Wednesday, December 28, 2011

Added Bengali language support in Indic typing booster

   Finally done with Bengali support for indic-typing-booster, working from last 5-6 days, actually we could have done this yesterday itself but initially i thought probhat (mostly used in community) is phonetic layout like itrans but later understood it is one-to-one mapping layout. So thought its better if to add phonetic layout as well since mostly new user find phonetic layout more user friendly.

   Hunspell word list made life little bit easy, but its huge wordlist around 3Lakh (.3 Million), not sure how much actually useful for Booster IME. If we can delete few words not related or very similar it can help to reduce Database size. Will discuss this in Bengali community and will take some inputs.

   Unlike first released of other language, from onward User no need to install all layouts rpm together, done sub-packaging in spec file. Users familiar with Probhat should install bengali-typing-booster-probhat, users familiar with inscript should install bengali-typing-booster-inscript and phonetic one should install bengali-typing-booster-phonetic. Considering db size hoping this will help to reduce download and install size. Rpm's are available at koji.fedoraproject.org/koji/taskinfo?taskID=3608241, scratch build. Raised Fedora new package request hoping to get quick review and build it for Fedora 17 and Fedora 16
   This is Beta release 0.9.0, with some bugfixes and comment will release 1.0.0 soon.

Monday, November 21, 2011

Wikimedia hackthon 1st day

Amazed with the passion and spirit of wikimedia team during hackthon. Eric gave good starting explanation of basic things of wikimedia and other ongoing activities, such that offline support, mobile support, internationalization.

Hackthon project list was excellent http://www.mediawiki.org/wiki/India_Hackathon_2011#Topics

I was interested in all but since all activities were parallel thought better to check Lohit fonts support for remaining Indian languages.

I was the leader for font testing activity http://www.mediawiki.org/wiki/India_Hackathon_2011/Schedule_notes#I18N
Glad to see 4-5 hackthon attendees shown there interest in font testing.

Pre lunch session went in introduction of project, group forming etc. Actual work start after lunch.

first 1 hrs i was just enabling students OS for Indian language, couple of had Fedora in there machine so did it quickly, santhosh helped for enabling Ubuntu for Indian languages.

It was great to see there expression's when they actually started writing in Indian languages. One of attendees (Jatin) father works in CIIL :)

Then explained students important of activity they are going to do, then finally we started.
http://etherpad.wikimedia.org/LohitFonts This is the page where finally we put all the testing reports.

In between we had good offline discussion regarding Lohit Tamil fonts, Discussion with Amir regarding on 1 language 2 script problem.

Ended first session with Group Photo  http://upload.wikimedia.org/wikipedia/commons/a/a7/Hackathon_Mumbai_2011_Groupshot.jpg
 

Friday, November 18, 2011

My 1st day at Wiki conference at Mumbai university

  First thing notices is the security, did not allowed anyone without checking it photo ID and wiki invitation, even inside university there was good security. (Many Police)    

  Later understood this is due to controversy created by BJP http://www.newsreporter.in/bjp-youth-activists-detained-for-protest-against-wikipedia anyway that's other part.     

  From last couple of days seeing excellent coverage by media for wiki conference(even there was new in local Marathi language newspaper Sakal) and same noticed even in conference. First row was mostly occupied by Media.     

  University convocation hall was small compared to COEP Pune conference hall, it proved land/space crunch in Mumbai. ;)    

  Met with Ramki and Santosh and introduction with wikimedia developer. I was interested in attending tech talk session but seminar hall was full, even few people sat on ground. So went outside and had good discussion on indic computing issues and plan for hackthon. Problems and development of Lohit fonts.
     
  Attended Post lunch session about introduction of wikipedia, in that understood wikipedia also facing same problem, we faced 1 year back, one language multiple script, so asked question regarding same in Q&A. I think they are planning good to transliterate language content of one script to another script, so it will save lots of effort and will help people knowing either of language. Since limited time decided to take further discussion offline, so will discuss more in tomorrow's hackthon session.     

  From couple of session i attended wikipedia made it clear regarding there high priority task is to provide fonts, input method for at least 22 official languages of India. It is like providing food. cloth and shelter Only few are missing though, i am working on this from last 1-2 year and my dream to announce sometime "Fedora now support all 22 Indian languages", only 4-5 language support is remaining now major part is testing and confirming the support, so in tomorrow's hackthon planning to highlight this testing points. If any bugs raised i am definitely there to fix. I am sure withing next 5-6 month we will able to announce support for 22 Official Indian languages.
     
