ubuntu-manual team mailing list archive

[Ask Hjorth Larsen <asklarsen@xxxxxxxxx>]

On Thu, Aug 5, 2010 at 9:56 PM, Phil Bull <philbull@xxxxxxxxx> wrote:
> Hi Kevin,
>
> On Wed, 2010-08-04 at 18:17 -0500, Kevin Godby wrote:
>> On Wed, Aug 4, 2010 at 5:56 PM, Phil Bull <philbull@xxxxxxxxx> wrote:
>> > Sorry to change tack slightly, but I'm wondering whether the format
>> > discussion should be deferred until some more important issues are
>> > resolved. If we choose an XML format, we can always transform to other
>> > formats with a bit of work.
>>
>> That's true to a degree.  But if one format has more granular tags
>> than another, it's hard to convert from the more general format to the
>> more granular format.  (It's easy to swap one tag for another or to
>> remove unnecessary tags, but it's hard to add them.)
>
> That's certainly true in theory, but I don't believe that it's a serious
> problem for us in practise. In my experience, few tags are actually
> required to give good semantic coverage. Additions to the
> frequently-used "core" set of tags tend to cover rare edge cases, where
> existing tags could be slightly abused with no ill effect. Often, the
> extra tags are displayed in the same way as an existing tag anyway.
>
> I don't think that there's a significant risk of choosing something
> that's semantically too poor. We could even use HTML with different
> classes for spans and divs if we wanted to!
>
>> (Did I stray wildly off-topic and completely forget to answer your question?)
>
> Eek, yes, although your discussion of translations was definitely
> valuable. I still think the editing issue needs to be addressed with
> some urgency, though. I'm worried that, after editing has finished, few
> messages will have been left intact, and the translators will have to
> redo pretty much everything.
>
> Thanks,
>
> Phil

The problem here is that Launchpad (rather scandalously unfortunately)
does not support the concept of 'fuzzy' strings in gettext.  As most
of you probably know, if a string is changed only a little, most
gettext-related translation software will mark the existing
translation as fuzzy, meaning the translators will know to review and
adjust the translation.

Here is a possible, although not ideal course of action:

  Make whatever string changes are necessary
  Export the translation po-files
  Use msgmerge to get all the fuzzy translations back into the file
  A reviewer will then have to go through all those changes, then
upload to Launchpad.  With a bit of scripting and fancy coloured
diffs, this could be reasonably efficient.

The drawback is that a lot of different people will have to perform
this possibly command-line-heavy procedure, so it requires a lot of
communication and documentation.  But I think this is much better than
having translators redo a lot of work.

Regards
Ask Hjorth Larsen