W3C home > Mailing lists > Public > www-html@w3.org > July 2004


From: Jukka K. Korpela <jkorpela@cs.tut.fi>
Date: Wed, 28 Jul 2004 15:29:40 +0300 (EEST)
To: www-html@w3.org
Message-ID: <Pine.GSO.4.58.0407281459550.4067@korppi.cs.tut.fi>

On Wed, 28 Jul 2004, Anne van Kesteren wrote:

> it wasn't really a link, but just a translation of a Dutch poem, but
> there isn't a way to say that the original poem was written in Dutch.

If I understand you correctly, you were looking for a way to say such a
thing in markup, without necessarily pointing to the original in any way.
While this might be relevant in some situations - it can be useful to know
that some text has been translated from a particular language, since this
may help to understand some odd formulations - it's hardly useable to
programs at the present level of the state of art. And to humans, it can
be expressed in prose after or before the quotation.

> > Irrespectively of this, I don't think a hreflang or a citelang
> > attribute is needed much. Your example refers to a document via HTTP,
> > and presumably to an HTML document; this means that one can use both
> > the Content-Language header and <html xml:lang="nl"> if desired.
> How can I show that to someone who reads my page?

You don't; it's not a property of your page but of the referred document.

> > The latter need not be known when the citing document is processed,
> > and it may in fact change independently of the citing document. For
> > example, you might wish to change the declaration of the language
> > from "nl" to "nl-NL" or "nl-NL-officialese"
> I don't really follow this.

Language markup is a very confusing topic. I was trying to make the point
that the language of a piece of text can be identified at different levels
of specificity. The language code could indicate just a "language" (which
is a debatable term - is a language a dialect with an army and a navy, or
is air force needed too? :-)), or it could additionally specify a
particular form of a language, such as 'Dutch as spoken in the
Netherlands', or even more specifically as particular variant of a
variant. I think it is up to the author of the document to make decisions
on this for the text in his own document.

> You could also want to extract the information using CSS generated content:
>    blockquote.poem[citelang=nl]::after{
>     content:"The original was written in Dutch.";
>    }

Yes, but if such information is relevant, it is probably relevant
independently of rendering style, i.e. it should be expressed using actual
content, not "generated content".

> Note that HREFLANG has changed in XHTML 2.0.

Yet another change in the current draft that I have missed to notice, and
a most unfortunate change.

> If I point to it like:
>   <a href="http://example.org/content" hreflang="nl">Content</a>
> ... the user agent should change it's 'Accept-Language' header to
> 'Accept-Language:nl', which is imo a great change.

Great indeed. And a fundamental confusion. The Accept-Language header is
meant to be used for interaction between a client and a server in the
interests of the user, according to the user's language preferences.

If there is a particular reason to refer to a particular version of a
document, when it exists in different versions selectable by content
negotiation, then an author should simply link to the document using a
language-specific URL. Surely the versions should have their own URLs as
well. One might be tempted into thinking that hreflang="..." is more
robust since the method of naming the different variants might change in
future, so that http://example.org/content.html.fr might become
http://example.org/content?language=fr for example. But cool URIs don't
change*), and for uncool URIs, the problem exists no matter what we do
with language selection.
*) That is, once you have announced a URL, you make sure it keeps working
as long as the referred document exists on the Web in any form, by using
whatever redirection or remapping mechanisms are available.

> HREFTYPE works in a similar way.

Then it has similar problems.

Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/
Received on Wednesday, 28 July 2004 08:30:37 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:16:00 GMT