  Second part was discussion with the attendees, met with dhananjay Aditya, saw him while discussing regarding adding pali language contents in wikipedia with "Alolita sharma". I am also very interested in doing something for Pali language, we have lots of great Buddhism content available in Pali language and high time to digitize it. Had good discussion with Dhananjay he is very passionate regarding Pali looking forward to work with him together in between Dhananjay is Admin of "Superstition Eradication Committee" http://www.facebook.com/groups/ansindia/ and very active in many activities.           Was interested in showing i18n development, specifically Marathi language to IBN lokmat representative Amruta, gave 5-10 explanation to her. Lets see if anything happen positive for raising awareness of Open Source indic language computing activities.

Saturday, November 05, 2011

Fudcon second day

    It was excellent day, unfortunately was not able to attend till last session. Started with the session on Fedora security, it was interesting and definitely fedora user felt proud to see effort of security team.

    My session with Anish was from 2 to 3 PM at seminar hall 2, started bit late since it was experience from first day that audience need some time to change rooms. It was about 20-25 audience but good part was most of them were interested in doing something for Indian language computing.
    Session was very interactive. following questions were raised during session.
    1. Can we add barah layout support? i think yes we can
    2. Since same data used across different users, is there any chance is privacy issues? No. since each user will have separate user db saved in his home folder
    3. is this available in debian? still not need one packager, but still one can install and use it on ubuntu and debian easily since ibus is already there.
    Best part was when in session after showing demo and architecture we asked regarding Advantages of Booster ime, audience able to tell its advantages perfectly.
    Sessions slides are available on fudcon.in.
    Outcome: Had good interaction with all. Decided some project to work with students. Looking forward to sometime visit COEP again and actually work on some task with student.        
Thanks fudcon for giving me this opportunity.

Friday, November 04, 2011

Excellent first day at Funcon 11 in Pune!!

    Jared smith gave excellent talk on Fedora and motivated lot to student to come forward and start contribution. He explain just using in not means contribution to community but one should come forward and report bugs give some feedback is taken very well.
    Saw many desktop powered by Fedora 15, gnome 3 :)
    Amit shah's session on git was very interactive and i think it was required some more time but yes he managed well in available time.
    Had a small running chat with Jared smith, briefed him on Fedora indic i18n activities and number of Indian languages and script. (Lohit fonts). Will be good if tomorrow get some time i can show some demos to him.
    Met with Prof Abhijit (COEP) was interested in meeting him from long time. He is great supported and promoter of opensource. I had meeting with some of his project student working on Gnome Terminal indic rendering stuff. Had a nice talk on indic language computing project. There is still lots of things to do and that also interesting for Computer science student. Will give show project demos and plan for some project tomorrow.
     Big news came after lunch with the Inauguration of fuel website, it is done by Satish Mohan. There are many open task for student do go through it if one wants to learn and contribute in opensource.
    Attended talk Vaidik kapoor's on how development of Fudcon.in  happened Case study was great and overall excellent work in limited time. I am sure we will do great next time and can use same experience for other such events. COD with Drupal is really excellent module for developing event website. Liked this session.
    Last talk of Aman was nice, He explained Fedora QA process very nicely. Still i think there need some motivation for students to test and reports bugs for Fedora. May be in Fedora new somewhere name regarding the person name with number of bugs he reported.
    My session is tomorrow and very excited about it.

Pravin Satpute
Fedora i18n team

Thursday, October 20, 2011

Soon bodo (brx_IN) locale will be available in glibc

few days back saw note on tdil-dc.in regarding submission on Bodo language locale in Unicode CLDR.

found it www.unicode.org/cldr/trac/browser/trunk/common/main/brx.xml

bug for glibc http://sourceware.org/bugzilla/show_bug.cgi?id=13282

Sanjib Narzary and me working on this. Hoping soon glibc will push in in upstream git repo

Wednesday, August 24, 2011

Nastaleeq script font now available in Fedora

  Just completed packaging of Urdu Arabic script fonts from CRULP for Fedora. Good news is now Fedora has fonts for Nastaleeq script as well.
  Following is its view with Harfbuzz NG
  Looks good to me, might be person belonging to Arabic script can comment more on same.

   Following are the name of packages added to Fedora.

nafees-naskh-fonts-2.01-2.fc15.noarch
nafees-nastaleeq-fonts-1.02-2.fc15.noarch
nafees-tehreer-naskh-fonts-1.0-2.fc15.noarch
nafees-riqa-fonts-1.0-2.fc15.noarch
nafees-pakistani-naskh-fonts-2.01-2.fc15.noarch

  cheers..

Wednesday, May 25, 2011

Indic internationalization new developments and improvement in Fedora 15 (Lovelock)

   Fedora 15 (Lovelock)  got release with excellent Gnome shell. I liked gnome-shell so much that i have started using it after F15 Beta release itself ;)
   I am writing some of the noticeable development and improvement specially from indic internationalization point of view in F15.

