Author Archive

« Older Entries

A FUEL for Gujarati Localization!

Wednesday, December 30th, 2009

Localization of Open Source Software into Gujarati language was started more than five years ago. Number of people joined the Gujarati localization efforts and many left. Every release of open source distributions like Fedora, Debian, OpenSuse got better localization into Gujarati. Despite continuous efforts by individuals or groups in Gujarati Localization there was no real 100% acceptance of many translations. Even today there is no 100% satisfaction in the Gujarati end users’ group for localized Gujarati distributions.

Similar issue has been faced by other (Indic) languages as well. FUEL (Frequently Used Entries for Localization) project was started by Mr. Rajesh Ranjan from Red Hat as a part of the community contribution to overcome the above mentioned issue. FUEL project is a community project initiated by Red Hat. It is hosted on fedorahosted.org. FUEL tries to provide a standardized and consistent computer interface for users. FUEL evaluation has been successfully completed for Hindi, Maithili, Marathi and Tamil languages before Gujarati. More information regarding FUEL can be found here: https://fedorahosted.org/fuel/

FUEL Evaluation workshop for Gujarati was organized on 26th – 27th December, 2009 in Ahmedabad @ Infineon Infotech Pvt. Ltd. office, where participants from Gujarati Localization group and Linguists from Ahmedabad have taken part. Participants were:

* Mr. Siddharath Bhatt (Linguist)
* Mr. Prashant Dave (Linguist)
* Mr. Yogendra Vyas (Linguist)
* Mr. Ketan Upadhyay (Department of Languages in Gujarat Govt.)
* Ms. Maitri Shah (Infineon Infotech Pvt. Ltd.)
* Ms. Shruti Amin (Infineon Infotech Pvt. Ltd.)
* Ms. Sweta Kothari (Red Hat Inc.)
* Mr. Ankit Patel (Red Hat Inc.)
* Ms. Deval Vyas (Infineon Infotech Pvt. Ltd.)
* Ms. Padma Jadav (Infineon Infotech Pvt. Ltd.)
* Ms. Sumaiya Vohra (Infineon Infotech Pvt. Ltd.)

First day, 26th December, started around 10:15 AM with a round of introduction and then I provided a generic idea about the FUEL, since it was a new thing for many of the participants. By 11 AM we were ready to start the evaluation of Gujarati translation of FUEL entries. Sweta and Maitri began to take us through FUEL entries. Initially it used to take lot of time for each FUEL entry, since the terms are based on computer terminology and it’s difficult for us to explain them practically each and every word, as language experts were not really used to computers. In the first two hours we did complete not more than 15 fuel entries, while our goal was to complete all 578 entries of FUEL within two days. Mr. Ketan Upadhyay has played a major role there to explain things in a better way that experts understand. We all were really amazed by suggestions from Mr. Prashant Dave and Mr. Siddharth Bhatt when they started bringing new better translation words and explained us some grammatical rules to remember while translating from English. We were taking reference of Saarth Dictionary and Mr. Ratilal Chandaria’s website GujaratiLexicon.com for clarifying some doubts. By 6 PM we were done with evaluation of 248 entries. But we knew that it’s going to be faster next day, since the rhythm of evaluation has already come in and everybody was on the same track.

Next day Mr. Yogendra Vyas, who is another great linguist, joined us for the evaluation. It has become an easy and quick next day since we had already gone through single word translation evaluation and now it’s time to complete two words and three words translation evaluation, where majority of the single words were common. So, we just needed to pick-up the translations from earlier evaluated words. Moreover, Mr. I Felix, who is the member of the FUEL upstream project provided contextual information for each FUEL entry. That has made our job little easy. End of the day, we were done with evaluation of all 578 entries of FUEL.

During the event, we realized that there were some words which are not really commonly used, but difficult for the end users to understand. So, FUEL entries has to be revised as well in my opinion. Another thing we noticed that, some words in English itself are improper at their usage location, which needs to be addressed and resolved. Standardization towards English words is necessary.

