indic-computing-devel Mailing List for The Indic-Computing Project (Page 25)
Status: Alpha
Brought to you by:
jkoshy
Menu
▾
▴
indic-computing-devel — Discussing the development of tools for Indian language information processing
You can subscribe to this list here.
| 2001 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(14) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2002 |
Jan
(25) |
Feb
(90) |
Mar
(41) |
Apr
(16) |
May
(8) |
Jun
|
Jul
(37) |
Aug
(35) |
Sep
(62) |
Oct
(37) |
Nov
(22) |
Dec
(7) |
| 2003 |
Jan
(16) |
Feb
(19) |
Mar
(10) |
Apr
(5) |
May
(26) |
Jun
(11) |
Jul
(35) |
Aug
(4) |
Sep
(14) |
Oct
(5) |
Nov
(5) |
Dec
(10) |
| 2004 |
Jan
(25) |
Feb
(2) |
Mar
|
Apr
(1) |
May
|
Jun
|
Jul
(10) |
Aug
(2) |
Sep
(2) |
Oct
(1) |
Nov
(9) |
Dec
|
| 2005 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(2) |
Aug
|
Sep
|
Oct
(1) |
Nov
(1) |
Dec
(1) |
| 2006 |
Jan
|
Feb
|
Mar
(1) |
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(1) |
Dec
|
| 2017 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(4) |
Dec
|
|
From: Tapan S. P. <ta...@ya...> - 2001-12-27 16:12:19
|
I would be really interested to hear about the people who are active on = these lists so far and get an idea of who they are and what they are = working on. I am a graduate student in CS at the Univ of Washington in Seattle. I = have done research in the areas of programming languages and compilers = and I have also done some work in development environments at IBM = Research in Hawthorne, NY. I did my Masters thesis designing a = high-level data representation language for modern compilers. I became interested in indic language computing while on leave from UW = last year, when I worked with the Honey Bee Project in Ahmedabad. = During this work I became very frustrated with the lack of support of = Indian languages in computing (esp. using CDAC's tools), and I thought = that this is definitely something I would like to contribute to. =20 I am thinking of turning my Ph.D. research in this direction, and = looking for good research problems that could lead to a thesis. I am = developing some of my own ideas for research projects that I would be = happy to discuss. So far I have been looking at software engineering / = programming language design issues in localization, as well as support = for local language from an application level. =20 Id also like to hear from other people working in this area to see what = kinds of research problems are out there that others are working on to = see if there is anything I could contribute or collaborate on. Also if = I do take up some work at the application level, Id like to make sure I = am coordinated with the ongoing work at the systems level you guys are = working on... I guess this is a bit off-topic from the current discussions on this = list, so excuse the intrusion please... Regards, Tapan ------------------------------------------------------ Tapan S. Parikh Department of Computer Science University of Washington http://www.cs.washington.edu/homes/tapan |
|
From: Guntupalli K. <kar...@fr...> - 2001-12-26 10:57:49
|
On Sat, 15 Dec 2001 17:10:08 +0530 "sunil" <su...@in...> wrote: > Dear Colleagues: > > An article on Dr. Pavanaja and the work of KGP. Since they are keen on open source - > should we collaborate with them for Kannada. > Yes for sure. Since KGP is backed by the local govt, any kannada implementation should support those standards. Best would be to get all info from them & put it online here (or help them put it on their site & provide a link). This could be best done by those based in bangalore itself. Regards, Karunakar |
|
From: Guntupalli K. <kar...@fr...> - 2001-12-25 07:44:12
|
