[accessibility] What are good reasons for not releasing the source in 2019?

[Christopher Rowley]

Ross makes a very timely point here:

> One thing that we cannot do is release code that is not sufficiently robust,
> which would allow authors to create broken PDFs without being aware of the defects.

This has led me to try and clarify the current status of this work in this posting.  I hope the following achieves this aim. 

IMPORTANT NOTE: mainly for The TUG President and Office.  
You may find some of this useful in your funding bids:

——————
Ross has been working in this area for many years now, but there is still a lot to do.  The LaTeX Team (see www.latex-project.org), which includes myself and Ulrike, are helping with the basic low-level TeX programming needed to integrate this work into the “LaTeX system”, so that LaTeX can easily be used to produce “non-broken pdf” that passes all tests for being pdf/ua, and many other accessibility requirements.  

Funding —
This is a long (since complex) process that, so far, has had very little direct funding of the highly expert human resources it needs.  TUG is currently attempting to raise money to make such funding feasible.  However, this can only be useful if we can also find people to work with us who have the correct combination of specific knowledge of, for example,  pdf2.0, TeX, plus (preferably) LaTeX3 programming syntax and techniques.  

They will also need general expertise in software design and implementation plus (probably most important) being able to work productively, largely via the internet, with the existing team members, most of whom will continue to work as volunteers and thus have limited time but will nevertheless need to be fully aware of your work and are very good at contributing their insights and ideas.

Experimental work —
Thus all work in this area for sometime yet must be considered experimental and the details, and even sometimes the basic concepts, will be liable to change.  Thus is another reason why we are not widely publicising the existing possibilities and we definitely do not want the current implementations to be widely distributed.

The future —
We thus have to complete the following two things:

1.  Some essential reworking of the “LaTeX kernel” to support tagging (in general) and the output of “good pdf”; this reworking also needs to be fully accepted by the very wide “LaTeX community” so we must take great care that it does not clash with the very, very wide range of current uses of LaTeX and the hundreds of LaTeX class files and packages in use.

2.  Knowledge of how to use this reworked LaTeX kernel to produce “non-broken and accessible pdfs”.

On,y then will we be able to say that we have a stable system for the production of “accessible pdf”.

Although, for many reasons that come from politics, history etc., we are committed primarily to output pdf, we also wish to support fully other outputs such as EPUB3 so we shall not do anything that prevents LaTeX being used for these and we want to know all about the technical stuff that would help the “New LaTeX” to support such alternative outputs.

Apologies for the length and complexity of this posting.  Please question me about anything that needs further explanation.

Chris