Re: [gnuspeech-contact] Document formats & repository

[David Hill]

Hi Jason,

On Apr 4, 2011, at 5:37 PM, Jason White wrote:

Paul Tyson <address@hidden> wrote:

I understand and appreciate your concern for aesthetically pleasing

documentation--particularly for user manuals.  A lot of docbook-based

pages are plain-looking, but they can be dressed up quite a bit with

custom stylesheets.  Nevertheless there are limits, and it is difficult

to have total layout control.

LaTeX, ConTeXt, or even plain TeX can offer a high degree of control over
presentation, and there are tools such as TeX4HT that can convert documents to
HTML/XHTML.

For people with print disabilities, structured formats are far preferable to
PDF alone.

Fair point. For people with visual disability access to TouchNTalk could be really useful:

HILL, D.R. & GRIEB, C. (1988) Substitution for a restricted visual channel in multi-modal computer-human dialogue. IEEE Transactions on Systems, Man & Cybernetics 18 (2), 285-304, Mar/Apr (J)

I am surprised no-one has ever commercialised the work, or even shown interest, but this has been my experience, working in the field of computer access for the disabled. There's little money for research, or to introduce innovative solutions.

All the tests we did, formal and informal, on the original TextToSpeech system that underlies gnuspeech indicated that the method we developed produced speech that was more acceptable than anything else available and -- in particular -- was much easier to listen to for extended periods, but further development as GPL software has not attracted the interest I would have expected -- speech is the Cinderella of computer I/O.

But I am getting off-topic.

I get more and more put off using markup-based document preparation systems the more I think about it. Perhaps I am spoiled, but, like many volunteer workers, I haven't enough time for all those things to which I am already committed and, rightly or wrongly, think that taking a step back to a markup basis for document production would be counter-productive, whilst the results (possibly because of my own inexperience) would probably be a lot less than satisfactory. If I were producing mathematical text books I probably would have to use TeX, but I'm not. I have frequently used HTML in the past, but the appearance depends on the browser used to read it, and HTML has significant limitations. I also used nroff and troff long ago, as well as other less well known m-u-l's.

Your feedback and suggestions are much appreciated, but I haven't seen anything to change my view -- yet!

All good wishes.

david

---------