On Mon, 24 Dec 2001 21:37:25 -0800 (PST) jk...@Fr... (Joseph Koshy) wrote: > > > >>>>> "as" === "Arun Sharma" writes: > > [On how to ensure we have some open-source fonts] > > as> Is there an alternative we're considering ? Since open type and true > as> type fonts are documented open standards [1], most font editing > as> tools should be able to import glyphs and reedit. > > Quite right. The point was that the internal storage format of > PfaEdit, though human readable, does not get us much more than using > TT/OpenType directly. > To be truely open source, apart from the TTF/OTF , the font tables (cmap, base, gsub, gpos etc) should be provided. In opentype context this becomes important as some tools (some??, VOLT is the only tool) dont allow modification of existing tables, they discard existing opentype tables to create new ones. So to add few more rules to an existing opentype font would not be possible in VOLT, if only the final font is avialable. > As for alternatives: METAFONT sources exist for a number of indian > scripts today. However, getting these to work outside of the TeX > environment is something I have not been successful at. There seem to > be utilities available to convert TTF fonts to METAFONT sources but > not vice-versa. > There is a MF to postscript tool at ftp://ftp.radio-msu.net/pub/tex/tex-archive/fonts/utilities/mf2ps/ This means convert mf to ps (type1) , then type1 to truetype. ftp://ftp.radio-msu.net/pub/tex/tex-archive/fonts/utilities has the full set of convertors. One useful thing that can be done is to convert MF to type1 or TT . Since MF were created from a doc writing purpose, they would be usesful in their TTF versions also. Regards, Karunakar |
|
From: Arun S. <ar...@sh...> - 2001-12-25 07:43:49
|
Arun Sharma wrote: >http://eagle:8180/arun/kandict/lookup_word.jsp >http://eagle:8180/arun/kandict/add_word.jsp > Oops, Make it: http://www.sharma-home.net:8180/arun/kandict/lookup_word.jsp http://www.sharma-home.net:8180/arun/kandict/add_word.jsp -Arun |
|
From: Guntupalli K. <kar...@fr...> - 2001-12-25 07:25:38
|
On Mon, 24 Dec 2001 20:28:31 -0800 (PST) jk...@Fr... (Joseph Koshy) wrote: > > PfaEdit is an interesting tool. > Its the only one good enough to do interesting stuff. Others like Spif (http://www.gh.cs.su.oz.au/~matty/Spif/ , still very primitive to do useful stuff), GOTE (Gnome opentype editor, gote.sf.net, dev is stalled) are nowhere near to pfaedit. And there are umpteen BDF editors. In windoz world also most tools are commercial, only good enough shareware one is Font Creator (www.high-logic.com), its a 30 day shareware, but no feature limitations (even after 30 days). Fontographer costs about $350 . So for free font desigining there is very little choice of tools. > I'm concerned that relying on its internal storage format as a primary > font specification method may be risky. Internal formats change. > Also, the internal format seems rather low-level; a human would > probably find it difficult to read and change it directly. > Consequently, people will not be able to take and improve these fonts > to the extent that we wished they could. > In that case Type1/TTF format is open enough, for anyone to write tools (or use available ones) to edit them. So fonts can be given as TTFs itself. Regards, Karunakar |
|
From: <jk...@Fr...> - 2001-12-25 05:37:27
|
>>>>> "as" === "Arun Sharma" writes:
[On how to ensure we have some open-source fonts]
as> Is there an alternative we're considering ? Since open type and true
as> type fonts are documented open standards [1], most font editing
as> tools should be able to import glyphs and reedit.
Quite right. The point was that the internal storage format of
PfaEdit, though human readable, does not get us much more than using
TT/OpenType directly.
As for alternatives: METAFONT sources exist for a number of indian
scripts today. However, getting these to work outside of the TeX
environment is something I have not been successful at. There seem to
be utilities available to convert TTF fonts to METAFONT sources but
not vice-versa.
