[Nexus] XML questions

[Johnson, G (Greg)]

Hi folks,
  I've been delving into XML for the first time, and trying to understand
the NeXus format in that context.  One of the things that has struck me is
that DTDs appear to be on their way out.  XML Schema are the suggested way
forwards by many authors (at least of the books I've read!) for a number of
reasons.  In particular, the ability to specify data types for elements and
attributes.

As a learning exercise, I have taken the NeXus format, and written it (as
much as I can find of it) as an XML Schema.  As someone (Ray, I think) made
the comment on the website, DTDs are not easy to read.  Nor, frankly, are
Schema - but they do make it very easy to check that your XML is valid
against the agreed standard.  

Am I correct in thinking that a correct XML Schema would be a useful way of
validating _all_(HDF4/5 & XML) NeXus files?

Due to my inexperience with XML, I'm sure the Schema could use some work and
perhaps some design changes.  But for now, I think it works.  It seems to me
that using NeXus XML lowers the barrier to 'full' NeXus (HDF) usage - for
small data files, the XML tag overhead is not huge, and the benefits of
being semi-human readable are a good thing for scientists who want to be
able to put their data into Excel without using a special extracter.

I'd be interested in hearing other people's feedback about XML and Schema -
it seems like the email list is very quiet.  Are all the real discussions
going on in private email?

With regards,
greg

PS if anyone wants to have a look at the Schema, let me know.  Perhaps we
can stick it on the website, or something.

============================================================================
Greg Johnson                              Manchester Materials Science
Centre

Phone (Manchester): +44(0)161 2008959     Manchester University
Fax (Manchester):   +44(0)161 2003586     Grosvenor Street
Phone (St Andrews): +44(0)1334 463080     Manchester M1 7HS
Phone (mobile):     +44(0)773 0068991     UK
Email:          greg.johnson at ieee.org
============================================================================