Re: Loading an external file of index

Subject: Re: Loading an external file of index
From: Claudio Sacerdoti Coen <sacerdot@xxxxxxxxxxxxxxxxxxxx>
Date: Tue, 27 Jun 2000 19:33:06 +0200
On Mon, Jun 26, 2000 at 10:09:48 +0100, Kay Michael wrote:
> XSLT section 12.1 states explicitly (in response to a comment from me, I
> might add), Two documents are treated as the same document if they are
> identified by the same URI. It explains that this means
> generate-id(document("foo.xml"))=generate-id(document("foo-xml")).
> In my original comment I suggested the clarification
> count(document("foo.xml") | document("foo.xml"))=1
> which I think is even clearer. If you call document() twice with the same
> URI, you should get the same tree back each time.

 I understand (this is clearly functional), but is this really reasonable?
 I mean, if the URI is an URL of a dynamic document that changes every time
 it is getted (e.g. some CGI), then I have no possibility at all of
 retrieving the newest version.
 A simple example: we can imagine of implementing a simple CGI that returns
 an XML page with the root element having a random number attribute.
 Every document() call should issue a new run of the CGI and return me a
 new nonce.

 I currently use this "feature" (but I should say bug!!!!) to implement
 "xsl-extensions" in a standard (but I should say non-standard!) way.

 I would prefer this requirement to be got rid of in the XSLT spec.
 Instead, I would put a special PI (or something similar) telling the
 XSLT processor that the document retrieved is ever the same.

 What do you think about this? Am I brain-damaged or what?


Real name: Claudio Sacerdoti Coen
Undergraduate Computer Science Student at University of Bologna
E-mail: sacerdot@xxxxxxxxxxx

 XSL-List info and archive:

Current Thread