Re: [CSS21] response to issue 115 (and 44)

>   - Most HTTP servers don't send the charset param, we're not going to
>     change that overnight.

I don't really see any prospect at all as the problem isn't technical.
It arises from:

1) service providers who only give full server metadata access as
a premium service (possibly bundling content only access with
basic connectivity).  This means that people learn web authoring
without access to such meta data, and never realise that the 
option may exist.  (Popular text books tend to address this
audience.)

2) there seems to be a strong psychological imperative to have single
technology solutions[1] (CSS only gets past this because people use
the term HTML to include it and several other things, as well as HTML).
Therefore, if things seem to work without access to server metadata, 
people will not try to learn about it.

These is in addition to the historical issue that people outside the
USA and Western Europe learned on browsers that treated the transfer
character set, the SGML character set and the display code page to
be the same thing, and ignored charset information.

> @charset, the heuristic that works best in practice is to assume the
> same encoding as the document that linked to this one. Another problem

Agreed, but with the added subtlety that one should assume that it is
the same, but unknown, ASCII superset if one gets a host document with
no character set specified.  (Although CSS capable browsers are likely
to resolve this by selecting a character set, rather than passing 
data through, raw, to the display font encoding.)

[1] This is why standards bloat.  HTML 2 is something of a counter
example, but SVG seems to be going the whole way.

Received on Friday, 20 February 2004 17:56:41 UTC