but may represent variant or canonicalization evidence. - LOAD guard must preserve no top-level LOAD page while preserving scoped CODASYL LOAD and WORKSPACE LOAD. - SET-family canonicalization remains deferred unless separately repaired or accepted. - AGGS remains internal or family-owner evidence unless explicitly accepted as public command surface. - HELP, CMDHELPCHK, and metadata can guide command reference review but cannot settle runtime or source canaries alone. - Command reference hygiene packages must not delete generated command pages during manual draft assembly.

Purpose of this section

This section explains how the Developer Manual should treat generated command pages, command-reference assembly, aliases, variants, canonical commands, and generated-page hygiene.

It follows the Command Surface, HELP/Metadata/CMDHELPCHK Alignment, and Runtime Evidence sections because those sections established the boundaries needed here. Command reference work needs all three: a model of command surface and entry variants, an alignment model for HELP and metadata, and an evidence model for runtime/source/canary closure.

The goal is not to publish the final command reference in this section. The goal is to define safe rules for assembling the future command reference from generated pages and other evidence without losing alias/variant information, publishing internal scaffolding, or deleting generated evidence prematurely.

Authority model for command reference assembly

The same doctrine applies:

Generated command pages are an evidence lane. They are not the authority by themselves.

Safe wording: - Generated pages feed review. - Generated pages do not replace review. - Generated pages may expose duplicates, aliases, scaffolding, or internal owners. - Final command-reference prose needs dedupe, alias/variant review, HELP/metadata alignment, and runtime/source checks where behavior or ownership is claimed.

Generated command pages

Generated command pages are useful because they preserve broad command-surface evidence.

They may show: - command names; - aliases; - variants; - generated slugs; - usage text; - HELP extraction; - metadata alignment candidates; - duplicate rows; - command-family scaffolding; - internal owner entries.

But generated pages can also expose draft artifacts.

Safe wording: - Generated command pages are draft evidence. - They are not final command reference prose. - They should be reviewed, deduped, and crosswalked before publication. - They should not be deleted or rewritten during ordinary manual draft assembly.

Duplicate command rows

Duplicate command rows can mean different things.

A duplicate may be: - a true duplicate that should be collapsed later; - an alias; - an entry variant; - a compatibility spelling; - a scoped subcommand; - a generated slug collision; - an internal owner or command-family scaffold; - a public command that shares a root token with another command.

Safe wording: - Duplicate rows are review input. - Duplicate rows should not be collapsed automatically. - A duplicate row is not proof that one page is wrong. - Dedupe should preserve evidence until canonical command, alias, variant, and scope are reviewed.

Aliases and variants

Aliases and variants are not automatically canonical commands.

An alias may be a shortcut spelling. A variant may be a compatibility form, app-style entry point, scoped command form, or user-facing convenience form. The command reference should preserve those distinctions until metadata and evidence review decide how they should be presented.

Safe wording: - Aliases and variants must not be treated as canonical commands without review. - SYSENTVAR should eventually organize aliases, variants, shortcut spellings, compatibility forms, generated page entries, and app-style entry points. - Generated command pages may represent aliases or variants rather than separate canonical commands. - Manual prose should avoid treating every distinct token as a separate command until dedupe and variant review are complete.

Canonical command identity

Canonical command identity is the reviewed identity used for final reference organization.

A canonical command should eventually align with: - SYSCMD command identity; - handler or source ownership where relevant; - public surface status; - help topic; - argument model; - aliases and variants; - generated page crosswalk; - runtime/source evidence where behavior or ownership is claimed.

Safe wording: - Canonical command identity should be reviewed. - Canonical command identity should not be inferred from slug text alone. - Canonical command identity should not erase useful alias or variant evidence. - Metadata can organize identity, but runtime/source evidence may still be required for behavior and ownership claims.

Slug collisions

A slug collision occurs when two different generated entries map to the same or confusing page slug.

Slug collisions may happen because: - spaces and underscores collapse; - punctuation is normalized; - aliases share words; - variants differ only by formatting; - scoped command names are flattened; - generator rules remove important distinctions.

Safe wording: - Slug collisions are review items. - Slug collisions should not be resolved by deleting generated pages during manual draft assembly. - Slug collision repair should preserve source evidence, generated evidence, and future metadata crosswalks. - A later command-reference hygiene pass may choose canonical slugs after review.

APPEND BLANK and APPEND_BLANK

APPEND BLANK and APPEND_BLANK are useful review examples.

They may map to the same slug or appear related in generated evidence, but the manual should not force a final answer here. They may represent: - a command plus argument form; - a compatibility form; - a generated normalization issue; - an alias or variant; - a canonicalization candidate.

Safe wording: - APPEND BLANK and APPEND_BLANK should remain variant/canonicalization review examples. - Do not treat one as wrong solely because of slug shape. - Do not publish a final command-reference rule until evidence review decides canonical command, variant, and argument treatment.

LOAD guard

LOAD is a known generated-page guard canary.

