Re: PaceMustBeWellFormed

[Walter Underwood <wunder@xxxxxxxxxx>]

--On Wednesday, July 14, 2004 5:27 PM -0400 Mark Pilgrim <pilgrim@xxxxxxxxx> wrote:
> If my wrestling with RFC 3023 has taught me anything, it is that HTTP
> has primacy.

That is true. The spec also has no way to handle the following
common case:

>     * The document author does not have any way to configure the
>       server to send the proper HTTP Content-Type header

This is a part of the HTTP spec which is simply not practiced.
Clearly, document authors need to specify encodings, and most
servers do not provide an easy way to do that.

Atom can specify that they must control it, but the spec won't
change how the web really works.

This is like the ROBOTS meta tag compared to the robots.txt file,
if that analogy makes sense to anyone. Authors are guaranteed to
have control over their own documents. They are not guaranteed
to have control over anything else.

Has anyone ever deployed a transcoding server? I've never seen
on in real use.

wunder
--
Walter Underwood
Principal Architect, Verity