1) New Indian Rupee Symbol
    Jul 15, 2010 Govt. of India announced about New INR symbol, http://blog.foradian.com/ made a mess by adding it on non-standard (on ) codepoint in excitement.
 During Fedora 15 Development cycle Unicode approved U+20B9 to New INR symbol and it is available in Unicode 6.0 (i guess quickest one in Unicode history thanks to DIT)
 On Qwerty keyboard AltGr+4 (Right Alt Key) allocated to it.
 Now Fedora 15 supports it and most importantly with Standardize way !!

Question: How to use it?

2) Indic Typing Booster
     This is predictive text input method for Indian languages, presently it supports Hindi, Marathi and Gujarati with itrans and Inscript layout. (Other languages will get supported soon)
    Though this is in Alpha stage it has great potential and presently in active development stage. Do give a try to this.
    Screencast of this http://www.youtube.com/watch?v=CTYNVP7p-xY

3) New Fonts packages:
i) pagul-fonts:
    A TrueType Font, which allows you to read and write in Saurashtra Script.
 ii) tabish-eeyek-fonts:
    A TrueType Font, which allows you to read and write in Meetei Mayek script.
Note: Pagul and tabish both are in Fedora 15 testing repo, install it with
$sudo yum install --enablerepo=updates-testing  pagul-fonts tabish-eeyek-fonts
 iii) nhn-nanum-fonts:
    Nanum fonts are collection of commonly-used Myeongjo and Gothic Korean font families, designed by Sandoll Communication and Fontrix. The publisher is NHN Corporation.
 iv) nhn-nanum-gothic-coding-fonts:
    Nanum Gothic Coding fonts are set of Gothic Korean font faces suitable for source code editing, designed by Sandoll Communication and published by NHN Corporation.
 v) thai-arundina-fonts:
    Arundina fonts were created aiming at Bitstream Vera / Dejavu compatibility, under SIPA's initiation. They were then further modified by TLWG for certain aspects, such as Latin glyph size compatibility and OpenType conformance.

3) Indic Rendering Improvement:
Lohit Devanagari Release 2.4.5:
          Added new ligature द्ध्र्य , Resolved shirorekha problem ग्मि, Customized glyphs for Nepali with locl feature (First indic fonts to have 'locl' feature implemented)
QT/Pango:
    Fixed most of the qt/harfbuzz indic rendering issues, In pango smaller improvement in Backspace processing for some indic characters.

4) Language Support
   i) Kashmiri
         a) select kashmiri keyboard layout from ibus-preferences (Kashmiri->inscript)
         b) Lohit Devanagari supports Kashmiri Characters recently added in Unicode 6.0
   ii)  Sindhi
     a) select kashmiri keyboard layout from ibus-preferences (Sindhi->inscript)
     b) Lohit Devanagari supports Sindhi Characters
5) bharati-m17n
    bharati is an enhancement to input method based on Inscript layout and implements unique transformation rule of deleting on the fly the previous dependent vowel for Indic languages.
6) Packages Latest Version:
Liberation Fonts:
Added U+040D and U+0400 for Bulgarian language.
For Font Developers:
glyphtracer-1.3-1.fc15 (Glyphtracer takes an image of letters. It detects all letter forms and allows the user to tag them.  They are then vectorised and passed on to Fontforge for fine tuning.)
IBus Sayura 1.3.1 release. (The Sayura engine for IBus platform. It provides Sinhala input method.)

In this mostly writing indic related stuff but considering overall for i18n there are lots of things available in Fedora 15, so enjoy Fedora 15 with excellent and very useful i18n updates as well !!

Thursday, April 28, 2011

Fedora 15 support Kashmiri language using Devanagari Script !!!!!!!!

This is really excellent feeling, it was 2007 when we first identified problems with Kashmiri Devanagari
  1. Few key Matras (U+0956, U+0957) and corresponding vowels (U+0976, U+0977) were missing in Unicode (Now available from Unicode 6.0)
  2. localedata was not available for Kashmiri Devanagari (http://sourceware.org/bugzilla/show_bug.cgi?id=6856)
  3. Not much knowledge on how to deal with issues of One language multiple script in glibc locale (using lang@script )
  4. No Fonts supporting Kashmiri  (Now its Lohit)
  5. No Keymaps (ks-inscript.mim)
    Finally today i can say that all these issues are resolved and one can type kashmiri words without any problem.

    It took long time though but standardization is time consuming process :(. most of the time taken in proposal acceptance by Unicode.

    Thanks to all who supported this, specifically from , Department of Information Technology (Swarn Lata Madam, Director & HoD, TDIL Programme and Manoj Jain) also from community M.K. Raina, Raman Kaul and Rakesh Pandit. I really appreciate effort done by Mr. M.K Raina in Kashmiri language. We used those as a reference in standardization process.

    Kashmiri Devanagari community will be happy to know Fedora 15 support Kashmiri Devanagari, Now one can see kashmiri content, create Kashmiri content in Fedora. This all in standardize way with Unicode.

    Surely this will boost kashmiri localization activities.