The guard should preserve: - no top-level LOAD page where that surface is intentionally suppressed; - scoped CODASYL LOAD where supported; - WORKSPACE LOAD where supported; - evidence rows that explain why top-level LOAD is guarded.

Safe wording: - LOAD guard must preserve no top-level LOAD page while preserving scoped CODASYL LOAD and WORKSPACE LOAD. - LOAD guard is not permission to delete all LOAD-related evidence. - Scoped LOAD forms should remain visible for review if they are evidenced. - A guard is a publication boundary, not an evidence deletion rule.

SET-family canonicalization

SET-family canonicalization remains deferred.

SET-family entries may include: - SET ORDER; - SET INDEX; - other SET-scoped command surfaces; - compatibility forms; - generated pages that flatten or split command/subcommand identities.

Safe wording: - SET-family canonicalization remains deferred unless separately repaired or accepted. - SET-family generated pages should be treated as draft evidence. - Indexing owns order/tag semantics; command-reference hygiene owns command identity and reference organization. - Final reference organization should not be settled by prose alone.

AGGS boundary

AGGS is a known command-reference canary.

Current doctrine: - AGGS is intended as an internal owner or aggregate-family grouping. - Direct aggregate verbs such as SUM, AVG, MIN, and MAX are the intended user-facing aggregate command surface where evidenced. - AGGS appearing in generated command pages may indicate internal owner exposure, scaffold leakage, or family grouping evidence. - AGGS should not be published as a public command surface unless explicitly accepted.

Safe wording: - AGGS remains internal or family-owner evidence unless explicitly accepted as public command surface. - Generated AGGS evidence should be retained for review. - Generated AGGS evidence should not become final public command-reference prose by default.

Internal owner and public surface

Command-reference assembly must separate internal owner evidence from public command surface.

An internal owner may be: - a command-family grouping; - handler scaffolding; - parser dispatch label; - generated metadata row; - development/debug surface; - owner used for help organization.

A public surface is what users are intended to invoke.

Safe wording: - Internal owner evidence is useful. - Internal owner evidence is not automatically public command surface. - Public surface requires intent and evidence. - PUB_SURF, VIS, DISP_REACH, HELP, runtime behavior, and source ownership may all participate in review.

HELP and CMDHELPCHK

HELP and CMDHELPCHK are important command-reference inputs.

HELP can explain: - usage; - concepts; - examples; - warnings; - aliases; - command families.

CMDHELPCHK can validate: - HELP/catalog consistency; - missing topics; - generated HELP drift; - command catalog alignment.

Safe wording: - HELP explains command reference intent. - CMDHELPCHK validates command-reference alignment. - Neither HELP nor CMDHELPCHK replaces runtime proof or source ownership. - HELP and CMDHELPCHK can guide command-reference hygiene.

Metadata feeders

Future and current metadata feeders should remain visible.

Expected feeders: - SYSCMD for canonical command identity, canonical name, handler, visibility, public surface, display reach, owner, source authority, and help topic. - SYSSUBCMD for scoped subcommand identity such as SET ORDER, SET INDEX, scoped LOAD forms, and command-family subcommands. - SYSENTVAR for aliases, variants, shortcut spellings, compatibility forms, generated page entries, and app-style entry points. - SYSARGS for command argument shapes, scopes, deleted filters, predicates, tag names, workspace files, and rebuild options. - SYSHELP for curated and generated command help text. - SYSMSG for parser, syntax, unknown command, invalid argument, and ambiguity diagnostics. - SYSFUNC for function-command bridge cases where scalar functions overlap command-like syntax.

Sparse feeders are still alignment lanes. They should not be ignored because they are incomplete.

Command-reference crosswalks

Crosswalks should connect generated pages to reviewed command identity without collapsing evidence.

Useful crosswalks include: - generated page to SYSCMD canonical command; - generated page to SYSENTVAR alias or variant; - generated page to SYSSUBCMD scoped subcommand; - generated usage to SYSARGS argument model; - HELP topic to SYSHELP; - command diagnostic to SYSMSG; - function-command bridge entry to SYSFUNC; - generated slug to reviewed final slug; - command-family owner to public command entries.

Safe wording: - Crosswalks may be candidate, partial, or verified. - Crosswalks should preserve uncertainty. - Crosswalks should not delete evidence. - Crosswalks help future metadata absorb temporary generated-page evidence.

Publication readiness

A generated command page is not publication-ready by default.

Before a command page becomes final reference prose, review should check: - canonical command identity; - alias and variant treatment; - duplicate row handling; - slug selection; - HELP alignment; - argument model; - public surface status; - internal owner exposure; - runtime behavior where behavior is claimed; - source ownership where ownership is claimed; - metadata crosswalk; - CMDHELPCHK consistency; - no unresolved canaries.

Safe wording: - Command-reference publication is a later gate. - Draft generated pages are review material. - This section does not publish the final reference. - Publication should be guarded by evidence, crosswalks, and human acceptance.

No-delete and no-mutation safety

Command-reference hygiene must preserve generated evidence during manual draft assembly.

