Sent by Jeremy Dunck on 13 June 2002 22:10
>From: "Joe Howley" [EMAIL-REMOVED]>
>To: [EMAIL-REMOVED]>
>Subject: RE: [css-d] CSS Rollovers - the saga continues with an example
>Date: Thu, 13 Jun 2002 11:07:56 -0400
>
<snip/ >
>Another Attack of the MIME Types, it seems - the server is sending .css
>files as text/plan when it should be sending them as text/css . Moz
>will only read stylesheets that are text/css, because, well, it's like
>that (yeah, I'm sure there's a better reason). I think it's pretty easy
>to fix; I have some basic instructions for apache somewhere if you want
>them.
<snip />
Actually, there is a very good reason for Mozilla to behave the way that it
does.
It's correct, according to the relavent standards (see references).
I could go on a rant here, but I'll just say this: fault-tolerant
algorithms in browsers encourage bad authoring habits, and keep blame for
bad implementations from being properly assigned. Here are a couple
relavent links for -that- train of thought. [a] [b]
You might ask why the standards were made the way they were.
You might say "it's silly for me not to have authority over a link that I
choose to make!". You'd be wrong.
1) On the web, server-side content negotiation [1] is possible on every URI.
That means it is possible that the representation that is returned to the
UA of you (the developer) is different than the representation of all of
your intended audience.
2) Further, resources controlled on my domain are under my control. Your
resource on your domain has no authority over me. [2] My server's
Content-Type takes precedence of your document's _expected_ content type.
Indeed, it may be that my current stylesheet could be replaced tomorrow by
-semanticly- the same resource (a different presentational mechanism) but
rather than being text/css, it is now XSL FO.
The point here is that just because you think some server must present a
particular representation, it doesn't mean it -will-.
There were a large number of bugs filled on Mozilla for this issue, and they
were right to stick by their guns. [3]
Now then, if you want to talk about the fact that it's difficult to remotely
configure some web servers to serve an arbitrary MIME type for some
arbitrary file, then let's go bother those server makers about it.
Certainly, don't blame the browser for being standards-compliant.
-Jeremy
[1] HTTP server-side content negotiation (Section 12.1):
http://www.ietf.org/rfc/rfc2616.txt?number=2616
[2] HTML 4 type attribute for LINK element:
http://www.w3.org/TR/html401/struct/links.html#adef-type-A
[3] Mozilla bug report:
http://bugzilla.mozilla.org/show_bug.cgi?id=150926
[a] Tag Soup discussion:
http://lists.w3.org/Archives/Public/www-html/1999Dec/0010.html
[b] The possible beginnings of the SGML working group that brought us XML:
http://www.nyct.net/~aray/htmlwg/stds.html
_________________________________________________________________
Chat with friends online, try MSN Messenger: http://messenger.msn.com
