Enjoy.
Cheers.
On 1 June 2011 12:41, Anivar Aravind <anivar.aravind(a)gmail.com> wrote:
On Wed, Jun 1, 2011 at 11:58 AM, sankarshan
<foss.mailinglists(a)gmail.com>
wrote:
On Sat, May 28, 2011 at 8:23 AM, Anivar Aravind
<anivar.aravind(a)gmail.com> wrote:
There is major problem in depending upon Non-FOSS
solutions.
There is. However, the line that comes immediately next might just
need to be couched as a personal opinion in the absence of cited facts
around it :)
> It seems google used api's for collecting crowd sourced input for
> improving their database & now shutting it down.
We have seen its evolution & improvement of service with wide adoption
over time.
One curious factor i used to notice always is the formation of words
with chillu characters in transliteration API .
Since there is a standard difference between Unicode 5.0 and 5.1
Chillu Characters it is easy to identify. Google is officially
following most recent unicode version, but since most of the malayalam
users uses chillu letters in unicode 5.0 specifications, it is
periodically changed in that way .
> Input methods such as Transliteration does
not need online services . It
can be easily done
without
google or microsoft. And popularise extensions such as
Narayam existing on mediawiki is the best way to move forward. In
addition, it will be good if Indic wikipedian's can develop a
stylebook for correct usage and link it with extensions such as
narayam. Then the possibility of typing errors can also be reduced.
I am more interested in the line in Sudhanwa's mail which seems to
suggest that WMF should be providing APIs (and, hosting a service
therefore ?) around this aspect.
Yes, it is an interesting possibility
And, I am afraid I also do not comprehend the
usage of the term
'stylebook' - what does it mean in this context ?
In google transliterate , it took a lot of time to correct spellings
in word formations
Since wikipedia have a large DB , and editors , who are aware about
correct usage , a style book of correct indic language spellings for
transliterated inputs can help input tools like narayam to auto
correct common spelling errors . It was a major feature of google
transliterate . This can be also implemented in API level as well
In my
opinion Translation Service& its api of google for indic is not
that great anyway , and usable for wikimedia projects . So i believe
it will not make much issues in wikimedia indic projects
The usage of the Google Translate API has been, for a while, discussed
under the umbrella of whether the translated content can be 'freely'
(as in libre-ly) used across domains.
_______________________________________________
Wikimediaindia-l mailing list
Wikimediaindia-l(a)lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikimediaindia-l
--
Regards,
ME.
Wear a Lungi, Support the Movement
My infrastructure invasion... plus other images
too.. on Wikimedia Commons.