On Friday, July 27, 2001, at 09:47 AM, Michael Gerdau wrote:
<snip>
> ...and that's certainly not my intention.
>
> Give me some time (a week or two) to finish a current deadline. Meanwhile
> I'll try to get started with texen/torque/velocity so that at least I'm
> able to correctly use the words ;-)
>
> Once I know better what I'm talking about and of course see what's already
> there I will try to replace the existing xsl scripts (the ones in the
> proposal aren't up to date anymore :). This should function as a first
> "template programming lesson" and I'd be happy to share the results and
> get pointed out poor use of concepts etc.
sounds reasonable :)
here's something to help get you started...
there are two general approaches to using texen to generate source from
xml (at least that i know about). the first is to use the xml more or less
directly by placing a generic document object model (for example DOM or
JDOM) and a walker in the context. the other is to create a (specialized)
object model mapped from the xml which can be used directly. the second
method requires more code but produces simpler templates. i have code
(that originated over in jakarta-ecs) which uses texen to generate source
code (usually beans) from DTDs (via xml-schema). i'm keen to follow up
jason's suggestions and look at generating digester rules so that you can
take a DTD, generate java beans and then generating processing rules. sort
of automating the initial production of an object model from a schema.
so i might have a go at generating a bean-based object model and digester
rules from your DTD. this should save you a bit of time. i can also let
you have a look at some of my texen example code to get your started.
(texen doesn't really have any really good examples at the moment). i have
some pretty nice ant build scripts for use with texen which you might also
like to have.
- robert
