Re: [relax-ng-comment] We need an RNG schema inference tool

["Anthony B. Coates" <abcoates@TheOffice.net>]

** Reply to message from David Rosenborg <david.rosenborg@pantor.com> on Thu,
24 Oct 2002 18:57:43 +0200

> > We could use an Open Source tool that accepts arbitrarily many XML
> > documents and emits the "best" RNG schema consistent with all of them.
> > "Best" would involve balancing simplicity, terseness, and narrowness
> > (schemas should not accept more than necessary).
> > 
> > Microsoft has just announced such a tool for W3C XML Schema.
> 
> Apart from beeing an interesting programming task, what are
> the real world use cases for such a tool? Similar tools for DTDs
> have been around for years, but I've never felt even remotely
> tempted to use them. But that's just me ...

My experience is that a lot of people start developing their ML by writing a
DTD/schema instead of starting with sample instance documents.  People tend to
overrate their ability to write schemas and theoretically visualise the instance
documents will be like (just as most programmers overrate their ability to just
start coding without needing to plan what they are actually setting out to do).
Once you convince yourself that you should start the process with prototypical
instance documents, an instance-to-schema generator is a good, quick way to get
a starting schema document which you can then refine into the real thing.  I
would argue that this is a good development process, often overlooked.

	Cheers,
		Tony.
====
Anthony B. Coates, Information & Software Architect
mailto:abcoates@TheOffice.net
MDDL Editor (Market Data Definition Language)
http://www.mddl.org/