Re: Linguistics versions

>M. T. Carrasco Benitez writes:
> > How should the server send the list of available linguistic versions:
> > 
> >  1) Alterates
> > 
> >  2) Content-Language
> >     Could be confusing as it would be the languages in the document and 
> >     the available linguistic versions.

According to HTTP/1.1, you must never use Content-Language for this: it
applies to the enclosed information only.  You can use Alternates from
draft-holtman-http-negotiation-04, but note that this document is still just
an internet draft, so do not use Alternates in a production server yet, only
in an experimental server.

> > Tomas
> > 
> > 


Gary Adams - Sun Microsystems Labs BOS:
>
>I think there is a fairly good proposal for handling variants of
>monolingual documents in the transparent content negotiation
>specification.  
>
>http://gewis.win.tue.nl/~koen/conneg/draft-holtman-http-negotiation-04.cm.html

Thanks for your support of this proposal.

>What may be missing is a way to describe a multilingual compositional
>document. e.g. This document is mostly English but contains a
>paragraph of Chinese text.

This is indeed missing, both from HTTP/1.1 and from transparent content
negotiation.  You _can_ label a document as `understandable to both english
and chinese language readers', but you cannot label it with `the reader must
understand both english and chinese'.  I don't know if this feature will be
missed though.  If at one point it turns out that people miss this kind of
labeling, some extension mechanism could be defined.

> Assuming the charset and language headers
>and tags have been declared properly, what must the browser do to be
>alerted to the fact that multiple fonts may be required in the
>rendering of the document? Is this even an issue?

If font support is an issue (and the opinions seem to be divided on that) it
could be negotiated on with feature negotiation, which is a part of the above
draft.

Koen.

Received on Tuesday, 14 January 1997 11:01:04 UTC