[Project-ideas] [Anubad] Glossary Tool to index available strings

[Arani Bhattacharya]

Hi,

Sorry for my late reply. Before discussing things, I wanted to get
familiar with the existing tools.

On Tue, Mar 20, 2012 at 9:58 AM, Sankarshan Mukhopadhyay
<sankarshan.mukhopadhyay at gmail.com> wrote:
> Hi,
>
> We've recently formed a new list to discuss the project ideas. I've
> copied that on this email, I'd request you to subscribe to the list as
> well. The subscription interface details are at
> <https://plus.google.com/b/111332687752284580880/111332687752284580880/posts/3ziPwcKomXD>
>
> On Mon, Mar 19, 2012 at 10:46 PM, Arani Bhattacharya <arani89 at gmail.com> wrote:
>
>> I am Arani Bhattacharya, currently a first-year M.Tech (Computer
>> Science) student of Indian Statistical Institute. I would like to
>> participate in this year's Google Summer of Code program. I found the
>> idea of creating a glossary tool to ensure consistent translation to
>> be interesting. I would be grateful if someone could kindly clarify my
>> doubts on this matter:
>
> I'll ask Runa to add more in her response to your queries. I'll add a
> few initial details to get started.
>
>> 1) As pointed out, one of the major initial tasks is to study any
>> existing glossary, and analyze its working, advantages and
>> disadvantages. Could someone point me to some existing tool?
>
> Translators use tools like lokalize, virtaal or, online tools like
> Pootle, Zanata, Transifex. All these tools allow addition of a
> glossary file to the translation memory in order to aid in enhancement
> of translation. One aspect of the project idea is that while the
> translation tools do their best to ensure consistency during the
> "translation activity", this specific tool will be of aid as a
> reviewer tool as well as will continue to keep itself updated and
> corrected with the various upstream repositories of content.
>
This week I tried to get familiar with virtaal and Pootle. I also got
familiar with using the Bengali prabhat layout. What I also discovered
was that we already have a fairly large collection of Bengali words in
the central Pootle server that can be used.

>> 2) Which programming language should preferably be used to implement the tool?
>
> We do not have any specific preference for a programming language. A
> language/framework in which you are comfortable and have high
> confidence in achieving the objective of your proposal would suffice.
> The initial idea on the page is a problem statement. What we are
> looking forward to is a proposal which indicates that you have thought
> about the project, have assessed the path you'd want to take and are
> now in the phase of initiating a conversation.
>
>> Apart from the above technical issues, there is one other issue that I
>> would like to clarify:
>>
>> 1) Our institute has a 10-week window for pursuing internships. This
>> period ends on 24th July. This is about 25 days before the final
>> submission date given on Summer of Code's website. Is this likely to
>> create any problems?
>
> I can see two ways out of the situation - one where you follow the
> GSoC schedule and work with your university to figure a proper way to
> handle the mismatch of nearly a working month. The other is where you
> follow the schedule of the university but you end up exerting more
> effort because you will not be able to continue to work for around 25
> days (ie. you get 25 days less than everybody, if selected).
>
> I'd say that you are best placed to make a choice on that matter.
Before laying out a formal proposal, I will talk to my institute's
authorities on this and then keep you informed. While I will be able
to work even after classes open, I think it would be best to complete
as much as possible before that. If selected, I will work in weekends
to ensure that.
>
> /sankarshan
>
>
> --
> sankarshan mukhopadhyay
> <http://sankarshan.randomink.org/blog/>

Please correct me wherever I am wrong. I am now laying out my own idea
about the problem.

As of now, we already have an index of the most common words. However,
we do not have any tool that can automatically translate the common
commands and use standard words for them. This can result in
inconsistent translations. What we would like to have is a tool that
can go through the available English commands (including click-able
items of GUI), and suggest standard translation terms to the
translator. To do so, it should use its own index, which can initially
be borrowed from other sources such as Pootle's server index.

As I said, this is my understanding of the idea as of now. I will be
happy to modify it depending on your suggestions and feedback.

With regards,
Arani Bhattacharya