[Savannah-register-public] [task #5082] Submission of Hindawi Vernacular Programming System

[Abhishek Choudhary]

Follow-up Comment #13, task #5082 (project administration):

Hello,

Thank you for approving Hindawi. And thanks for reading my last message! (It
was a bit long!) It shall now be much easier for me to allow people into the
project. Many have been enquiring how they may contribute their effort to
Hindawi. Besides, at LinuxAsia 2006 I'm sure I'll find enough vounteers for
every aspect of Hindawi. For now I have to complete the background task of
identifying the sub-projects and drawing up a feasible time-line for
completion. I can now work on submissions for Romenagri and APCISR (tasks
#5083 and #5084). Without these two background projects Hindawi would not be
possible; besides, they can be of benefit to all human-languages which do not
have native programming systems. Sebastian <kickino> is looking into them, but
maybe you could provide me a few suggestions or critical views since you've
been through much of the discussion related to those projects as well. It
would be a great help if those projects could also be approved before
LinuxAsia 2006 (Feb 8-10 http://www.linuxasia.net)

I have really appreciated your queries, and through them have been able to
accomplish a task hitherto procastinated - that of creation of a prelim
technical FAQ for Hindawi. Even the IT people in India, who are actually used
to computing with the latin script, often raise similar questions.

As about the UTF-8 option, I had initillaly tried it, but it has a few
drawbacks. First, although UTF-8 is backward compatible to 7-bit ASCII, it
requires 8-bit character codes and hence cannot be optimally used for
keywords, as I have explained earlier. 
Second, existing library symbols are encoded in 7-bit ASCII and and a
completely new library had to be created if UTF-8 was used instead of the
Romenagri encoding. 

Third, reading UTF-8 requires rendering support, which is available only
under GUIs for now. APCISR does support UTF-8 in text mode, but only after
converting it to 8-bit ISCII; besides, APCISR may not be available on all
systems (on those which text mode font cannot be changed, at least).

This discussion implies that UTF-8 would be uneconomical in terms of
lexer/parser speed if used for 'core' compilers, and, on systems which lack
support for rendering UTF-8, such source would not be human-readable, whereas
Romenagri does not suffer from either of these two drawbacks.

As about GCC command-line option for Indic variable names and keywords etc. -
well, whatever technology / encoding is used (Romenagri, UTF-8, Big/Little
endian UNICODE etc.), I would really like to see gcc support Shaili Guru,
Shaili Shraeni, Shaili Vyakaran or other shailis; i.e. Indic C, C++, yacc
etc. It would be wonderful to see a command like "ghin -o namaste_duniya
namaste_duniya.hin"; i.e. GNU Hindawi driver used for compilation of Shaili
Guru or other shaili (namaste_duniya = hello_world). I hope Hindawi, Romenagi
and APCISR are accepted as parts of GNU someday soon; then gcc can interface
to ghin or GNU Hindawi compiler, just as it does to g77 for now. I shall get
in touch with the authors / maintainers of gcc/flex etc. for possible
outcomes. I also hope some of the developers joining Hindawi and related
projects are interested in such a scenario. Personally I wish to devote my
effort to the technological developement towards a possible future of
programming languages which is human-language and paradigm neutral. I am also
academically interested in natural language programming. Once Hindawi has
started off properly, I can devote my resources for NLP and
paradigm-neutrality. 

Presuming that I have roused your interest regarding Hindawi, I expect you
shall continue to watch its course, and, what I need more, continue to
question, advise and criticise!

Thanks a lot,
Abhishek

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/task/?func=detailitem&item_id=5082>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/