Default boundary: - no generated command page deletion; - no HELP mutation; - no META mutation; - no CMDHELPCHK mutation; - no catalog apply; - no source edits; - no runtime data mutation; - no production SelfDoc metadata promotion; - no final publication.

Safe wording: - Review can flag generated pages. - Review can recommend later repair. - Review must not delete generated pages during manual draft assembly. - Production mutation requires explicit authorization.

Slow-lane canary tracking names

The slow-lane review tracks these canaries by exact name. These names are review anchors, not final user-facing prose.

These anchors preserve the canaries that the prose discusses in ordinary language. They should remain until the section is promoted through evidence review.

Review notes before PIP-003

This is a slow-lane section. Before generic PIP-003 is allowed to create a reviewed-candidate path, an MDO slow-lane evidence review should check:

Recommended required tokens for later PIP-003: - command reference - generated command pages - HELP_COMMANDS - HELP - HELP GIANT - CMDHELP - CMDHELPCHK - SYSCMD - SYSSUBCMD - SYSENTVAR - SYSARGS - SYSHELP - alias - variant - canonical command - slug collision - duplicate command - APPEND BLANK - APPEND_BLANK - LOAD guard - SET-family - AGGS - internal owner - public surface - manualgen - PIP - crosswalk - report-only - no mutation

Boundary

Pippets used: - PIP-001 Target Selection - MDO-196 Target Selection - MDO-197 Draft Fill

Evidence boundary: - Runtime proves behavior. - Source defines implementation and subsystem ownership. - HELP explains. - Metadata organizes. - CMDHELPCHK validates. - SelfDoc preserves provenance. - Manualgen assembles.

Slow-lane warning: - This section touches promoted draft review, review packet practice, candidate note cleanup, reviewed candidate status cleanup, header normalization, path repair, canonical path verification, table of contents checks, section order checks, final publication boundaries, generated command page no deletion, and no mutation safety. - Do not send this directly to generic PIP-003. - Run a slow-lane promoted-draft readiness review first.

Evidence tokens under review: - promoted draft - review packet - inspection packet - human review - candidate note - reviewed candidate status - header normalization - publication readiness - final publication - table of contents - section order - section count - path repair - canonical path - slug - generated command pages - no deletion - HELP - CMDHELPCHK - metadata - SelfDoc - manualgen - PIP - report-only - no mutation

Draft notes: - This is conservative manual prose for slow-lane evidence review. - Promoted draft workspace assembly is not final publication. - Candidate note headers in promoted draft sections must not be normalized casually one section at a time. - Reviewed candidate status blocks in promoted draft sections must be handled by a systematic promoted-draft header normalization pass. - Header normalization must not rewrite substantive prose without explicit review. - Path and slug verification must use canonical section ids and inspectable files. - Table of contents and section order must be verified against actual section files. - Review packets should remain the preferred human-inspection surface before authorization. - Generated command pages and evidence artifacts must not be deleted during publication-readiness work. - Publication readiness may recommend repairs, but it should not silently publish or normalize.

Purpose of this section

This section explains how the Developer Manual should review a promoted draft workspace, prepare for header normalization, and separate publication readiness from final publication.

It follows the Command Reference Assembly section because that section exposed a concrete promoted-section path problem. MDO-194 reported a successful promotion, but MDO-194A was needed to repair and verify the canonical promoted section path. That experience proves that status reports are not enough. A promoted draft must be inspectable, path-checked, section-counted, and reviewed as a workspace.

The goal is not to publish the manual. The goal is to define a safe review lane between promoted draft assembly and any later final publication step.

A promoted draft workspace is an assembled manual draft.

It may contain: - reviewed candidate prose copied into section files; - candidate note headers; - reviewed candidate status blocks; - promoted section paths; - generated section ordering; - table of contents material or future table of contents inputs; - evidence of promotion history; - known draft-workspace debt.

Safe wording: - A promoted draft is not final publication. - A promoted draft is a reviewable workspace. - A promoted draft can contain review artifacts that are useful during assembly but inappropriate for final publication. - Final publication should require a later explicit gate.

Inspectable files

The user should be able to inspect the actual prose.

Inspectable files matter because status reports can be green while path or filename issues remain. MDO-194A showed why the canonical path must be checked directly.

Safe wording: - Review must include actual section files, not only status CSVs. - Canonical paths should be opened or tested directly. - A readable section file is stronger evidence than a report that only says a section exists. - Review packets should make prose easy to inspect before authorization.

Review packets and inspection packets

Review packets are human-inspection surfaces.

They should provide: - the prose to read; - a checklist; - known canaries; - a summary; - a hold, repair, or accept decision point.

Safe wording: - Review packets should remain the preferred human-inspection surface before authorization. - Review packets should not create human decisions by themselves. - Review packets should not promote. - Review packets should help the reviewer decide HOLD, REPAIR, or ACCEPT_FOR_PROMOTION.

Candidate note headers

Candidate note headers are useful during review.

They preserve: - origin of the candidate; - promotion-gate provenance; - review warning; - draft status.

