Lists Home |
Date Index |
- From: Paul Tchistopolskii <email@example.com>
- To: "XML-Dev (E-mail)" <firstname.lastname@example.org>
- Date: Wed, 13 Dec 2000 19:38:49 -0800
From: Eric van der Vlist
> Tim Bray wrote:
> > Probably Dynabase, just like infoworld.com; it's the only product
> > I know of that serves HTML pages with the .xml extension (a practice
> > that seems more than a little weird to me). -T
> Most of the XSLT servlets are doing so, at least in their basic
> As Didier mentioned, the source document is really a XML document that
> is transformed on the server.
... And I agree with Tim. This *is* weird. ( And I was stupid.
I'm blaming myself here, because Hiawatha also returns html result
of server-side XSLT rendering when receving request for some.xml ).
But there is some twist with Hiawatha.
Because most of links from outher space to my site are .html ( and because
I'm trying to keep the legacy of my URLs so that people will not get 404
from my site ever ), Hiawatha also got a URL rewriting layer.
I think in 'ideal world' URL rewriting should be a *practice* not
the exception, so that just looking at some URI we can see
what will be a mime-type of the response.... for example ...
Well ... and maybe we can also get some other information
from the URI ? I mean, for example, what protocol will
be *acepted* by that URI ? Oh, no....
Let's face it.
URI is very important chunk of the information. I wish some day
we'l get some *recommendations* for *possible* patterns to use
I'm talking about the soft *recommendations*. I'l be strongly
against W3C *restricting* URI string in any way.
But I feel some need in recommendations. For example,
when I was hacking XT to understand document( "/! sql request here" ),
I got a feeling that the freedom with URIs is not actually good. For example,
the "/! " signature was kind of unavoidable, because of some Java
restrictions. I mean that some recommentations for 'custom' URIs
will be more forgiving to the legacy code than others.
Also I think that it was not so important when every server
were HTTP and every content was HTML.
The era of HTTP / HTML-only Web is gone. I think we should start thinking
in terms of many protocols and many mime-types.
> It can be seen as weird, but not more that serving HTML documents with
> .php or .asp extensions since in this case it's also the result and not
> the source that is sent to the browser.
I think this practice exists only because there is no simple and clear
URL-rewriting layers in most of the HTTP servers I've seen.
I think situation is somehow similiar to HTML. What you ( and many
others, including myself ) are saying : "URI is confuizing? Big deal -
it works" , is in fact close to saying : "HTML is not well-formed?
Big deal - it works".
PS. Oh, gosh. It comes to the basics again. "What is URI" ?
"Why namespaces are using URIs - is it really reasonable?"