http://jeff.cs.mcgill.ca/~luc/fonts.html
Regards,
Koshy
<jk...@fr...>
|
|
From: Arun S. <ar...@sh...> - 2001-12-25 05:04:08
|
I wanted to test the capabilities of postgresql in storing indian language text in unicode format. And much to my pleasant surprise, everything seems to be working! Here is an attempt to build kannada dictionary (you could store any indian language dictionary in this db). http://eagle:8180/arun/kandict/lookup_word.jsp http://eagle:8180/arun/kandict/add_word.jsp (The first word I added is: "Internet") I've found that Mozilla 0.9.7 still doesn't do the right thing in doing a HTTP POST with UTF-8 characters. I'll take it up on their i18n newsgroups. IE 6.x seems to work just fine though. I know, I know, before I blink, the database will have plenty of unprintable language from visitors on the net. I intend to keep a "clean" copy of the dictionary, to which I intend to push user submitted entries after screening. Feel free to contribute to the dictionary! -Arun |
|
From: Arun S. <ar...@sh...> - 2001-12-25 04:41:15
|
On Mon, Dec 24, 2001 at 08:28:31PM -0800, Joseph Koshy wrote: > PfaEdit is an interesting tool. The Kannada font sahyadri I (pre) announced yesterday was created using pfaedit. > > I'm concerned that relying on its internal storage format as a primary > font specification method may be risky. Internal formats change. > Also, the internal format seems rather low-level; a human would > probably find it difficult to read and change it directly. > Consequently, people will not be able to take and improve these fonts > to the extent that we wished they could. Is there an alternative we're considering ? Since open type and true type fonts are documented open standards [1], most font editing tools should be able to import glyphs and reedit. For eg, sahyadri is based on another kannada font called baraha (http://www.baraha.com/). I imported baraha.ttf glyphs into pfaedit and reencoded them with unicode, apart from changing the glyphs. Also interesting is a tool called VOLT from Microsoft. It's a free download from MS (no source) and helps create substitution tables in an open type font file (for eg: replace "glyph123 glyph456", with glyph 789). -Arun [1] The specifications are freely available but don't know about patents. |
|
From: <jk...@Fr...> - 2001-12-25 04:28:35
|
>>>>> "gk" == "G Karunakar" writes: gk> Pfaedit stores all the font data as a spline font database (.sfd) (it uses cubic splines) , which is plain text. A sample for a glyph definition is given below > ------------------ > StartChar: exclam > Encoding: 3073 3073 > Width: 387 > Flags: W > VStem: 221 37<108 140> > Fore > 388 -16 m 1 > 387.333 -18.6667 379 -20 363 -20 c 0 > 333.667 -20 305.333 -10.3333 278 9 c 0 > 240.667 35.6667 221.667 73.6667 221 123 c 0 gk> So this .sfd can be cosidered as font source and can be released . If there are any reservations about the tool, I would say at present this is the only free good enough tool to do scalable font design in *nix platform. PfaEdit is an interesting tool. I'm concerned that relying on its internal storage format as a primary font specification method may be risky. Internal formats change. Also, the internal format seems rather low-level; a human would probably find it difficult to read and change it directly. Consequently, people will not be able to take and improve these fonts to the extent that we wished they could. Regards, Koshy <jk...@fr...> |
|
From: Guntupalli K. <kar...@fr...> - 2001-12-24 14:16:32
|
On Thu, 20 Dec 2001 01:23:11 -0800 (PST)
jk...@Fr... (Joseph Koshy) wrote:
>
> g.k> We need free & open source fonts
> g.k> - Define open source font .
> g.k> - font licensed under an open source license
> g.k> - all font data available in portable format , this includes foll.
> g.k> - outline data
> g.k> - mapping tables
> g.k> - opentype tables
>
> Right. As of now, I'm only aware of METAFONT technology that allows
> fonts to be described in "source" form. Is there any equivalent for
> {True,Open}Type fonts?
Truetype/Opentype font data is in binary format. tools can be written to extract outline info from them.
In this context I would like to point to Pfaedit (http://pfaedit.sourceforge.net) . Currently it is the only complete (reasonably stable, but still under development) opensource tool to design Type1/Truetype fonts.