But candidate note headers can become clutter in a final manual.

Safe wording: - Candidate note headers in promoted draft sections must not be normalized casually one section at a time. - Candidate note headers should be handled by a systematic promoted-draft header normalization pass. - Candidate note headers should not be removed until the manual has a defined normalization rule. - Header normalization should preserve provenance in reports even if public-facing prose is cleaned later.

Reviewed candidate status blocks

Reviewed candidate status blocks are also useful during review.

They say that a section passed a reviewed-candidate lane, but they are not necessarily final publication text.

Safe wording: - Reviewed status blocks should not be removed casually one section at a time. - Reviewed status blocks should be recorded in normalization reports if removed from public-facing prose. - A section can be accepted for promoted draft assembly without being final-publication-ready.

Header normalization

Header normalization is the process of turning review-oriented section headers into final-manual section headers.

Header normalization may include: - removing Candidate note blocks from public-facing copies; - preserving provenance in reports; - standardizing title and section metadata; - checking section order and table of contents alignment.

Safe wording: - Header normalization must not rewrite substantive prose without explicit review. - Header normalization must be systematic. - Header normalization should report every section changed. - Header normalization should preserve originals or backups. - Header normalization is not final publication by itself.

Substantive prose boundary

Header normalization should not become stealth editing.

Substantive prose includes: - conceptual explanations; - evidence doctrine; - command behavior claims; - canary language; - examples; - scope boundaries; - future metadata feeder descriptions.

Safe wording: - Header normalization may clean review headers. - Header normalization should not change substantive prose without explicit authorization. - Substantive repairs should go through a repair/review path, not a normalization pass. - If a header-normalization script changes body prose, that is a failure unless explicitly authorized.

Path repair and canonical path verification

Promoted drafts should verify canonical paths.

Path repair and canonical path verification should check: - section id; - expected filename; - actual file existence; - readable content; - hash match with source candidate where expected; - legacy or non-canonical path leftovers; - duplicate section files.

Safe wording: - Path and slug verification must use canonical section ids and inspectable files. - A status report is not enough if the file cannot be opened. - Path repair should preserve evidence and avoid deletion unless separately authorized. - Canonical path verification should be part of publication-readiness review.

Slugs and section ids

A slug or filename is not just formatting. It controls whether humans and scripts can find the section.

Safe wording: - Slugs should be derived from reviewed section ids. - Slug changes should be reported. - Non-canonical slugs should be flagged before publication. - Slug repair should preserve old evidence until cleanup is explicitly authorized.

Section count

Section count is a basic workspace integrity check.

A promoted draft review should report: - total section files; - expected required sections; - missing required sections; - extra or duplicate sections; - sections with candidate headers; - sections with reviewed-status blocks; - sections needing header normalization.

Safe wording: - Section count is a workspace health check. - Section count does not prove prose quality. - Section count should be combined with inspectable file checks.

Section order

Section order matters because the manual should read coherently.

A section-order review should check: - onboarding and orientation before deep internals; - data model before navigation; - navigation before indexing; - indexing before expressions where appropriate; - command surface before command reference; - evidence doctrine before publication readiness; - appendices or reference sections after concepts.

Safe wording: - Section order should be explicit. - Section order should be reviewable. - Section order should not be inferred only from directory listing order. - A future table of contents or manifest should own final order.

Table of contents

A table of contents is the public navigation surface for the manual.

Publication readiness should eventually verify: - every TOC entry has a section file; - every required section file appears in the TOC or is intentionally excluded; - titles match; - order is intentional; - section ids and slugs are stable; - no draft-only sections appear accidentally; - no evidence-only reports appear as public manual chapters.

Safe wording: - Table of contents checks belong to publication readiness. - TOC readiness is not the same as final publication. - TOC generation should not delete source sections. - TOC generation should preserve auditability.

Final publication boundary

Final publication is separate from promoted draft workspace assembly.

Safe wording: - Final publication requires explicit authorization. - Promotion into a draft workspace is not final publication. - Header normalization is not final publication. - TOC readiness is not final publication. - Publication readiness may recommend repairs, but it should not silently publish or normalize.

Generated command pages and evidence artifacts

Generated command pages and evidence artifacts remain protected.

Publication-readiness work must not delete: - generated command pages; - HELP evidence; - metadata reports; - CMDHELPCHK reports; - review packets; - PIP reports; - source evidence reports; - canary reports; - promoted draft history.

Safe wording: - Generated command pages and evidence artifacts must not be deleted during publication-readiness work. - Cleanup recommendations may be reported. - Deletion or mutation requires a separate explicit authorization path.

HELP, CMDHELPCHK, and metadata boundaries

Publication-readiness review may read HELP, CMDHELPCHK, and metadata evidence.

It must not mutate: - HELP; - META; - CMDHELPCHK; - catalogs; - source; - runtime data; - production SelfDoc metadata.

Safe wording: - HELP explains. - CMDHELPCHK validates. - Metadata organizes. - Manualgen publication-readiness review remains report-only. - No mutation occurs without explicit authorization.