At last we discussed the strategy to implement the evaluated translations into the existing work. Under the wonderful co-ordination from Ms. Maitri Shah has helped us to complete the event successfully on schedule.

FUEL evaluated translations for Gujarati is available from fuel-gujarati page: https://fedorahosted.org/fuel/wiki/fuel-gujarati

In simple words, it was a great event for Gujarati Localization community and evaluated translations going to play an important role for future localization activities. Hats of to all who participated in the event and spend their precious time and efforts to make Gujarati Localization better through FUEL evaluation workshop!

Some snaps of the event:
Infineon Infotech Pvt. Ltd
Mr. Prashant Dave
Mr. Siddharth Bhatt and Mr. Ketan Upadhyay
FUEL Gujarati participants
FUEL Gujarati participants

/Ankit

Posted in Gujarati, Localization | No Comments »

Turing on x86_64 bit machines

Monday, September 28th, 2009

You might have read about implementing Captcha on your website using a Ruby on Rails gem called Turing in my earlier blog post. Turing is really very good captcha system suited for my needs and may be many of yours.

Recently I have faced an issue with Turing on 64 bit machine (server, where I host my website). It did throw "500 an internal server error" when trying to access a webpage which shows captcha image using Turing, while the webpage used to work perfectly on i386 machine locally. I tried running the code manually from the console and it did gave "/usr/lib/ruby/gems/1.8/gems/gd2-1.1.1/lib/gd2/font.rb:234: [BUG] Segmentation fault" error indicating some issue with gd2 font.rb file.

After digging through the error message I came across two important urls that had solved the problem. The urls to refer are:

http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-talk/227097
https://boxpanel.blueboxgrp.com/public/the_vault/index.php/Turing_Gem

The problem isn’t exactly within Turing or gd2 gem but ruby needs a better code to work with these gems.

URLs above describe the problem, reason of the problem and solution as well. Tried to write them in a much simpler way, that could ease the process of getting the problem solved.

## 1. Download the latest ruby source code instead of ruby-1.8.6.tar.gz
wget -q ftp://ftp.ruby-lang.org/pub/ruby/1.8/ruby-1.8.6.tar.gz

## 2. Extract the downloaded tarball
tar xzf ruby-1.8.6.tar.gz

## 3. Change directory to extracted source
cd ruby-1.8.6

## 4. Change directory to place where the patch has to be applied
cd ext/dl

## 5. Make a backup of original file where the patch is going to be applied
cp sym.c sym.c.orig

## 6. Download the patch
wget http://www.ankit644.com/ror/sym.c.patch

## 7. Apply a patch
patch sym.c sym.c.patch

## 8. Change directory to parent directory
cd ../../

## 9. Configure with below command
./configure –prefix=/home/local –enable-pthread –with-readline-dir=/home/local

## 10. Make the binaries
make

## 11. Install the binaries to your system.
make install

When your website is on a shared hosting web server, your hosting provider will not allow you to install a custom patch for solving an individual needs. So, you have to install custom ruby in your local directories and make your custom settings. e.g.

  • Installing newly compiled ruby into your local directories (as per ‘make’ and ‘make install’)
  • $PATH environment variable should take your custom built ‘ruby’ command
  • "#!/usr/bin/ruby" (a shebang line) in public/{dispatch.cgi, dispatch.fcgi} has to be changed to "#!/home/local/bin/ruby"

And you are done!

Please do provide your feedback, if you think issue could be solved in a better way.

Thanks for reading!
Ankit

Posted in Ruby On Rails | No Comments »

Towards standardization…

Tuesday, August 11th, 2009

Till date, we have seen many efforts towards standardizations of translations, like Glossaries, Dictionaries, FUEL, etc. across Open Source projects and language communities. However, one question always come to my mind why translations are not yet standardized even though these many efforts are already in place!

Tried to figure out the possible reasons, like

  • not all languages have efforts being made in the direction of standardization
  • not all languages have glossaries, dictionaries, fuel, etc.
  • not all translators follow standard terms of translations being used in Glossaries, Dictionaries, FUEL, etc
  • not all translators use terminology management tools
  • not all translators spend time standardizing translations

it could be any of the above. But there is another important issue I just realized that we have different group of translators working for almost every different open source project for this matter.

