Re: [xml-dev] XPath 2.0 Best Practice Issue: Graceful Degradation

["Andrew Welch" <andrew.j.welch@gmail.com>]

On 28/01/2008, David Carver <d_a_carver@yahoo.com> wrote:
> Rick Jelliffe wrote:
> > One of the many troubles with XSD-style schema systems is that they
> > encourage a big bang mentality, where you have to try to make all
> > sorts of decisions about structure and type that are unnecessary or
> > which can be left to emerge, while making it difficult to add
> > constraints based on feedback from processes about what the real
> > problems might be.
> Rick, I'm going to have to disagree here.  This isn't a XML Schema
> problem, it's a problem with any poorly designed data model.  This can
> happen with SchemaTron and RelaxNG as well, not just XML Schema.
>
> You can apply agile techiniques, with continuous integration, and
> adaptability with XML Schema...

Indeed its all about being prepared for change, and having the energy
and willpower to do it.  It's nothing to do with the limitations of
XML Schema - you can rewrite the schemas, parsers, seralisers and the
object model at any point.  It's a lot of work, but it's better in the
long run than wedging extra information into an existing schema when
the requirements change.

Fwiw, I worked for a company while it was becoming ISO 9001 and cmm
level 3 accredited... it was all pretty meaningless, just a tick-box
requirement from those that bought the software to ensure we worked to
certain "standards" (commenting of code was one iirc).  The software
was still late and buggy, and because it was made using the waterfall
based inpenetrable-million-page word document, it was often not quite
what the customer thought they were getting.



-- 
Andrew Welch
http://andrewjwelch.com
Kernow: http://kernowforsaxon.sf.net/