Re: ITS IG Teleconference on Tuesday Apr-28-2009 At 15:00 UTC/GMT

Yves Savourel wrote:
> Hi Camille,
>   
[...]
>> Concerning packages for localization that highlight modified parts, with context,
>> I must say I have no idea how that is made, can you recommend a simple source of
>> information about that?
>>     
>
> Simple diff algorithms should provide you with the sequences to highlight.
> A starting point could be http://en.wikipedia.org/wiki/String_searching_algorithm
>
> I don't know any library that does directly compare and highlight, but other may.
Thanks, and once you have the diff information, how do you specify in
the XLIFF what parts have changed, what is context, etc.?
But maybe I'm misunderstanding: what do you call exactly a "package for
localization"? Sorry for being so ignorant in this field...

I guess I must add some context myself about how we intend to manage this:

- Before the original file is translated, automatically add a unique
xml:id to every element that is translatable (according to ITS)
- Make a first translation
- Update the original
- Add xml:id where missing if necessary
- Increment a revision number associated to every xml:id of elements
that were modified

This way we can know exactly what elements need translation update. Does
this make sense to you? What would contain the translation package in
this case?

Best regards,

Camille.