Re: [xsl] visual representation of imports/includes

Subject: Re: [xsl] visual representation of imports/includes
From: "Darcy Parker" <darcyparker@xxxxxxxxx>
Date: Wed, 21 May 2008 09:29:50 -0400
A thought:

I don't know of a particular solution, but perhaps you can create the
visual using graphviz?

I use XSLT to transform content into DotML and from DotML to the
graphviz .dot language, and finally to an SVG.  (Graphviz also
supports GXL, so it is possible to skip the step of going to DotML and
then .dot before creating the SVG visual output.  I go to .dot because
I like to fine tune some parameters that I wasn't able to control
easily in GXL.  And I found DotML easy to use for rendering .dot
files.)  With a combination of javascript and .svg, a nice visual
navigation could be built.  (but unfortunately .svg support in web
browsers is not great.)

http://www.graphviz.org/
http://www.martin-loetzsch.de/DOTML/

Darcy

On Wed, May 21, 2008 at 9:11 AM, Andrew Welch <andrew.j.welch@xxxxxxxxx> wrote:
>
> 2008/5/21 Dimitre Novatchev <dnovatchev@xxxxxxxxx>:
> > XSelerator has been doing this for more than 5 years.
>
> That's not quite what I meant (remember I was an Xselerator enthusiast too :)
>
> I was looking for an interactive diagram of the import/include
> hierarchy: the specific transforms at the top with lines down to the
> common transforms (for a simple set - the hierachy I'm looking at at
> the moment contains 100s and is a mess imho)
>
> Ideally each transform would be represented as a box, which could be
> expanded to show its constituent top-level elements
> (params/keys/vars/templates etc) by double-clicking.
>
> I haven't looked at it for too long, but Ken's XSLStyle appears to be
> able to handle import precedence and priority, so if the interactive
> diagram could show where a template/variable/param etc is overridden
> and provide a link to it (and vice-versa)  that would be an extremely
> useful feature.
>
> Does that sound feasible?
>
> What do others use for explaining a large codebase of xslt to someone?
>
> --
> Andrew Welch
> http://andrewjwelch.com
> Kernow: http://kernowforsaxon.sf.net/

Current Thread