Re: [xsl] hard problem

Subject: Re: [xsl] hard problem
From: Peter Flynn <pflynn@xxxxxx>
Date: Mon, 02 Feb 2004 14:10:49 +0000
On Sun, 2004-02-01 at 16:31, Nikolas Nehmer wrote:
> Hi,
> I have a really hard problem. Imagine you have a set of publications,
> consisting of books, articles, PhDThesises,... all modeled in XML(and
> corresponding XSD) files. All those publications have different
> characteristics, like different elements,...Every publication class is
> linked with a specialized XSL File for visualization purposes, for
> example a book with an ISBN Number is displayed differnt from an
> article,... Every publication is saved in it's own file.

In other words, a normal document collection :-)

> Now I want to create a list of all those publications without creating
> an XSL file with too much overhead. 

$ find . -name '*.xml' -ls | awk 'BEGIN {print "<docs>"} {print "<doc>"
$11 "</doc>"} END {print "</docs>"}' >doclist.xml
$ saxon doclist.xml stylesheet.xsl doclist.html

where stylesheet.xsl says something like

  <xsl:template match="doc">
    <xsl:value-of select="document(.)//title[1]"/>

You may need to detect the type of document via the XSD in order to
get the correct element holding the title, if they are named 
differently across document types. 

> In common object oriented
> programming languages like Java I would say OK every Object in my list
> knows how to visualize, so I just call every object's visualization
> method and I have created the list. 

In effect, this is what the above does.

> But unfortunately this is XML and
> not Java ;-( 


> So far my list consisted of several Xlinks pointing to the publication
> XML files. A specialized list.xsl file visualized those links and when
> you klicked the links, the spezialized publication visualization opened.

I think it's a bad idea to think of this from the idea of
"visualization". You don't want to visualize these files,
you want to extract some piece of data from them (the title).

> But what I want to do now is to make something like to call the
> publication's visualizer. In my opinion xlink:show="embed" would be a
> very simple solution to my problem but as I know this function is not
> implemented yet in browsers like Mozilla. Does anyone have some
> suggestions? Maybe something like simulating the xlink:show="embed"
> functionality on XSL level (in the list.xsl) could be a solution?!

Think like a tree, not like a chainsaw :-)


 XSL-List info and archive:

Current Thread