Re: Regional iTLD Question

["Martin J. Duerst" <mduerst@ifi.unizh.ch>]

On Thu, 28 Nov 1996, Administracio del Sistema wrote:

> The main problem:
> 
> Catalonia has about 7 million inhabitants.
> Most of them use catalan language, which is a very
> different language than Spanish (nothing to do).

I don't want to start a discussion here, but to be
exact, another language used in Spain, namely Basque,
has absolutely nothing to do with Spanish nor with
any other language in Europe. Catalan is a member
of the Roman language group, and goes together with
Latin, Italian, French, Spanish, Rumenian, and so on.
Although somebody knowing only Spanish will not be
able to read Catalan, (s)he will be able to guess
quite a bit.

> Most of the web sites here are in Catalan. Our addresses,
> have the .es suffix. Lots of people of overseas enter
> these webs and they say 'Ough ! this is not spanish !!'.

On a political scale, it might be fancy to have a .ct
suffix. But probably we would then have lots of regions,
states, and provinces also want to have their own TLD.

However, if you look at it on a global scale, language
does not coincide with political boundaries anyway
(I can give dozens of examples)
and even where the language boundaries more or less
coincide with political boundaries, universities
and other people may publish documents in many other
languages. Language therefore should be treated
orthogonal to political boundaries. For HTML,
to which you are referring, this actually has been
done:


As one of the coauthors of the html i18n standard
(draft-ietf-html-i18n-05.txt, soon to become an RFC),
I may point to another, probably more appropriate way
of solving this problem:

HTTP defines the "Accept-Language" header. Readers
can set this on their browsers. People that read
Spanish, but not Catalan, should set the preferences
accordingly. If there is only a Catalan version of
a document, the server will still be serving this.
Actually, it might be a good idea; more people will
get aware of the fact that there is a Catalan language,
and will try to read and understand the documents.
On the other hand, the server can (and should) be set up
to serve the document in the preferred language of the user
if several languages are available. Even if you don't have
a Spanish version, you might have an English version,
and some speakers of Spanish may prefer English over
Catalan, or Catalan over English, depending on their
knowledge of English.

HTTP also defines a "Content-Language" header. The
server can indicate the language of the document
with this header. The language can also be indicated
inside the document with the LANG attribute, which can
go with any HTML element. In this way, multilingual
documents can be marked up to any level of detail.

It is of course important for you to correctly
configure your server (e.g. the latest version
of Apache) and to correctly label your documents.
Anybody should do this.

Search services should take advantage of this markup
and serve only Spanish web pages for a user looking
up something in Spanish.

All this is currently well defined, and deployment is
continuing. With correct labeling and setup, you can help
advance multilinguality on the web.


Regards,	Martin.