Lists Home |
Date Index |
- To: Michael Kay <firstname.lastname@example.org>, 'Tatu Saloranta' <email@example.com>, firstname.lastname@example.org
- Subject: RE: [xml-dev] Bad programmers use recursion, xslt (Offtopic...)
- From: Michele Vivoda <email@example.com>
- Date: Mon, 27 Mar 2006 14:08:07 +0200 (CEST)
- Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.it; h=Message-ID:Received:Date:From:Subject:To:MIME-Version:Content-Type; b=Nyer4Ncmay5F6YDbVkz4aj/qSP7+b/qKyhg+J4h0FcixWSnoWWuFsDG85i58usRma0u08zSKw8KvSurklzOqkf3SMsVuAucWJSLtkmqnd5i/1eziZjXFqTZX7FYpoZbFnpNBJmy/tx5vCPBG7HBaoxvLlbEtPxYMFtvTa48Hu5Y= ;
> > if one does not define very well the
> > "xml interfaces" to communicate between
> > (the structure of the xml trees input of each
> > stylesheet) it becomes very difficult to
> > document and debug them.
> Agreed: there's a lot to be said for defining a
> schema for each intermediate
> document in the pipeline,
Exactly, if these intermediate formats already exist
is easier, otherwise one must define them, write their
> and using schema-aware
> stylesheets to trap the
> errors as soon as they occur, rather than allowing
> them to propagate to
> later stages in the pipeline.
Doing with schema is one option. Any option
one uses he needs to be very precise and methodic
and be well aware of the complexities involved.
Writing a schema or documentation is time consuming,
one should weight all these aspects.
I found myself sometimes like I was playing
chess in jumping from one stylesheet to the other.
For complex stuff a tecnique I use is printing
err:@* attributes during transformations
when something goes wrong and even as warnings.
Then I read two times the
output: in the first I collect errors and warnings,
then I present them to the decison-maker and in case
I continue with the next step, that simply ignores
ther err:@* attributes.
> > I tried to find something that allows you to
> > do multiple transformations in a declarative
> > way but I did not find much.
> There's a W3C activity starting up to define a
> pipeline processing language
> which will make it possible to do this in a portable
Almost scared by this ;-) joking.
It would be nice to do it
with the almost the same syntax of xslt but with
state inside this pipeline stylesheet.
> Meanwhile there are pipeline processors for example
> from Orbeon and from
> Markup Technologies (http://www.markup.co.uk/), and
> other people are using
> tools like Ant.
Thanks. I follow 2 approaches: or I append with the
the output of the first transformation
to the source, or I copy also the original tree
with the first stylesheet. In any of the two cases I
have to use some code.
Yahoo! Mail: gratis 1GB per i messaggi e allegati da 10MB