W3C home > Mailing lists > Public > www-international@w3.org > January to March 1997

Re: Language label

From: M.T. Carrasco Benitez <carrasco@innet.lu>
Date: Fri, 28 Feb 1997 15:57:03 +0100 (MET)
To: Gary Adams - Sun Microsystems Labs BOS <gra@zeppo.East.Sun.COM>
cc: WInter <www-international@w3.org>
Message-ID: <Pine.LNX.3.95.970228153703.3213C-100000@localhost>
> I'm not sure I agree that the server should always be configured to
> parse a document when it is transmitted. On the other hand it would be 
> easy for a site administrator to preprocess the documents so a "foo.html"
> becomes a "foo.html.asis" in the case of the Apache server, this document
> would then have a preconfigured set of HTTP headers.

Docs should a internal language label that could be preprocess by the doc
management system.  But "virtually" the server is reading the internal
language level.

> Since very few clients have been updated for RFC2070 HTML constructs,
> I imagine even fewer servers are ready to support "<HTML LANG=xx>".
> On the other hand most servers do support the "<META HTTP-EQUIV...>"
> workaround.

This was more or less my line of reasoning, but I am happy with (almost)
any syntaxt.

> So if you want to deploy documents today that are portable across
> most current HTTP servers the META workaround is probably the most
> widely deployed solution. If you want exact control over the 
> HTTP headers, then you might be better off with a specific vendor's
> HTTP solution, like the Apache "asis" mechanism.

 - 1) Reach a consensus on how to put the language label in the docs.
 - 2) Dissemination of the agreed language label technique.

It will take a while (1  or 2 year) to disseminate, but at least
eventually a significant number of docs (probably non English) will be
properly label.

I would like to avoid vedor's specific solutions.

> In the long run if you want to be RFC2070 (should I say HTML 4.x)
> compliant and you want to be able to transmit the documents over
> something other than HTTP, then you want to use the HTML LANG 
> attribute.

So come back to <HTML LANG=xx>.  The relation with HTTP is that whaever
syntax is used, the server put it on the header.  For other protocols it
would have to be defined, but it should behave in a similar fashion.

> (And lobby the client and server vendors to support the full I18N spec
> as soon as possible).

I have the feeling that they will move in this direction.  By the way. the
doc producers also.

Received on Friday, 28 February 1997 09:55:18 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:40:40 UTC