> From: Code for Libraries [mailto:[log in to unmask]] On Behalf Of
> Ross Singer
> Sent: Thursday, July 16, 2009 11:07 AM
> To: [log in to unmask]
> Subject: Re: [CODE4LIB] Open, public standards v. pay per view
> standards and usage
> On Wed, Jul 15, 2009 at 8:57 AM, Ray Denenberg, Library of
> Congress<[log in to unmask]> wrote:
> > Ross, if you're talking about the ISO 20775 xml schema:
> > http://www.loc.gov/standards/iso20775/ISOholdings_V1.0.xsd
> > It's free.
> It's also not a spec, it's a schema. If the expectation is that
> people are actually going to adopt a standard from merely looking at
> an .xsd, my prediction is that this will go nowhere.
> I mean, I'm wrong a lot, but I feel pretty good about this reading
> from my crystal ball.
Not saying you're wrong Ross, but it depends. People adopted MARC-XML
by looking at the .xsd without an actual specification. Granted it's
not a complicated schema however, and there already existed the "MARC 21
Specifications for Record Structure, Character Sets, and Exchange Media"
so it wasn't a big leap to adopt MARC-XML, IMHO.
Generally I agree with your conclusion Ross. It's difficult for people
to just pick up an .xsd and understand what the semantics are for each
element and attribute in the schema and which element(s) should be used
for the document element. This is mitigated by annotations in the .xsd
for the elements and attributes and also mitigated by using the Russian
doll schema approach, that MARC-XML uses, so it's clear what elements
can be used for the document element. Also tools like XMLSpy that
provide a graphical representation of the .xsd can provide insights
into how the schema should be used.
But these are a lot of if this and that was done, and you have appropriate
tools. A freely available specification detailing each element and
attribute along with their semantics is much better for understanding a
schema than the schema itself, but obviously the schema is the definitive
authority when it comes to generating conforming instance documents.