Pfaedit stores all the font data as a spline font database (.sfd) (it uses cubic splines) , which is plain text. A sample for a glyph definition is given below
------------------
StartChar: exclam
Encoding: 3073 3073
Width: 387
Flags: W
VStem: 221 37<108 140>
Fore
388 -16 m 1
387.333 -18.6667 379 -20 363 -20 c 0
333.667 -20 305.333 -10.3333 278 9 c 0
240.667 35.6667 221.667 73.6667 221 123 c 0
219.667 176.333 239.667 216.667 281 244 c 0
309.667 263.333 340.333 273 373 273 c 0
380.333 273 385.333 272.667 388 272 c 1
379.333 270 367 267 351 263 c 0
288.333 244.333 257.333 199 258 127 c 0
258 68.3333 278 26.6667 318 2 c 0
332.667 -6.66667 356 -12.6667 388 -16 c 1
EndSplineSet
MinimumDistance: x8,-1
Ref: 255 N 1 0 0 1 0 0
Ref: 255 N 1 0 0 1 0 0
EndChar
-----------------
When we are done with design , we use the Generate Font option to generate the final font in Type1, Postscript, Truetype, bitmap or bdf formats.
So this .sfd can be cosidered as font source and can be released . If there are any reservations about the tool, I would say at present this is the only free good enough tool to do scalable font design in *nix platform.
Regards,
Karunakar
|
|
From: <jk...@Fr...> - 2001-12-20 09:23:13
|
g.k> First of all we need to use a common notation method & terminology.
g.k> TODO --> So two more things to add (notation technique & glossary
g.k> of terminology)
Point taken. A common notation & terminology used across all
languages would make a big difference to the clarity of our
communication.
This notation should probably be documented in one of the introductory
chapters of the Handbook. Could you contribute a write up on a
suitable notation? Something aligned with popular existing practice
would be ideal.
g.k> Few sugg.
g.k> - Use Unicode naming technique for script/language alphabets.
g.k> - Suitable abbreviations for commonly used phrases.
Ok.
I've used "language information" loosely in the roadmap.
g.k> Define What we mean by language information ?
g.k> For a script :
g.k> a. The basic character set - Consonants (C), vowels (V),
vowel modifiers(M), diacritics (D), punctuation etc.
g.k> b. Glyph set for this basic set.
g.k> c. Rules for syllable formation -
g.k> consonant-vowel - C V , C V M, C V D
g.k> consonant-consonant - tuples - C C , triples - C C C ,
quads - C C C C etc.
g.k> List all possible (which are actually used , now or were at
some point of time) with suitable
g.k> examples
g.k> how modifiers , diacritics , punctuation are used?
g.k> d. complete Glyph set for (c)
g.k> e. Writing rules
g.k> writing direction
g.k> word breaking
g.k> line breaking
g.k> f. Collation rules
g.k> Sort order
g.k> g. TODO - add more
I'll add this list to "Language Information" chapter of the Handbook in
section "What kind of information are we looking for".
g.k> We need free & open source fonts
g.k> - Define open source font .
g.k> - font licensed under an open source license
g.k> - all font data available in portable format , this includes foll.
g.k> - outline data
g.k> - mapping tables
g.k> - opentype tables
Right. As of now, I'm only aware of METAFONT technology that allows
fonts to be described in "source" form. Is there any equivalent for
{True,Open}Type fonts?
g.k> - List out complete glyph set for each language, which can be used as base
by font designers.
g.k> A minimal Glyph set usable for 8 bit encoded fonts
g.k> A maximal set having all possible glyphs, used by multibyte encoded
fonts - unicode fonts, opentype fonts etc.
g.k> - A howto/tutorial on creating an opentype font from an existing TTF one -
supplemented by examples, with font data.
g.k> - TODO - add more
Will add.
