Re: XML Schema

[John Palmieri <johnp martianrock com>]

James Henstridge wrote:

> On Sat, 23 Jun 2001, John Palmieri wrote:
>
> > Sorry for double posting but my computer's clock was messed up.
> >
> > I'm getting a little bit of a feel for XML Schema's.  The documentation
> > islittered with new terminology that takes a bit getting use to.  The
> > examples look straight forward.  I think they are a better way to
> > represent the shapes instead of DTD's.  I will work on moving the
> > widget.dtd to UrShape.schema once the C++ stuff James is doing hits the
>
> Which C++ stuff is this?  I don't really see the benefit of changing any
> of the existing infrastructure over to C++ -- it sounds like a lot of work
> for not much gain (people have to learn the new infrastructure, a lot more
> work to write language bindings, etc).

No C++.  James Lowden wrote in one of his e-mails - "I'll next define such an
object notionally in C++ for us"  I'm assuming he just needs to define it in
his own terms so that he can better understand the problem.  I would have
preferred UML but whatever gets his brain moving is good with me.  The end
result will be coded in C within the same format as Dia.

> > streets.  About the validator - I'm going to fiddle around with this
> > (note the fiddle part).  Schema's is a whole other project and the specs
> > are just way too complicated for us to be worrying about that.  If
> > anything, I might be able to squeeze something out that is tailored to
> > our purposes (implements only the types and tags that we use).  I
> > suspect that libxml will include support for this in the future since it
> > is very useful.  One thing that gets me is the regular expressions.
> > What notation of RE are they using? Perl uses different notation from
> > python and lex.  Are their libs for this already? Well I don't think
> > that we are going to have to use restrictions or unions or any of the
> > more complex parts of the spec so I think we can just ignore them.  Well
> > I'll create the schema file and see if I want to tackle validating it
> > also.  If we get past that we can have it automagicly create a tree to
> > conform to our mini-DOM interface.  Comments?
>
> Remember that a dtd or schema is only a very small part of these sort of
> changes.  It is just a definition of a file format.  Coding up a new
> feature is the hard part (and you will probably find that the file format
> needs changing as things are implemented).  Don't spend too much time just
> planning things.

The whole Schema thing is just a side project for me.  Something I can work on
separate from the other guys since I can't work with CVS at the moment.  In
the long run a Relax-NG (the schema dialect I am using) could make development
faster since a compiler could output code whenever a change is made the file
format.  Again, I said fiddle around with it.  Once I get code from the other
guys and I set up the network in my parents house I will be able to code on
the more important stuff.

> James.
>
> --
> Email: james daa com au
> WWW:   http://www.daa.com.au/~james/

--J5