Specs and manuals: keep figure/table numbering, multilevel lists, and print-ready legibility aligned
`spec-visio-vsd-docx` covers embedding legacy .vsd diagrams into specifications, interface descriptions, and ops manuals. Those documents depend on heading structure, captions, and cross-references—if an insert breaks SEQ fields or poisons styles, whole-book updates derail. Legacy .vsd is prone to nonstandard fonts and dense legends; pilot with print preview and a TOC refresh. Bind each figure–text–table triple to a chapter id and source .vsd version in a mapping table to prevent silent text edits without diagram updates.
Spec workflow: freeze heading and caption rules first, then validate TOC fields and print margins
- In Visio, reserve caption placeholders aligned with the Word style sheet; before export, verify legend text is not below paper-readability limits so interface ids stay legible after archival printing.
- In Word, refresh TOC and figure lists on pilot pages; confirm SEQ and REF targets still resolve—if fields break, fix sectioning and styles instead of hand-editing field codes.
- Put doc id, applicable release, and linked .vsd checksum on the cover; any clause change that affects figures must bump figure or appendix versions and log a new hash.
VSD to DOCX (spec) FAQ: captions, cross-references, style pollution, text–diagram consistency
We want automatic cross-refs between multilevel body numbering and figures—will .vsd→DOCX need manual bookmarks?
Usually yes: insert captions and REF fields with a pilot refresh; full automation is unreliable—prime the template with the smallest viable field skeleton before mass apply.
Figures updated but prose still cites old interface ids—what evidence gap blocks internal audit?
Missing paired .vsd/DOCX hashes and change-ticket linkage; enforce “figure change forces text change” or attach a delta note—otherwise treat as configuration drift.
Mixed CJK/Latin terms wrap badly—fix in Visio text boxes or Word paragraph styles?
Loosen boxes and unify East/West paragraph styles in Visio first; in Word check Asian typography grids, splitting language appendices if compression persists.
Paper sign-off differs from on-screen pagination—who owns the baseline?
Use the target printer’s print preview as the signed baseline and write that into policy; screen zoom is not a substitute for paper readability QA.
Multiple appendices and .vsd sources—how do we stop prose from citing an unregistered figure build?
Map every appendix letter and figure id to a source checksum; broadcast release manifests; ban personal-drive stray copies posing as official.