Report-only publication readiness

Publication readiness should start as report-only.

Report-only readiness can produce: - section inventory; - candidate header inventory; - reviewed status inventory; - canonical path check; - slug check; - table of contents check; - section order check; - review packet inventory; - unresolved canary inventory; - publication blocker list; - recommended repairs.

Safe wording: - Report-only readiness is safe by default. - Report-only readiness can recommend repair. - Report-only readiness must not silently normalize, publish, or delete evidence.

Human review

Human review remains the decision point.

Human review should decide: - HOLD; - REPAIR; - ACCEPT_FOR_PROMOTION; - READY_FOR_HEADER_NORMALIZATION; - READY_FOR_PUBLICATION_REVIEW; - NOT_READY_FOR_PUBLICATION.

Safe wording: - Human review should inspect prose, not only reports. - Human review should be recorded. - Authorization should follow inspection. - Publication should not be implied by prior section acceptance.

Slow-lane canary tracking names

The slow-lane review tracks these canaries by exact name. These names are review anchors, not final user-facing prose.

These anchors preserve the canaries that the prose discusses in ordinary language. They should remain until the section is promoted through evidence review.

Review notes before PIP-003

This is a slow-lane section. Before generic PIP-003 is allowed to create a reviewed-candidate path, an MDO slow-lane evidence review should check:

Recommended required tokens for later PIP-003: - promoted draft - review packet - inspection packet - human review - candidate note - reviewed candidate status - header normalization - publication readiness - final publication - table of contents - section order - section count - path repair - canonical path - slug - generated command pages - no deletion - HELP - CMDHELPCHK - metadata - SelfDoc - manualgen - PIP - report-only - no mutation

Boundary

Runtime Evidence, Source Verification, and Canary Closure

Pippets used: - PIP-001 Target Selection - MDO-180 Target Selection - MDO-181 Draft Fill

Evidence boundary: - Runtime proves behavior. - Source defines implementation and subsystem ownership. - HELP explains. - Metadata organizes. - CMDHELPCHK validates. - SelfDoc preserves provenance. - Manualgen assembles.

Slow-lane warning: - This section touches runtime proof, source verification, canary lifecycle, legacy evidence cautions, report-only boundaries, and no-mutation safety. - Do not send this directly to generic PIP-003. - Run a slow-lane runtime/source/canary review first.

Evidence tokens under review: - runtime - source - SOURCE - SOURCE_FACT - SelfDoc - manualgen - PIP - canary - canary ledger - smoke test - shakedown - regression - proof - verification - evidence - build - release - HELP - CMDHELPCHK - metadata - crosswalk - report-only - no mutation

Draft notes: - This is conservative manual prose for slow-lane evidence review. - Runtime proof should include concrete commands, output, build context, and date where possible. - Source verification should identify implementation ownership and should not be demoted to sidecar evidence. - Canary rows remain visible until closed with evidence. - Closing a canary requires current evidence, not optimism or old design intent. - Legacy documents may be useful context but should not be treated as current fact without verification. - Manualgen and SelfDoc remain report-only unless explicitly authorized. - HELP, CMDHELPCHK, and metadata can guide review but cannot close runtime or source canaries by themselves. - Evidence packages should not mutate HELP, META, CMDHELPCHK, catalogs, source, or runtime data during manual draft assembly.

Purpose of this section

This section explains how the Developer Manual should use runtime evidence, source verification, and canary closure.

It follows the HELP, Metadata, CMDHELPCHK, and Manualgen Alignment section because that section established the authority doctrine. This section operationalizes that doctrine into daily engineering practice: how runtime evidence is captured, how source verification is attached, how canaries remain visible, and how canaries are closed only with evidence.

The goal is not to turn every manual claim into a test case. The goal is to prevent unsupported claims from becoming permanent prose and to make sure important uncertainties remain visible until they are resolved.

Evidence doctrine in practice

The working doctrine remains:

In practice, this means a claim may need several evidence lanes before it is safe for final manual prose.

Examples: - A command behavior claim needs runtime evidence. - An implementation ownership claim needs source verification. - A HELP wording claim needs HELP evidence. - A metadata alignment claim needs metadata evidence. - A consistency claim may need CMDHELPCHK evidence. - A provenance claim needs SelfDoc or manualgen run evidence.

Runtime evidence

Runtime evidence proves observed behavior.

Useful runtime evidence includes: - exact commands that were run; - exact output or relevant output excerpts; - build configuration; - release/debug mode; - dataset or workspace used; - command path or script path; - date and session context; - pass/fail result; - any unexpected warnings or canary behavior.

Safe wording: - Runtime evidence proves what was observed in that run. - Runtime evidence should be labeled with enough context to reproduce or understand it. - Runtime evidence does not automatically define implementation ownership. - Runtime evidence can close behavior canaries when the observed behavior is current and relevant.

Examples of runtime evidence lanes: - smoke test; - shakedown; - regression run; - build output; - release verification; - manual command transcript; - script output; - comparison before and after a repair.

Source verification

Source verification attaches implementation and ownership evidence.

Useful source verification includes: - source file path; - header path; - function or class name; - source-comment contract; - source-miner report; - SOURCE_FACT row; - SelfDoc source-comment evidence; - handler or parser ownership note; - subsystem boundary note.

Safe wording: - Source defines implementation and subsystem ownership. - Source verification should identify the relevant file or owner where possible. - Source may verify why runtime behavior occurs. - Source is not merely a sidecar; it is authority for implementation structure.

Source verification is especially important for claims about: - parser dispatch; - command handlers; - storage ownership; - relation traversal ownership; - expression evaluation ownership; - memo payload lifecycle; - message emission; - backend boundaries; - generated/help/metadata alignment.

Runtime and source together

Runtime and source answer different questions.

Runtime answers: - What happened in this run? - What behavior can be observed? - Did the smoke/shakedown/regression pass?

Source answers: - Which subsystem owns the behavior? - Which implementation path handles it? - Is the observed behavior intentional, scaffold leakage, compatibility behavior, or a bug?

Safe wording: - Runtime proves behavior. - Source defines ownership. - A strong manual claim often needs both. - HELP, metadata, and CMDHELPCHK can guide the search, but they do not replace runtime/source evidence.

Canary lifecycle

A canary is a visible uncertainty, risk, or proof-sensitive behavior that must not disappear silently.

A canary can be: - open; - deferred; - narrowed; - reproduced; - partially closed; - closed with evidence; - superseded by a better canary; - converted into an issue or work item.

A canary should not be: - hidden because it is inconvenient; - closed because the expected behavior seems obvious; - closed because old design intent says it should work; - closed using HELP, metadata, or CMDHELPCHK alone when runtime/source proof is required.

Opening a canary

A canary should be opened when evidence shows a behavior, mismatch, ambiguity, or risk that could affect manual accuracy or project correctness.

Good canary records include: - short name; - observed behavior; - evidence source; - date or session; - affected subsystem; - proof needed for closure; - current status; - next recommended action.

Examples: - SET ORDER active tag reporting. - x64 workspace ERSATZ load reporting. - MIN/MAX command/function ambiguity. - AGGS internal family owner exposure. - memo backend attach on workspace open. - generated command page duplicate or slug collision.

Keeping canaries visible

Canaries remain visible until closed with evidence.

Visibility matters because older project documents, generated reports, and runtime sessions can overlap. A canary keeps the manual honest by saying: this claim is not ready for final prose, or this behavior needs proof-sensitive wording.

Safe wording: - Canary rows remain visible until closed with evidence. - Canaries can be deferred, but deferred does not mean resolved. - Canaries can be narrowed when evidence shows only part of the risk remains. - Canaries should be carried forward in review reports until closed, superseded, or explicitly transferred.

Closing a canary

Closing a canary requires current evidence.

Closure evidence may include: - runtime run showing the behavior; - source verification showing the implementation path; - regression test output; - smoke test result; - CMDHELPCHK report when the canary is about HELP/catalog consistency; - metadata readback when the canary is about metadata coverage; - manualgen report when the canary is about assembly output; - explicit human decision when the canary is a policy or surface decision.

Safe wording: - Closing a canary requires evidence. - The closure record should name the evidence. - The closure record should preserve enough context to audit the decision later. - Old design intent is not closure evidence by itself.

Legacy documents

Legacy documents are useful context.

They may contain: - old design intent; - old architecture notes; - historical problems; - early terminology; - superseded assumptions; - useful examples; - project memory.

But legacy documents should not be treated as current fact without verification.

Safe wording: - Old documents remember. - Recent summaries steer. - SelfDoc verifies. - Runtime proves. - Source defines. - Manuals explain.

Legacy evidence should be labeled when it is used: - historical context; - design intent; - superseded note; - unverified claim; - candidate canary; - verified current behavior.

Smoke tests, shakedowns, regressions, builds, and releases

Different runtime evidence lanes serve different purposes.

Smoke test: - quick proof that a feature or path starts and behaves basically as expected.

Shakedown: - broader exploratory runtime evidence, often with manual commands and transcript notes.

Regression: - evidence that a previously fixed or expected behavior still works.

Build: - evidence that code compiles and links in a given configuration.

Release: - evidence attached to a release-ready state, usually after build and smoke checks.

Safe wording: - Build success is not behavior proof. - Smoke success is limited behavior proof. - Shakedown is useful runtime evidence but should be scoped. - Regression evidence is strong for previously known behavior. - Release evidence should identify build configuration and major checks.

REGRESSION and TEST as proof launchers

REGRESSION and TEST make repeatable proof entry points discoverable, but neither command makes a run self-proving. Retain the exact command, selected script or suite, build identity, transcript, and result counts when using either launcher as evidence.

