[Freecats-Dev] Project-related communication, development tools, etc.

[Henri Chorand]

B'jour tout l'monde,

>> If the majority of team members prefers another name, why not.
>> The actual name is "présentable", rather self-explanatory (for
>> a translator) and furthermore, I like cats :-)

> Well I just wanted to stir some thought. I reckon language
> professionals should do better than everyone else when it comes
> to using language.

Sure. Well, calling for creativity, who wants to work on a logo?

>> Marilyn Courté is presently writing up (in English) a document
>> intended for developers who don't know anything about CAT, to
>> help them download Trados software's trial version and translate
>> a small sample file while understanding how it all happens, with
>> a few screen captures.

> Hmm, I'd be inclined to try and balance that with a second
> application demonstrating an alternative paridigm to the "do
> everything in Word" system used by Trados and Wordfast. The SDLX
> or Trans Suite systems for example, make a total abstraction of
> the text to translate so you can leave your word processor at home.

Good point. The majority of ideas expressed up to now are as follows:

- We don't want to work from within MS Word. It's not an ethical choice, 
nor does it allow portability (we do want a Windows + Mac + Linux GUI 
thing), and seeing things from Wordfast's mailing list, it's also a 
nightmare from a coding perspective (VBA is slow, buggy and 
under-documented, I tried it a few times). Also note that Open Office 
still does not have a macro language, a Sun, Java-based project might be 
started to address this, but we can't wait that long.

- The initial version of the translation client will feature a 
standalone text editing area (an enhanced multi-line text control). 
Furthermore, our bilingual working format is going to be tagged text, 
and we plan to pre-segment the source document when converting it to BWF 
(while providing some flexibility for re-segmenting by allowing the 
translator to interactively:
    merge the currently opened translation unit with the next one
    split the currently opened translation unit.

I reckon it's a long way from Wysiwyg, but I feel the integration within 
a text processor has its own limits (especially when trying to increase 
the number of native document formats processed).

As we plan to use HTML layout tags, at some stage, we should be able to 
implement some form of document preview through a Web browser's display 
engine. While we would like to protect tags, a bit like in Trados Tag 
Editor, we might implement color coding (like in code editors) at first.

So, yes, there is life outside of Trados, and to give a comparison, we 
prefer the Tag Editor approach (stand-alone translation tool) to their 
Word-integrated Workbench.

The first version of the specifications document which everybody should 
have received (and read) contains a few ideas along this line.

Some of the ideas above were developed during our first project team 
meeting in Quimper, so this e-mail might also be the only "official" 
meeting report I'll produce.

> I believe they also offer free trials -- they certainly used to.
> This would help stop developers picking up preconceived ideas
> about the ONE TRUE WAY. (Also of course, there are a fair few
> developers out there who do *not* have Word).

Well, as there is no Linux tool available today, we guessed the best was 
to pick one of the most comprehensive proprietary tools.

Note that Julien Poireau suggested to build a working document which 
would include screen captures from all major CAT tools, for comparison's 
sake. This would also help making sure we try to provide a "best of 
worlds" approach as far as the interface is concerned. Feel free to 
contact him (jpoireau AT club-internet DOT fr) so as to provide him 
feed-back about Wordfast.

>> The idea of such a document comes from Simos, the most experienced
>> developer among all of us, who believes it can greatly help raise
>> interest about the project among free software developers.

> Talking about developers, where does Tcl/Tk come into this? Is this
> what the existing developers feel most comfortable with? Is it a
> question of writing everything in Tcl or just making a Tcl front
> end for something written in C for example?

I began looking for C/C++ tools with a GUI IDE that can write portable 
software. I eliminated Kdeveloper simply because its installation was 
not bug-free on my recent Red Hat and because I was not sure whether 
it's KDE-only or also Gnome-compatible (a major potential source of 
conflict in the free software world).
I then eliminated Qt, even though it looks a very well-finished product, 
because its XML functions library is only included in the (non-free) 
enterprise version, when David Welton, from Apache Tcl project, 
suggested me to look at Tcl.
I contacted David when I discovered an article he wrote nearly one year 
ago in www.advogato.com about using translation memories for translating 
free software, and he kindly provided me with feedback. See for yourself 
at http://www.advogato.org/article/450.html
See also a follow-up at http://www.advogato.com/article/454.html

I also contacted the authors of the two replies to ask if they were 
interested, without any result yet.

And, while looking for the above URL (too bad, no search engine), I also 
found the following about scripting languages:
http://www.advogato.com/article/550.html

Coming back to Tcl/Tk, I intend to prototype the interface with Tk. 
Please refer to Dev list archive to read the message I posted about 
Tcl's advantages for this project (strong string handling functions, 
access to Xindice & Apache available, easy integration of C (or whatever 
language) modules once we begin work on performance bottlenecks.

Of course, this is not to say we can't use other languages & tools.

> I've had a fair bit of experience with Tcl and though it has a
> solid body of followers and is very practical to use, it does
> require a certain "retro" mindset. I'm basically saying that maybe
> it's not sexy enough to attract developers. Have any alternatives
> been considered? Python + QT for example.

This really good news. Is there a chance we may call you soon a team member?

Again, I'm just trying to stir thought, which of course is my chief 
virtue :)

Yes, we would definitely need you.

>Would you be available to proofread it next week?

Sure if it's not _too_ long :-)

1903 source words. I only want to make sure it's readable English and 
that I properly convey my ideas [read: any ideas worth].

>> Kirk McElhearn already kindly accepted to proofread another
>> document which requests an article in Brave GNU world, which I
>> feel is also an important communication-related task. As all of
>> you know, we need a few extra (experienced) coders to speed
>> things up.

> Then why not go straight to source forge where there are always
> volunteers looking to pick up on a project?

Yes, it should be done.
A few people I initially contacted when launching this project warned me 
against Source Forge's possible drifts from the true spirit of free 
software (no flames here please). They also suggested Savannah, even 
though they said Savannah moderators could be quite selective when 
accepting projects. As Savannah immediately accepted Free CATS...

This is not to say we might not try to use Source Forge to recruit team 
members; I reckon SF hosts a VERY large number of projects today. 
Please, anybody, tell us if you happen to know how to use SF for this 
purpose. Launching a fake fork?  ;-)

>> Any ideas are welcome - apart maybe from shouting like Steve
>> Ballmer in a rather well-known conference, some time ago:
>> "Developers, Developers" ;-)

> Sorry I don't weigh 150kg and don't have sweaty armpits :^)

ROTFL (author's note: Rolling On The Floor Laughing, or "mort de rire")

Thinking about it, who would volunteer to see if we can have some media 
coverage via Freshmeat?

Last encouraging news: our development list has 8 readers. I hope we can 
reach (at the very least) 80 after the press articles, and 800 to, er, 
much more (seriously) when we launch our first prototype.

By the way, Kirk, we spoke about a short press release intended for 
translators, which we can ask to be posted to translators mailing lists, 
like TCR. if you still have time after proofreading the Brave Gnu World 
thing...


I'll stop for now. Good night, everybody!

Henri (a bit tired but very happy)