Re: Modified strings up to release date

[Claude Paroz <claude 2xlibre net>]

Le lundi 22 septembre 2008 à 09:53 -0500, Shaun McCance a écrit :
> On Mon, 2008-09-22 at 12:45 +0200, Claude Paroz wrote:
> > Le lundi 22 septembre 2008 à 11:01 +0200, Andre Klapper a écrit :
> > > Am Montag, den 22.09.2008, 07:01 +0200 schrieb Jorge González González:
> > > > there have been modifications of document strings up to the very same
> > > > date of release, this is very sad. We already talked about "freezing"
> > > > somehow doc translations so we could do our job, but developers keep
> > > > updating it up to today.
> > > > 
> > > > I know there is no freeze for docs, but still, you, developers, cannot
> > > > think we can translate like this.
> > > 
> > > I think we have this discussion in release-team every time we meet.
> > > It boils down to the problem "Having better (updated) english
> > > documentation vs. having more translated (but outdated) documentation".
> > > Pick your poison.
> > > For me it looks less useful to have translated but outdated docs.
> > 
> > Sorry Andre, but I completely disagree here. The choice is not the one
> > you presented. I could say the same with other freezes. E.g. with the UI
> > freeze : do you want better and more polished UI rather than well tested
> > and documented but minimal and uglier ones because of the freeze...
> 
> As I've pointed out before, this analogy doesn't hold water.
> It's not a matter of making the documentation more polished.
> It's a matter of making it correct.
> 
> If a program needs a string addition to give a user information
> in the case of some error, that's polish.  If the documentation
> is telling you to click on button XYZ, but button XYZ does not
> exist in the program, that's just flat out wrong.  There is no
> point in translating that sentence, because there are exactly
> zero users who would be helped by reading it.

I don't contest the need to correct this, but the timeframe to do it.
Moreover, freezes have exception processes, so obvious and critical
errors could still be committed.

> > It's a matter of process. When you put a freeze in place, you're simply
> > telling people that they have to do their job in a specific timeframe.
> > The objective here is to have updated AND translated docs.
> 
> Stormy had a great blog post a few days back, talking about
> how project releases are constrained by time, resources, and
> scope.  We currently have a fixed amount of time, and have a
> very difficult time increasing our resources (i.e. writers).
> Our scope (i.e. documentation quality) suffers.
> 
> If you decrease our time without increasing our resources,
> documentation quality will suffer.

Please, please... we're talking about a some (3?) days freeze in a
schedule of six months.

> Please don't claim there is no respect for translators, as
> you did in a previous email.  That's pure flamebait.

When you pass tenth of hours to translate a big document and you see
half of it unvalidated by an update some hours before the release, try
to imagine the feeling of the translator...

>   I put
> a lot of work (along with Danilo) into making translators
> able to do documentation with po files.  I wrote an entire
> DocBook toolchain in part because the existing solutions
> didn't serve our translators well.
> 
> I have put a lot of development time into making sure you
> can have properly translated documentation.  But I will
> not take precious time away from our few valiant writers,
> just so you can have a translated version of a document
> that's not even correct or helpful in English.

And you know that I'm also one of these few (albeit a minor one).

We both defend our respective position (doc writer/translator). IMHO
both are somewhat valid but unfortunately they conflict... and I don't
see this thread going to change anything right now.

Claude