Jean-Christophe Helary ha scritto:
(Boris or anybody else, I Cc to OLT's dev list because that is
relevant here. How hard do you think it would be to have OLT's filters
work with OOo's l10n format: .sdf ?)
On 23 déc. 07, at 18:00, Yury Tarasievich wrote:
On Sun, 23 Dec 2007 05:59:05 +0200, Jean-Christophe Helary
<[EMAIL PROTECTED]> wrote:
XLIFF files directly generated from the XML for the help files and
from the rest for the UI.
Here I should point out that preciously little tools work currently
with the 1.1 XLIFF (even Sun's OLT didn't process 1.1 in Spring 2007).
Anyway, such change, while disruptive to many, would also be at least
useless, if with it won't come extended meta-information coverage
(string context etc.)
So, possibly the real answer is "extend coverage of meta-information,
extending the SDF format appropriately".
Yuri, I don't want to sound rude but you have it all wrong. I don't
care about the SDF format.
What I need and what translators need is a tool chain that does not
_break_ the data as it does now with oo2po.
What translators need is tools that support the format they use with
the least transformation possible.
Open Language Tools' problem is not that is does not support XLIFF
1.1, its problem is that its filter does not support SDF...
Besides, if OLT does not support XLIFF 1.1 why not feed it with XLIFF
1.0 ? Have you thought about that ?
All the source files are in plain XML, have you thought about the
efficiency loss of converting an XML file to a non XML format in terms
of checks ? Have you considered the efficiency gains of keeping a
strictly XML based tool chain for the localization ?
Pavel says that SDF is so easy to transform, then, what about OOo l10n
developers took some time to improve OLT's filters so that we can use
them directly with SDF ?
_That_ would considerably improve the translation process. To a point
that OLT could actually be used with the TMX by any community that
wants to work with a professional grade tool...
Any taker ?
I agree with JC: it would be a great thing if we could step away from
database --> sdf --> po --> whatever_format_translators_use and back
again to sdf with all the inevitable errors the conversion and
back-conversion brings about (any time I run gsicheck I'm impressed by
the number of errors I find!). Unfortunately, I'm in no position to put
forward some sound proposal on the matter but I would really like to see
OOo (and OSS in general) adopt a more linear way to handle the content
to be translated.
I'd like to add one more thing on the OLT tools: to make them usable,
the TM engine should be improved. As it is now it takes me from two to
over five second to switch from one segment to the other when I load a
big TM (such as the one provided by Sun for the OOo L10N) - and I'm
running Windows Xp or Linux on a dual core machine with 1GB RAM.
Ale.
----------------------
Jean-Christophe Helary
http://mac4translators.blogspot.com/
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
--
Alessandro Cattelan
Freelance translator (EN, ES --> IT)
http://www.proz.com/profile/76355
Tel.: (+39) 338 1823554
Skype: acattelan
Yahoo! IM: alessandro.cattelan
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]