Fedora: http://fedoraproject.org/wiki/L10N/Teams
Debian: http://lists.debian.org/i18n.html
OpenSuse: http://en.opensuse.org/OpenSUSE_Localization_Teams
Gnome: http://l10n.gnome.org/teams/
KDE: http://i18n.kde.org/teams-list.php
XFCE: http://i18n.xfce.org/wiki/language_maintainers
Mozilla: https://wiki.mozilla.org/L10n:Teams
Openoffice: http://wiki.services.openoffice.org/wiki/Languages
Robot Project (for various other important open source applications): http://translationproject.org/team/
Pidgin: http://developer.pidgin.im/l10n/

and many more…

So, one of the major problems I see towards standardization of translations is: there are more than one translation community per language, which leads to different standards being followed in various open source projects. As a result, there is no standard translation terminology across (and within) open source distributions. This issue is obvious for closed source projects as well, but could be resolved for Open Source projects at least by,

  • integration of language communities of each language
  • working together collaboratively
  • following same standard, even if language communities decide to run separate

Not sure, how is it going to be possible to implement any of the above, but there should be some way to fix this issue.

Thanks!
Ankit

Posted in Localization | 2 Comments »

Social life on the Internet!

Wednesday, July 8th, 2009

$subject of the topic might not be new to anyone reading this blog post. But it might be new to know that not all Technology Savvy people use the internet or surf websites as frequently as the end users do. i.e. The people who build the technology use it less than the consumers. One can find this statement dangerous as well as loosely coupled, since I don’t have any statistical data to prove it, but I think it is the fact. I have been in this industry for more than five years now, but earlier I have never been involved with the social networks, like orkut, twitter, LinkedIn, Flickr, Myspace, YouTube, etc or blogs even, over the internet as frequently as I am using now. I know there are faces around, who take the advantage of the internet and the facilities less than what actual consumers do with respect to social networks. Why is that?

What are the reasons, many tech savvy people use social networking websites less than the end users?

  • introvert or don’t want to open thyself to the world (can’t help with)
  • not good at writing (that doesn’t mean you should not do it. you will learn it eventually)
  • not enough content or ideas to write (start with something, you will get it automatically as I do)
  • not enough time (I don’t agree with it. If executives, celebrities, or the biggies can find time to do it, why can’t you?)
  • don’t like it (no force, but you will like once you get involved with it)
  • what to write? (read what others are writing, will help you to understand the trend)
  • what to surf? (once you are connected to the social networks, eventually you will start getting updates about the latest technologies or happenings in the world)

There will be various other reasons of not getting involved in the social networks over the internet techies will have. Post them here and get the answers.

Thanks!
Ankit

Posted in Miscellaneous | No Comments »

Growth of Indic Localization for Firefox

Saturday, July 4th, 2009

In late 2004, Firefox localization work started for two of the Indic languages (Gujarati and Punjabi) and respective language communities managed to get both of the languages included with Firefox 1.x release by 2005. Look at the status of the Indic languages for Firefox today… 11 Indic languages! It’s amazing! Three cheers for all localization contributors as well as good support from the Mozilla and Mozilla l10n team.

Firefox Version 3.5 3.0.x 2.x 1.x
Indic Languages Assamese

Bengali (India) Bengali (India)

Gujarati Gujarati Gujarati Gujarati

Hindi Hindi

Kannada Kannada

Malayalam

Marathi Marathi

Oriya

Punjabi Punjabi Punjabi Punjabi

Tamil

Telugu Telugu
Total Indic Languages 11 7 2 2
Total Languages 70+ 60+ 40+ 30+

Download the latest version of Firefox 3.5 and provide your feedback to the Mozilla and L10n development communities to make your web experience better and better with Firefox.

Firefox downloads: http://www.mozilla.com/en-US/firefox/all.html

Cheers,
Ankit

Posted in Firefox, Localization | 2 Comments »

« Older Entries