Posted by Lopsotronic on Feb 04, 2020; 9:00pm
URL: https://discuss.asciidoctor.org/S1000D-tp6676p7570.html

kibukj, I am honored to be speaking with you "face-to-face". Your work is impressive and is probably the best S1000D-based repo on github.

It's amazing there's more than one of us interested in Asciidoc-S1000D cross-compat, but here we are.

PM me if this forum supports that, I would like to "meet" you in a more personal forum or even in person, geography depending.

Short answer to your question: we need to draft a permissive BREX for adoc that comes as close to stock S1000D SGML as we can get it. It's not out of the question to make "Ascii1000D", much like AsciiRFC project extends Asciidoctor to support RFC7991.

I'm trying to do something very much like what you're describing, using DMCs as filenames, ifdefs for "applicability", and tagged includes for CIRs (Common Information Repositories). If we could adopt "pure" asciidoctor pipelines aka prawn for pdf, textQL would be a gamechanger for parts information repositories, but the prawn PDFs . . ah well, see below.

The weak link is writers, who aren't familiar with git or version control or anything like that, at all. They're hitting the same file from different branches and getting angry about conflicts on the merge.

They're also needing more PDF customization than prawn can give them, but don't have the XSL chops to really customize the fopub docbook stack on their own. PDF customization eats a ton of my time.

At the end of the day this implementation will probably die, but this specific method is for sure going to be the way I do my next publications environment. It's just too damn efficient . . assuming you have a staff of writers who know how to git.