[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#913723: please confirm that dictionaries should move to section localization



[ Sorry for the dupe! Resending to the correct bug number, the debbugs
  mboxes do not track cloned bugs. :/ ]

Hi!

On Wed, 2018-11-14 at 11:12:04 +0100, Sébastien Villemot wrote:
> Le mardi 13 novembre 2018 à 17:33 +0100, Jonas Smedegaard a écrit :
> > Quoting Sébastien Villemot (2018-11-13 16:34:05)
> > > Le mardi 13 novembre 2018 à 16:12 +0100, Jonas Smedegaard a écrit :
> > > > I consider lintian an _aid_ but not authoritative.
> > > 
> > > I tend to consider lintian authoritative when there is no “higher 
> > > source of law” (such as the Policy, DevRef, team policies…).
> > 
> > Ok, so we agree that Debian Policy is more authoritative than lintian.
> > 
> > Debian Policy states § 2.4 says this about section definitions:
> > 
> > > For more information about the sections and their definitions, see the 
> > > list of sections in unstable.

Actually the ones responsible for the archive sections, their definition
and contents are supposed to be the ftp-masters. My impression though is
that they have stopped caring about that for some time now. :/

And given the above, I've been trying sporadically to cleanup and
correct the overrides after the fact (after NEW processing which is
when the correct overrides used to be determined). I track that with:

  <https://namespace.hadrons.org/arsenal/overrides-sublimator>
  <https://namespace.hadrons.org/arsenal/overrides>

should perhaps make that somewhat more visible, though. And in this
case I was meaning to send a mass override fixup after lintian got
the patch accepted, but it seems I forgot.

> > > The web page https://packages.debian.org/unstable/ says this about 
> > section "localization":
> > 
> > > Localization support for big software packages.
> > 
> > ...and this about section "text":
> > 
> > > Utilities to format and print text documents.

> > It seems sensible to me that hunspell dictionaries are treated as 
> > utilites to process text documents rather than localization support.
> > 
> > Localization is specific to mapping internationalization strings into a 
> > local context - purpos of a dictionary is far more general than that - 
> > and it seems all other maintainers of hunspell dictionaries besides you 
> > came to that same conclusion.

Actually I disagree with this characterization. Localization [L] is
not just translation, it's the entire task of adapting a project into
specific locales, be that with translations, image/media adaptation,
specific customs of the locale, such as telephone numbers, paper size,
or numeric notations, etc. In this case I see the spell checkers are
"text", because they *are* utilities (part of i18n) and the dictionaries
as "locatization" because they are definitely *not* utilities (they are
passive data :), and are relevant as part of the locale.

  [L] <https://en.wikipedia.org/wiki/Internationalization_and_localization>

> On the other hand, the main usage of dictionaries is to correct
> spelling within the context of _localized_ text editors. So it's not
> absurd to consider that their main purpose is localization, though I
> agree that they may be used in other contexts.

I think any other context is still locale-specific, even if it's not
to spell check a text. :)

> FYI, bug #874121 contains the patch that implemented this lintian tag,
> but unfortunately it contains no justification for it.

Sorry, should have added one, I guess I thought it was obviously
correct at the time. :D

Thanks,
Guillem


Reply to: