From: Peter Eisentraut <peter_e@gmx.net>
To: "Dave Page" <dpage@vale-housing.co.uk>,
	"Alexey Borzov" <borz_off@cs.msu.su>
Subject: Re: Requirements for updated site
Date: Thu, 15 Jan 2004 18:29:59 +0100
User-Agent: KMail/1.5.1
Cc: <pgsql-www@postgresql.org>
References: <03AF4E498C591348A42FC93DEA9661B8720474@mail.vale-housing.co.uk>
In-Reply-To: <03AF4E498C591348A42FC93DEA9661B8720474@mail.vale-housing.co.uk>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200401151829.59799.peter_e@gmx.net>

Dave Page wrote:
> > I presume that 99.9% of these is documentation. And it won't
> > be translated, at least not in *this* way.
>
> Why not? The whole point is to have *one* system for the whole site
> and not a mish-mash of different ways of building bits of the site.

Well, that's like saying we should edit all PostgreSQL source code in 
assembly, because there is some assembly in there and for reasons of 
consistency that should be the common denominator. ;-)

If, say, it were accomplished to extract gettext-style message catalogs 
from both XHTML and DocBook, then translators could use the same set of 
tools, and writers could use the same XML editors for everything.  Just 
the internal semantics are going to be different, but they will anyway, 
because you're writing two different things.  It would still be a 
pretty good "one system" to me.

I keep saying "gettext", because that is the direction that the 
documentation is most likely going to go in, because it's successfully 
practiced at KDE and all the tools are available.  Nothing is decided 
yet, but some translators have expressed strong desires to go that way.  
But I'm glad we're here to discuss it so we don't invent the wheel 
twice more.