[Freecats-Dev] Re: Can I join? - YES :-))

[Henri Chorand]

Hi Michael,

> Just ran across your project in the context of open
> source for translators.  I'm a translator, yes -- but
> I'm a programmer first, and an open-source programmer
> at that (see http://www.vivtek.com/wftk.html).
> So naturally I've been interested in the whole notion
> of open-source translation software for some time.

Great. Your help proposal comes at a very timely moment indeed - we are 
to begin the actual coding within a few days.

> Anyway, count me in.  I just now signed up for this
> Savannah thingy; my userid is vivtek.

You are now registered in the project team with admin rights.

I suggest you browse the mailing list archive in order to quickly see 
what we went through - mostly design issues, actually.

None of the original team member is a developer. Tim Morley, a free 
software developer who, like most of the original group, lives in 
Brittany, recently met us a few days ago and kindly offered to help. I 
have yet to sign him in the project team, but he is already subscribed 
to Free CATS mailing list - I'm CC-ing him so that you know his email.

Our present aim is to develop a HTTP-accessible translation memory 
server. I'm sending you our two design documents in a separate message. 
The most urgent tasks planned are:
- to validate the API's list of functions and parameters so that any 
errors are corrected
- to define our "database" format (we have yet to decide what we want to 
start from: a (free) native XML server, some search-engine's full-text 
indexing technology or a basic, filesystem level approach)
- to code the API's functions (using a modular approach that will enable 
us to refine things over time, like advanced indexing or non-Western 
languages)
- to code the HTTP communication layer (we are thinking about an Apache 
plug-in, a more basic HTTP server or a framework like Twisted or Cherry).

As we want reasonable performance and good portability, we selected 
Python to develop our software. The other reasons are that Python is 
"good" at strings and XML management, good for prototyping and easier to 
master than lots of other languages, but also that it's supposed to be 
great at integrating bits of code in various other languages - so if and 
when we meet performance bottlenecks, it should be easy to rewrite the 
critical bits in C/C++ (or any other, faster portable language).

That said, if experienced developers joining the project have good 
reasons to prefer other languages, why not - as long as we don't loose 
the language's requested assets. For instance, if a C++ expert proposes 
to build up all project modules in a snap, we won't object ;-)

We thought about Java for some time, but more or less rejected it 
because of its stiff learning curve (like me, a few of us translators 
want to have a chance to write bits of code once the general 
architecture is set up), its performance and also its possible 
license-related issues.

Once you have read the specification documents, I guess the best you can 
do is to tell us more in detail about your coding skills so that you and 
Tim decide how to best coordinate your efforts.


I tried to sum up things the best I could. Thanks again for your help 
proposal!


Cheers,

Henri