Regards,
Koshy
<jk...@fr...>
|
|
From: Guntupalli K. <kar...@fr...> - 2001-12-19 07:58:05
|
Hi, few suggestion on the roadmap ( http://indic-computing.sourceforge.net/roadmap/index.html ) Right now only covered first two sections. More to follow later. First of all we need to use a common notation method & terminology. TODO --> So two more things to add (notation technique & glossary of terminology) Few sugg. - Use Unicode naming technique for script/language alphabets. - Suitable abbreviations for commonly used phrases. ******************************************************************** 1) Language Information KN-i Collect language information for Kannada TL-i Collect language information for Telugu HN-i Collect language information for Hindi MT-i Collect language information for Marathi GR-i Collect language information for Gurmukhi ... repeat for other languages as volunteers contribute ... ----------------------------------------------------- This should actually go by Script -> Languages using the script . Some languages written using diff scripts can be taken seperatly. Define What we mean by language information ? For a script : a. The basic character set - Consonants (C), vowels (V), vowel modifiers(M), diacritics (D), punctuation etc. b. Glyph set for this basic set. c. Rules for syllable formation - consonant-vowel - C V , C V M, C V D consonant-consonant - tuples - C C , triples - C C C , quads - C C C C etc. List all possible (which are actually used , now or were at some point of time) with suitable examples how modifiers , diacritics , punctuation are used? d. complete Glyph set for (c) e. Writing rules writing direction word breaking line breaking f. Collation rules Sort order g. TODO - add more For each language - Info for all (a) to (f). Language rules may override script rules, esp. (c), (e) & (f). *********************************************** 2) Fonts: Create a collection of indic fonts Fn.1 Sponsor creation of fonts (corporate sponsorship?) Fn.2 Collect free fonts on the Internet Fn.3 Document existing fonts: availability, characteristics, encodings etc. Fn.4 package fonts for use in popular open-source OSes (ie. install documentation, user documents etc) ---------------------------------------------------------------- We need free & open source fonts - Define open source font . - font licensed under an open source license - all font data available in portable format , this includes foll. - outline data - mapping tables - opentype tables - List out complete glyph set for each language, which can be used as base by font designers. A minimal Glyph set usable for 8 bit encoded fonts A maximal set having all possible glyphs, used by multibyte encoded fonts - unicode fonts, opentype fonts etc. - A howto/tutorial on creating an opentype font from an existing TTF one - supplemented by examples, with font data. - TODO - add more *********************************************** Karunakar |
|
From: sunil <su...@in...> - 2001-12-15 11:48:38
|
Dear Colleagues: An article on Dr. Pavanaja and the work of KGP. Since they are keen on open source - should we collaborate with them for Kannada. Thanks, Sunil -- Sunil Abraham Team Leader - MAHITI Info-tech for the Voluntary Sector India Cares, Vijay Kiran 314/1, 7th Cross, Domlur Bangalore - 560 071. Karnataka. India Pager: +91 80 9624 279519 Ph/Fax: +91 80 5352003, 5350035 E-mail: su...@ma... Web: http://www.mahiti.org -----Original Message----- From: Frederick Noronha <fr...@by...> To: byt...@go... Date: Sat, 15 Dec 2001 01:40:52 +0530 (IST) Subject: From Kannada to keyboards... an Indian language enters the cyberage ********************************************************************** FROM KANNADA TO KEYBOARDS: AN INDIAN LANGUAGE ENTERS THE CYBERAGE ********************************************************************** By Frederick Noronha fr...@by... For Dr U.B. Pavanaja, an unlucky 1993 scooter accident turned out to be the proverbial blessing in disguise. For nine months as he lay immobilised in bed, the scientist learnt Visual Basic. Laying prostrate on his bed, with a computer alongside, he then went on to write the first versions of what is now his 'Kannada Kali' software programme. This is a game that helps a child or new learner of the Kannada language of the Southern Indian state of Karnataka to shape his alphabets properly. "I did it lying on the bed with a computer by my side," he recalls with a smile. Over the years, as he stepped up work on the issue of Indian regional language computing, the one-time scientist at India's prestigious atomic research centre finds his output increasingly relevant to the commonman. Currently he's at the helm of the Kannada Ganaka Parishat (or, Kannada Computer Association). This is a voluntary organisation formed by computer professionals, literary persons and others to promote the standardisation and usage of the Kannada language on computers. It's probably important not to underestimate the size of this task. Kannada is the