Re: Sun Translation tool (Was: Re: Heartsome XLIFF Translation Editor for document translation?)
- From: Tim Foster <Tim Foster Sun COM>
- To: Simos Xenitellis <simos74 gmx net>
- Cc: GNOME I18n <gnome-i18n gnome org>
- Subject: Re: Sun Translation tool (Was: Re: Heartsome XLIFF Translation Editor for document translation?)
- Date: Tue, 30 Nov 2004 18:38:11 +0000
Hi Simos & all,
On Tue, 2004-11-30 at 13:46, Simos Xenitellis wrote
> > Tim wrote :
> > order to get our translation tools (XLIFF editor and filters for a bunch
> > of file formats) open sourced
> Will this tool support a translation memory
Hmm, define "support" ? The translation editor includes a mini-TM system : probably doesn't scale to the millions & millions of segments that you'd expect a typical TM system to work to. For projects of perhaps tens of thousands of segments, it should be fine, but it's not industrial strength.
Now, that said, we communicate via XLIFF and TMX files, so there's no reason why you can't take input and send output from/to a larger TM system. We have developed such a system in-house, and we're also considering open sourcing this, but from a standpoint of not wanting to have people drinking from fire hoses (does that sentence make sense across cultures ?) we don't want to just offload *all* our code in one massive lump at the same time and have people struggle with it, when 90% of users would be interested in the editor - getting people focused on the editor and the idea of writing XLIFF filters I think would be a better aim of the project to begin with.
Initially, we're thinking of just releasing the filters and XLIFF editor, and then when people have digested them (burp!) we could release the server-side components. It's likely that if we do that, the underlying search engine that powers the TM system would only be released in binary format (since it was developed at SunLabs, I don't know what their position is wrt. releasing the code)
As I say, I think the releasing the industrial strength TM system is going to be a bit down the line. ( It's comparatively hard to deploy, we've only run it on SunONE Application Server + Oracle combinations -- before we release it, I'd like to be able to have a running version on (say) JBoss + MySQL, for example, and generally make it as easy as possible to start to use/develop on)
In comparison, the editor runs on any client machine that has a JVM, has a one-click installer and the XLIFF filters run on the command line, so all in all, they're much easier to start using than the TM system.
> As open-source there will be a lot of things one can experiment on, like
> Web-based translation, adapting existing tools (KBabel, poEdit,
> gtranslator) to the translation memory and so on...
Neato - certainly you could take the mini-TM component of the editor, and bundle it into other translation editors/servers if you felt like it (or perhaps port the algorithms we're using at the very least - otherwise you'll be getting into JNI calls to/from our java-based mini-TM, which may be more complex)
> Once this tool is released, there will be little to say about the
> efforts of Sun on open-source software. :)
Hey, from *my* point of view, there's little to say right now, but I'm biased obviously ;-)
http://blogs.sun.com/roller/page/eschrock/20041129#jonathan_on_opensolaris
cheers,
tim
--
Tim Foster - Tools Engineer, Software Globalisation
http://sunweb.ireland/~timf http://blogs.sun.com/timf
http://www.netsoc.ucd.ie/~timf
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]