Re: [Freecats-Dev] Re: Implementation

[Stanislav Visnovsky]

Dňa Štvrtok 06 November 2003 09:53 Henri Chorand napísal:
> Hi Stanislav,
>
> Nice to read you. I'm answering on the list, so that Dan and other (active)
> members also see this.
>
> > Is there still no code for FreeCATs? The question is, should I
> > start from scratch, or is there anything available (I mean for
> > FreeCATs project, not the other projects floating around).
>
> No code yet, I'm afraid, but Dan Urist already has something he would like
> to bring in.
>
> I have not heard from Tim Morley about his first experimentations with
> MySQL. Whatever the result, we would be interested to know.
>
> Dan is presently working on packaging his prototype software (see his
> latest posts on the list). Note that (especially when reading his setup
> warnings and since my current PC is under Windoze) I did not try to
> download it and look at the code yet, as I'm not familiar enough with the
> languages he used.

I know, I've check his older code. IMHO it will not help that much, but maybe 
the new one will be more interesting. 

>
> >From the point of view of Free CATS's goal, some aspects of it make it
> > look
>
> most valuable at first sight (especially the interface plug-in for Open
> Office).
>
> Dan used Perl and Java languages, not necessarily in line with my initial
> suggestions, but I mean, why not...
>
> In order to successfully integrate all or part of his code into Free CATS'
> framework, I think it will be important to agree on some common ground,
> part of it being, using the same words in order to designate the same
> items. This is why I think it's important to read our spec documents,
> including because they give a good idea of what CAT is to developers. It
> may look like a mere detail, but I think it's not.
>
> > I'd like to implement at least a prototype for the server API.
>
> Well, for a start, you could possibly work in team with Dan so as to review
> my API proposal and Dan's WSDL files so as to highlight discrepancies in
> terms (I saw some) and features (mostly minor ones, I guess), so as to call
> for a short set of votes and see them converge.This is a type of job where
> an "external eye" is quite valuable. Also, Dan's WSDL files is not exactly
> an API, as it's an implementation of an API. Again, I believe his work to
> be a quite valuable basis, but I think it's important for the future to be
> able to stabilize a first "core", even if it later evolves a bit.
>
> My API proposal is obviously not the work of a hard-core hacker, but it
> tries to synthetize the core features needed by any CAT software from the
> point of view of experienced translators. Because of this, I believe it
> should be globally taken into account (possibly extended and improved, of
> course).
>
> This is important, because other bits of code are likely to come up, and we
> ensure all of them comply to the same consistent set: one (compatible with
> itself) project. Otherwise, it could (and will) become a source of hassle.
> Once the API has been finalized, and every layer well-fefined, it will be
> easy to plug other modules. I hope I express myself right here.

OK, I will try to stick with those, but use the yours as a primary source.
IMHO, we should stop talking about specifications and start to write the code.

Stanislav