language of some 47 million people worldwide -- more than the number of Polish speakers in the globe, and just below the number of Ukrainian speakers. Besides, the lessons learnt with Kannada could have important implications for other prominent Indian languages whose speakers number in millions. For instance, Hindi (496 million), Bengali (215 million), Urdu (106 million), Punjabi (96 million), Telugu and Tamil (75 million each), and Marathi (72 million). "There is so much talk about computing for the commonman. But the main problem that everyone seems to overlook is that the commonman (specially in countries like India) speaks in languages other than English," as Dr Ubaradka Bellippady Pavanaja reminds us. (Both his first names are village-names, and in the South Indian style, are generally not spelt out in full.) So, for the past many years, he's been working sweating over this front. Some solutions are simple, why-didn't-we-think-of-it-earlier ways out. Others are attempts to do the groundwork and undertake standardisation that could have far-reaching implications for the future. So far, the standardisation has already been done, both on a uniform keyboard for Kannada, and also for the glyphs and glyph-codes. (The latter refer to the component parts that, when joined together in varying combinations, make up each alphabet.) There's a big difference between English and Indian-languages over the display and storage of information in computers. In the case of English, there is a one-to-one correspondence between the display codes and the storage codes. But in the case of an Indian language, say Kannada, the letters are made up of combinations of consonants and vowels. Using, for example, a consonant-plus-consonant-plus-consonant-plus-vowel combination. These characters have a unique storage code in ISCII, or the Indian Standards Code for Information Interchange. Display of these characters are accomplished by joining pieces of characters known as 'glyphs'. Codes for the storage characters and the display pieces (glyphs) are different. In addition, the number of characters which make the make the character (used for storage) and the number of display pieces which are used for the display of the letter simply don't have a one-to-one correspondence. An example: the Kannada language uses some 142 pieces to obtain all the possible combinations that can be obtained from the based 49 Kannada alphabets. In the past, Indian groups working on language-solutions -- like the Pune-based government backed C-DAC and Mithi, which specialises in local language computing, also from Pune -- have worked on similar work. But in earlier cases, everyone followed their own glyph sets. This meant data lacked 'portability'. Text composed on one computer could not be carried over, or understood by, another computer which did not share the same software. This was a great handicap in a world where the ability of computers to 'talk to one another' has made them into the powerful tool they currently are. "We feel the best solution is to have the storage in ISCII. Other solutions have attempted to tie up the user in their own software solutions," says Dr Pavanaja. He says that the Government of India's stand is that ISCII should have standardised glyph sets. "In our region, the Government of Karnataka has standardised glyph sets already. We have benchmark software too... to ensure that the software would work with any standard computer." Admits Dr Pavanaja: "Standardisation is something that has to be imposed (for the sake of moving ahead together)." At another level, the Kannada language has also pushed for what it calls the Kannada Standard Code for Language Processing. This is used for sorting, as per the Kannada order of alphabets. "Sorting is a very important job for computers. Can youthink of a single database operation without sorting and indexing?" asks Dr Pavanaja. "For all these years, using computers for Kannada-work meant simply using it for typing, making books, printing invites and DTP (desktop publishing) work. It has now changed," points out Dr Pavanaja. Sorting and indexing in the regional language, he argues, has opened up new possibilities. C-DAC (the Government of India-backed Centre for Development of Advanced Computing) earlier had solutions, but this, he says, was not particularly suitable for the Kannada language. This attempt evolved a national standard based on Hindi, whereas every language of India has its own specialities and requirements. At another level, the Parishad has been working towards a standardised Unicode for Kannada. "KGP general secretary Srinatha Sastry and myself put together a document, and sent it to the Unicode Consortium. It was partly accepted," says Dr Pavanaja. He underlines the importance of uniformity for the Unicode character table and collation code for this regional language. Incidentally, India's voting-member at Unicode Consortium is the Indian government's Ministry of Information Technology (MIT). But lack