REGRESSION

REGRESSION [USAGE|LIST|SHOW <name>|RUN <name>|<name>|ALL] selects from a curated set of stable regression and shakedown scripts. LIST and SHOW expose the curated entries; RUN <name> and the compact <name> form launch one entry; ALL launches the defined ordered set. The command delegates execution to DOTSCRIPT, so it is a catalogued launcher rather than a separate test engine. Developer reproduction canaries that are not in the curated set remain outside ALL.

TEST

TEST <scriptfile> [<logfile>] [VERBOSE] runs a specified test script through the shell test harness. The harness resolves the script path, handles supported inline comments and continued logical commands, executes those commands, and reports processed and error counts. A supplied logfile may be created or truncated. The script controls the resulting side effects, so TEST must not be classified as read-only merely because it is used for verification.

Evidence boundary

Launcher availability proves only that the entry surface exists. A proof claim still requires the retained run inputs and outcome. Cross-reference the Scripting and Control Flow section for script semantics and Runtime Operation, Invocation, and Automation for entry-path behavior.

HELP, CMDHELPCHK, and metadata in evidence practice

HELP, CMDHELPCHK, and metadata are powerful review guides.

They can: - identify expected command vocabulary; - expose missing help or command topics; - organize arguments, messages, variants, and functions; - detect catalog/help drift; - point to canaries; - guide runtime/source verification.

They cannot by themselves: - prove runtime behavior; - define implementation ownership; - close runtime canaries; - close source ownership canaries.

Safe wording: - HELP explains. - CMDHELPCHK validates. - Metadata organizes. - Runtime/source evidence closes runtime/source canaries.

SelfDoc and manualgen evidence

SelfDoc preserves provenance. Manualgen assembles.

SelfDoc evidence may include: - source-comment contract reports; - SOURCE_FACT rows; - source inventory; - classifier reports; - diagram reports; - canary ledgers; - provenance notes.

Manualgen evidence may include: - pippet run records; - target selection reports; - draft fill reports; - slow-lane reviews; - PIP-003 evidence gates; - PIP-004 reviewed candidates; - PIP-005 human decisions; - PIP-006 promotion patches; - promoted draft workspaces.

Safe wording: - SelfDoc and manualgen are evidence systems. - They default to report-only. - They must not mutate HELP, META, CMDHELPCHK, catalogs, source, or runtime data during manual draft assembly. - Production mutation requires explicit authorization.

Evidence crosswalks

Crosswalks connect evidence lanes without collapsing them.

Useful crosswalks include: - runtime transcript to canary row; - canary row to source file; - source file to SOURCE_FACT; - HELP topic to command identity; - command identity to SYSCMD; - function claim to SYSFUNC; - diagnostic claim to SYSMSG; - argument claim to SYSARGS; - alias/variant claim to SYSENTVAR; - manual section to evidence tokens; - PIP report to promoted draft section.

Safe wording: - Crosswalks preserve traceability. - Crosswalks can be candidate, partial, or verified. - Crosswalks should not pretend a weak source is strong evidence. - Crosswalks help future metadata feeders absorb temporary evidence.

Future META alignment

This section should eventually align with evidence metadata.

Expected future feeders and evidence lanes: - SOURCE_FACT for source/comment provenance and implementation ownership evidence. - SRCFILE, SRCBLOCK, SRCLINE, SRCUSAGE, SRCCLASS, SRCDISP, SRCALIAS, and MEMO_LINES for SelfDoc source-comment evidence where available. - SYSCMD and SYSSUBCMD for command and subcommand identity tied to evidence. - SYSFUNC for function evidence and function-command bridge canaries. - SYSMSG for diagnostic and message canaries. - SYSHELP for help alignment and curated help evidence. - SYSARGS for argument-shape evidence and validation surfaces. - SYSENTVAR for aliases, variants, and command/function entry variants. - Manualgen PIP reports for gate evidence and assembly provenance. - Canary ledger reports for open, deferred, reproduced, narrowed, and closed canary rows.

No-mutation safety

Evidence packages should not mutate production artifacts during manual draft assembly.

Default boundary: - no generated command page deletion; - no HELP mutation; - no META mutation; - no CMDHELPCHK mutation; - no catalog apply; - no source edits; - no runtime data mutation; - no production SelfDoc metadata promotion; - no final publication.

Report-only work is the default.

Slow-lane canary tracking names

The slow-lane review tracks these canaries by exact name. These names are review anchors, not final user-facing prose.

These anchors preserve the canaries that the prose discusses in ordinary language. They should remain until the section is promoted through evidence review.

Canary non-disappearance boundary

This exact review anchor is intentionally retained for slow-lane evidence review:

The anchor is not final user-facing prose by itself. It preserves the boundary already described in this section: canaries remain visible until closed, superseded, transferred, or explicitly deferred with evidence. ### Review notes before PIP-003

This is a slow-lane section. Before generic PIP-003 is allowed to create a reviewed-candidate path, an MDO slow-lane evidence review should check:

Recommended required tokens for later PIP-003: - runtime - source - SOURCE - SOURCE_FACT - SelfDoc - manualgen - PIP - canary - canary ledger - smoke test - shakedown - regression - proof - verification - evidence - build - release - HELP - CMDHELPCHK - metadata - crosswalk - report-only - no mutation

Boundary

Diagrams

12 promoted diagrams bound from the attachment matrix (docs/manuals/developer/manualgen/reports/diagram-publication-attachment-matrix-v1.csv), each placed at its manual target and carrying its proof level.

