Friday, November 13, 2009

brisbane visit (first post)

better to right this now since i think there are many more going to come

what's a lengthy travelling from Mumbai to Brisbane,
it took almost 31 hour for me from Mumbai to here Brisbane
i leave Mumbai on 6th Friday 1pm and reach here in Brisbane on 8 morning 10am (IST 5:30am)

though hospitality was good in emirates

now i really feel we should have taken route Mumbai-Singapore-Brisbane instead of Mumbai-Dubai-Brisbane.
but still some +ve point were
Dubai airport was excellent, its so huge and so many things around
lucky we got a good time to rest in Spring Hill Apt. on Sunday.

umm bit ok with food, but not very good though
i guess will need try some indian food restuarant...

Note: posting bit late, i came to Brisbane 8th Nov

Tuesday, August 04, 2009

FUEL @ CDAC GIST on 31st Jul and 1st Aug

What a nice two days on discussing FUEL (frequently used entries for localization) for Marathi, problem is "though Marathi is my mother tongue i don't use all the Marathi words regularly"
in fact i found that almost, may be more that 50% words of Marathi i never use in my day to day conversation.

It revised lots of my Marathi words.

It was event i attended ever on Marathi, i found Marathi language is so rich in words.

I think translating words from English to Marathi looks like very much one-to-many mapping like thing


in English we use one word at many location and meaning keeps on changing but

in Marathi even for same thing we have different word according situation

smell can be good or bad

in Marathi it can be
good smell = सुवास/सुगंध
bad smell = वास / दुर्गंध

at so many location we found this diversity and then finally decided to do by actual context of word :)

the one thing "Clean Up by Name" is nothing but sort according name so we used corresponding word for Marathi

we didn't took a big risk of translating into word which 90% of Marathi people will not understand and may be they will need to see in dictionary what this Marathi word exactly mean.

so we preferred to do transliteration instead of translation since these English word are so widely used by people that keeping them as it is will help people lot.

example: setting
even people in 1st class also say sometime setting ;)

We had a really difficult time in translating some words like:
Buddy Pounces, scenarios, detective, control, autotext and anchor.

Though there is corresponding word available in dictionary, but it should satisfy context as well, that was the main problem.

But finally we completed whole FUEL list on second Day.

thanks to GIST ,CDAC and specially Mr. Mahesh Kulkarni for coordination and off course Sandip Shedmake for his patience in listening all attendees suggestion.

some snaps:

Thursday, May 28, 2009

Still gcalctool not allowing indic numerals

presently gcalctool not supporting indic numerals, but i think it will nice if we enable it for indic digit also, so if someone entered indic numerals for calculating it should do a calculations and display a result in indic numerals.
I googled for this but didn't found any topic related this.
now i dont know complexity involve in doing this, but as far as i think it will just require conversion from Indic numerals to Latin numerals and vice versa. When somebody enters any expression and if it contains indic numerals it will convert it to latin, then processing as usual and then while display result it will do same thing back.

1) understading gcalctool code
2) utf-8 to unicode conversion
3) validation for indic numerals
4) converting to latin before processing
5) converting result back to indic after processing
6) unciode to utf-8

thats it

hope so there is no any standard not allowing doing this thing :), i will talk with upstream for this lets see

Wednesday, May 20, 2009

facing problem while using Policykit for user authentication

I am trying to use policykit in system-config-language for authentication, so thought lets try with some test application

while doing so i am following steps given in murrayc

everything is going well
def write_file():
fp = open('/etc/fonts/fonts.conf', 'a')
fp.write("i have wrote")

if __name__ == "__main__":
print "calling wrtie file"

#Call the D-Bus method to request PolicyKit authorization:
session_bus = dbus.SessionBus()
policykit = session_bus.get_object('org.freedesktop.PolicyKit.AuthenticationAgent', '/')
if(policykit == None):
print("Error: Could not get PolicyKit D-Bus Interface\n")
granted = policykit.ObtainAuthorization("", (dbus.UInt32)(0), (dbus.UInt32)(os.getpid()))

print granted

if granted:
print "no access"

in my code after calling

granted = policykit.ObtainAuthorization("", (dbus.UInt32)(0), (dbus.UInt32)(os.getpid()))

it is asking me for root password as per i stated in .policy file, but after entering root password, in next line when i am trying to update /etc/fonts/conf.d, it is not allowing me to do that :(

i dont know even after entering root password from policykit why my code is not able to edit above file

my all file are here

still trying for this, thought lets share it

Thursday, February 12, 2009

Malayalam Collation is now in glibc upstream

yesterday Ulrich upstreamed ml_IN collation patch

2008 nice year of Indic computing as we started targeting collation for Indic Languages in glibc,
starting point was Ulrich Drepper's India visit for 2008, he pointed out in his speech still there lots of work to do for Indic Languages in glibc
for collation and iconv
I started some work initially, studied already implemented things such that collation tables for ta_IN and as_IN, community event Indic Mashup in Pune helped lot, in that event almost all Languages LM were present and got a good input to start work

started collation with Devanagari mostly used script in India, then one bye one,

we have got a good community, they really worked very fast for this, for Sinhala and Malayalam harshula and santhosh completely wrote collation table i just reviewed and tested them.

so its really nice thing happen for Indic languages

we can see all menus for particular languages in sorted order,
also we can easily sort any long list using sort function provided by glibc and quickly search required item,
also like the way we search our required file in huge docs folder pressing initials of name, same thing now we can do for our language(i was missing that lot)

Fedora 10 already has support for most of the collation, now Fedora 11 will have support for Malayalam Collation also

collation table for above languages/script are available at glibc upstream

the only one remaining now is Bengali collation, Bug
hope so things for that also get sorted quickly and we can say now Linux support sorting for all Indic Languages :)

thanks to all contributed in this and special thanks to Ulrich for all support :)

Tuesday, January 13, 2009

now easily install ttf/otf fonts in fedora 10 using kfontview (kfontinst)

today i was trying to read news from, but it is using custom encoding font so i downloaded it..

I was planning to install it using regular mkfontdir, fc-cache command but was feeling bit lazy to do that..
just thinking can i just install it using just one click and yes now we can do that.

kfontview is now have that feature.

Just right click on any .ttf font file and open with kfontview

It will show you characters in fonts and also there is install buttton (It will be disables in font already installed ) just click on that, now restart your application in which you want to see same font and its d one

I am still remembering those days when windows user were calling me only for installing font for testing it in there Linux machine, It is very good for them now they can easily install and test font without any help :)

I think gnome-font-viewer is now disabled.

Friday, January 09, 2009

photofunia good one!!

today saw photos of my friends in some nice frames, tried photofunia
and its really funny

It quickly add you snap in nice frame, it uses face detection technology for doing this
some more of mine available at

for trying it visit :)