of uniform interests among the various Indian languages used for computing means that sometimes not much can be done on this front. In September 2000, Dr Pavanaja took part in a Unicode conference in California. "We explained the issues (involved in Kannada), and that was appreciated a lot. The MIT is waiting for all languages to come up with a decision. Only Kannada has done this much groundwork on Unicode. At least Kannada could be implemented on Unicode for now (instead of waiting for all Indian languages to finish their task)." Besides, the Parishad has developed a free Kannada script software. This was released in October 2001 in Bangalore. "It has got SDK (the software development kit) as part of it. But most importantly, it comes free (in terms of price)," stresses Dr Pavanaja. He suggests that this is important too in a price-sensitive region like India, where millions still live in poverty. Using this, developers can write Kannada database applications. It could, therefore, have applications linked to phone directories, ration cards, banking, libraries and even road-transportation operations. This spells immense fallouts for this large state of Karnataka, which has a population roughly the size of South Africa, and over half the area of Germany in land-mass. "Everyone needs good database applications. In Indian language computing, 90% of the uses are linked to DTP unfortunately. But in English, computers are overwhelmingly used for database applications," says he, stressing that the lack of applications also causes problems. Whether it's e-commerce, business transactions or public utlities and governance, all these sectors need good database applications, stresses Dr Pavanaja. One of this team's solution is called 'Kalitha'. It is a Kannada keyboard driver and font. "It also has a sorting engine, not just a sorting-facility. This is the first time that any Indian language had this facility," says Dr Pavanaja. This group led by Srinatha Sastry, has modified a Kannada keyboard-layout originated by K.P. Rao. It uses the 26 English-language keys for Kannada's 49 alphabets. "Even Bill Gates appreciated (the concept behind) such a layout for a keyboard," says Dr Pavanaja. But just how does it work? The 'shift' (or 'caps') key comes to the rescue. "English has 26 alphabets multiplied by two (with each using the caps key). This makes a total of 52. In Kannada, we need only a total of 49. It works well with the 'shift' and 'unshift' key," says he. This layout has been accepted and notified by the Karnataka government. In order to keep things simple for the typist and computer-operator, this keyboard makes things a "little more difficult" for the programmer. But once that is taken care of, things become simple in actually using this solution. Besides his technical work, this man's own story is also interesting. Dr Pavanaja, currently 42, is a PhD in chemistry. He was a scientist at the Bhabha Atomic Research Centre (BARC) in Bombay. "We used computers extensively, in lab-automation and we also experimented in connecting a lot of lab equipment to computers," he recalls. Using computers "as a tool" for his scientific work for awhile, he says he "got addicted". His own efforts took the chemical scientists closer to the computer in the early days of the PCs. "I soon became seen as a computer professional," he recalls of times in the mid-eighties, when the PC first began to make its appearance in the Indian scientific establishments. In BARC, a group to promote the Kannada language often faced difficulties in publishing technical articles in its Kannada-language science magazine. That set him thinking. "While doing our magazine 'Belagu' (whose name loosely translated to 'Shine' or 'Reflect Light'), we decided to buy our own DTP package." In 1995, a visit for advanced research to Taiwan revealed that computer professionals were heavily into computer use, but were overwhelmingly using Chinese. "If they could use their language, why not we?" thought Dr Pavanaja. Soon, he became active on Internet 'news' groups like soc.culture.indian.karnataka and also set up websites. What happened afterwards is narrated in terms of the output achieved and listed above. "When I was a scientist, I felt my doctorate had no use. I was hardly doing any (socially-relevant) work. Now, I don't feel guilty about that anymore," he says. He returned from Taiwan in 1996 and resigned from BARC in June 1997. In 1998, his work made Kannada one of the first Indian languages to use dynamic fonts. He explains: "Earlier, if you wanted to browse a web-site, you needed the (same font used by the site) to be installed on your PC." Obviously, a real dilemma in a region where there exist dozens or hundreds of non-standardised fonts for each language. This meant downloading the font. You needed to do it each time you used a different computer! Dynamic fonts solve the problem by residing on the 'server', not on the 'client' (or user's computer). When you browse a site, you