Diagram Source asset Manual target Proof
DIAG-SELFDOCCORE-001 docs/manuals/assets/diagrams/evidence_to_manual_pipeline_v1.svg DEV-14; DEV-15; DEV-19 generated-reviewed
DIAG-SDLCPARALLEL-002 docs/media/LabTalk_DotTalkpp_Systems_Storyboard_Deck/slide-*.png DEV-19 generated-reviewed
DIAG-SDLCCURATION-003 labtalk/diagrams/LABTALK_SDLC_DIAGRAMS_v0.md DEV-15; DEV-19 generated-reviewed
DIAG-CMDAUTH-004 docs/manuals/assets/diagrams/command_reference_harvest_v1.svg DEV-05; DEV-17 source-evidenced
DIAG-X64GEOM-005 docs/manuals/assets/diagrams/trinity_headers_v1.svg DEV-08 runtime-evidenced
DIAG-X64SELFDESC-006 docs/manuals/assets/diagrams/x64_self_describing_dbf_v1.svg DEV-08 source-evidenced
DIAG-INDEXARCH-007 docs/manuals/assets/diagrams/index_order_cdx_lmdb_v1.svg DEV-09 runtime-evidenced
DIAG-WORKAREA-008 docs/manuals/assets/diagrams/workareas_cursor_control_v1.svg DEV-07; DEV-12; DEV-13 runtime-evidenced
DIAG-ARCHLAYERS-009 docs/manuals/assets/diagrams/dottalkpp_architecture_layers_v1.svg DEV-04 source-evidenced
DIAG-LABCAMPUS-010 labtalk/diagrams/labtalk_campus_map_v0.svg DEV-19 references only generated-reviewed
DIAG-LABPROOF-011 labtalk/diagrams/labtalk_proof_path_v0.svg DEV-19 references only generated-reviewed
DIAG-MDOFLOW-012 labtalk/diagrams/mdo_manualgen_publication_dataflow_v0.mmd DEV-15; DEV-19 generated-reviewed

Appendices

part pending generator (manualgen build-dry-run (section candidate))

Glossary

Review candidate — terms harvested from command/function names and section headings; definitions are pending maintainer review (candidate mode). Note: the DotTalk++ GLOSSARY command is documented in the Command Reference, distinct from this back-matter glossary.

Term Definition
CDX Compound index file (multiple ordered tags).
DBF dBASE-family table file; x64base uses a 64-bit-widened variant (DBF_64).
FPT Memo (variable-length field) file paired with a DBF; x64base uses FPT64.
LMDB Lightning Memory-Mapped Database backend used for index/order acceleration.
MDO Master Documentation Organizer — lane/section/promotion structure.
RECNO Record number; x64base widens record addressing to 64-bit (RECNO64).
SelfDoc The report-only documentation-generation subsystem.
WAL Write-ahead log providing durability for buffered mutations.
ABS definition — review
ACOS definition — review
ALLTRIM definition — review
APPEND definition — review
AREA definition — review
AREA51 definition — review
ASC definition — review
ASCEND definition — review
ASCII definition — review
ASIN definition — review
AT definition — review
ATAN definition — review
ATC definition — review
AUTODBF definition — review
BETWEEN definition — review
BIBLETALK definition — review
BOTTOM definition — review
BROWSE definition — review
BROWSER definition — review
BUFFERING definition — review
BUILDLMDB definition — review
CANARY definition — review
CDOW definition — review
CEILING definition — review
CHR definition — review
CHRISTMAS definition — review
CHRTRAN definition — review
CLOSE definition — review
CMDREL definition — review
CMONTH definition — review
CNX definition — review
COBOL definition — review
CODASYL definition — review
COMMANDSHELP definition — review
COMMIT definition — review
CONCAT definition — review
CONTINUE definition — review
COPY definition — review
COS definition — review
CREATE definition — review
184 further harvested terms pending review

Index

A

B

C

D

E

F

G

H

I

K

L

M

N

O

P

R

S

T

U

V

W

Y

Z

Colophon — build provenance

This manual assembled itself. The record below is emitted by the assembler so the self-documentation claim is auditable end to end.

Assembler manual-assembler/0.1.0 (MANUAL-ASSEMBLY M3)
Manifest tools/manualgen/manual_assembly_manifest.yaml (dottalk.manual.assembly_manifest.v1)
Parts assembled 22 (of 22 declared)
Greenfield parts generated 8
Public source snapshot 8ee746dee
Python (build) 3.10.12 (target 3.12)
Build timestamp (UTC) 2026-07-20T19:17:26Z
Machine Alienware m16 R2 — MAINTAINER_ATTESTED
Accepted reader baseline EA2E12A9D3E1