Request for Pandoc heading IDs (was: "Dogfooding" the LwDITA spec ?)

["Wegmann, Frank" <Frank.Wegmann@softwareag.com>]

Please find a reply of Fred Baube on the topic of processing the LwDITA spec, originally forwarded to me by Carlos. That is certainly an interesting project for folks with enough spare time and if he comes back with results, I'll let the group know.

The latter half is about Pandoc-style heading identifiers, which I think we don't mention anywhere. It would be good if anybody could provide me with some historical information whether this has been discussed before so I can give a qualified answer.

Thanks,
f.

-----Original Message-----
From: f.baube <fbaube@iki.fi>
Sent: Wednesday, 16 November 2022 09:25
To: Wegmann, Frank <Frank.Wegmann@softwareag.com>
Cc: fbaube@iki.fi
Subject: Re: "Dogfooding" the LwDITA spec ?

Hello,

Thanks for getting back to me.

I see there is active work on the spec these days, including improved alignment (and reuse) vis-Ã-vis the new DITA spec, so I have held off on thinking too deeply about implementation details :)

As a "warm-up" for "dogfooding" the LwDITA spec, I could try writing a command line conversion tool, DITA => XDITA/HDITA. This could be applied to the source of the official LwDITA spec (and its drafts). Obviously, progress on alignment is reducing the scope of this task.

To go to your email,

> If you say that you work on a LwDITA processor, I take that you focus
> on processing LwDITA documents, which would not enable you to process
> a "richer" full DITA document. But I'm happy to discuss any ideas how
> to accomplish the original challenge, if time permits, and of course,
> even a non-automatic proof of concept is more than welcome.

You are correct that full DITA is not in scope for my effort. Currently my code takes any of XDITA/HDITA/MDITA and generates an AST in a unified format specific to my implementation. It should therefore be able to generate all three representations as outputs from any particular LwDITA AST.

I believe the goal for code I write should be to generate "Normalized LwDITA", in the manner of https://www.dita-ot.org/dev/topics/dita2dita.html, so that the DITA-OT can be used for output processing.

For MDITA parsing, I use the goldmark library (https://github.com/yuin/goldmark). Goldmark is based on CommonMark and so it might not on its own achieve full MDITA compliance.

For test data, you recommend the TRS-80 manual (https://github.com/DITAWriter/LwDITA_Code_Samples).  It does not appear to be tracking current spec work. I will watch the mailing list for more information.

* * *

Also I have one more question. I wonder if anyone has given any thought to using Pandoc-style heading notations in MDITA, i.e.

     {#identifier .class .class key=value key=value}

as described at https://pandoc.org/MANUAL.html#heading-identifiers/.

For example, to specify that a file describes a Task:

     # Installation { .task }



Best,

Fred

---
F.Baube
fbaube@iki.fi


Software AG â Sitz/Registered office: UhlandstraÃe 12, 64297 Darmstadt, Germany â Registergericht/Commercial register: Darmstadt HRB 1562 - Vorstand/Management Board: Sanjay Brahmawar (Vorsitzender/Chairman), Dr. Matthias Heiden, Dr. Stefan Sigg, Dr. Benno Quade, Joshua Husk - Aufsichtsratsvorsitzender/Chairman of the Supervisory Board: Christian Lucas - https://www.softwareag.com