automatically pull the font info the first time you browse it. Also, it works with any operating system you're using, Dr Pavanaja points out. "In English, you don't have the problem of clashing glyphs. If you use a fancy font, you can still read it at least in Times or Arial...," He notes. Pavanaja has also createD a Kannada version of LOGO. "LOGO stands for 'logic-oriented, graphic-oriented' programming. It is a language for children. It uses very simple commands, like 'forward', 'backward', and so on. School children of the fifth to eight standards (roughly 10 to 13 years of age) can use it effectively. I thought of Kannada-medium schools, and wanted something for them," says Dr Pavanaja. Work done by this group could make Kannada the first Indian langauge to get onto a palm-top computing device, believes Dr Pavana. "Much of the coding (for some of our projects) has been done by K.M.Harsha, a 22-year-old mechanical diploma holder from a village," he points out. This, says the scientist, only underlines the creativity of youngsters if given the chance. It challenges the myth that city-born children are more intelligent! One of the KGP's dreams is to have Kannada working with the 'free' and 'open source' Linux operating system, which was largely build up by volunteers worldwide. "But that could take some time," concedes Dr Pavanaja. "We need to have keyboard drivers, fonts, a toolkit for software developers, a free office suite like Star Office, and even the complete Linux working in Kannada," he adds. Getting legal copies of proprietorial software would cost millions for a state the size of Karnataka. "So far the KGP has been taking its funding from the government, semi-government institutions, corporate world and philanthrophy. We need to develop software and make it available freely (so as to make it affordable to the commonman in a country where millions still live in poverty). We don't sell anything," says Dr Pavanaja. Says Dr Pavanaja: "If you don't put Indian languages into the computer, all our tongues will get relegated to being just spoken languages in five to ten years time." Currently the editor of 'Vishva Kannada', which he terms the world's first Internet magazine in the Kannada language, Dr Pavanaja can be contacted at <pav...@vi...> This magazine's site can be visited on the World Wide Web at www.vishvakannada.com (ENDS) ------------------------------------------------------------------- Frederick Noronha * Freelance Journalist * Saligao 403511 Goa India fr...@by... * Phone +91-832-409490 Mobile 9822 12 24 36 ------------------------------------------------------------------- |
|
From: <jk...@Fr...> - 2001-12-14 10:55:37
|
One of the sub-projects that Indic-computing project aims to facilitate is the creation of a bootable OS that has packaged-in support for Indian Language computing. By this we mean that a relatively inexperienced user should be able to boot into this OS (say from a CD) and have at her disposal tools to operate in Indian Languages. Such an OS distribution could serve as a starting point for - demonstration CD showcasing the progress of indian language tools - the creation of customized versions of the OS for use in "embedded" applications like rural information kiosks and the like This mail is meant to start a discussion on the requirements of such a demo OS. We need to discuss if such a project is (a) a good idea (b) and if so, what kind of features and software we should target and making part of this OS. Here are some first cut requirements: a. The OS should be boot off 'common' hardware. In the indian context this means x86 architecture PCs. x86 PCs are very diverse in hardware components; we will need to support a 'reasonable' range of hardware (for some definition of 'reasonable'). b. Do we require the OS to be 'installed' or should it run directly of the CDROM? Running directly off a CDROM is useful for demos. However, unless the boot and device probing process is made fully automatic, we'll need to ask hardware configuration related questions each time we boot and this can get quite irritating. Q: Is a feature wherein the OS keeps mutable data (user data, log files, configuration data etc) on a hard disk, while keeping all other files on read-only media useful? c. The OS should present a GUI interface to the user from the start. We could have an X server and a graphical login start off at boot time, providing we can solve the issues of X Server configuration at boot/install time. d. The application mix provided on CD should work seamlessly together. Basic capabilities for a demo CD could include: - Web browsing / creation of web pages - editing of text files (word processing) - printing of indian languages - other text processing Since we are looking primarily at open-source tools, we'll need build and packages these applications in such a manner that they all work together. e. ...others... Regards, Koshy <jk...@fr...> |
3 messages has been excluded from this view by a project administrator.
