Re: LANG= for character-mapping

> - Only one charset in allowed per document.

Yes.

> - The "document character set" should Unicode; other are allowed.

No, others are not allowed for HTML. Maybe for other kinds of SGML.

> - The charset for transmission should be Unicode; other are allowed.

No, the charset for transmission should be whatever is appropriate for
the data, e.g., ISO-8859-5 for Russian might be appropriate.

> - The server should inform the client; charset = "UNICODE-1-1" (no sniffing).

Yes, the server should inform the client, but probably not
"UNICODE-1-1"; if you're using UNICODE-1-1-UTF8, then say so.

> - Transmissions transformations are for compressing, encrypting
>  (content-encoding) or "safe transport" (transfer-coding); but virtually
>  what it is sent is the charset.

I don't understand what "virtually what it is sent" means.

> - LANG is for higher functions, such as short quotations.

LANG is for marking something else; it says what language the
characters are intended to represent.

> - The server should inform the client with Content-Language.

When it's useful.

> - LANGs in the document overrides the Content-Language.

Yes.

> - There is no association between LANG and charset.

Not formally. Clearly, some charset parameters are more appropriate
for some languages.

> 2) HTTP needs some changes/clarifications

I don't think so. Your comments make it clear that you are looking at
the WRONG DRAFT, as the HTTP specification was revised substantially
in this area. If you look at
	draft-ietf-http-v11-spec-06.txt and .ps
you'll see that your comments no longer apply.

> This should be the ordered list of "prefered languages".
> The meaning of the quality factor "q" should be changed
>  from  "...estimate of the user's comprehension of that language..."
>  to    "minimun accepatable quality of the translation"

This was changed.

> - Content-Language
> This should be an ordered list; the first language should be the language 
> of the document transmited; the rest, the languages available.

No. Content-Language must describe the language of the
content. Anything else is not talking about the content, but about
alternatives.

> 3) HTTP should allow two type of conversations:

It allows more than that already.

Regards,

Larry

Received on Wednesday, 24 July 1996 05:05:23 UTC