[Animl-develop] AnIML Schema Documentation draft available on CVS

[Mark F. B. <sa...@co...>]

I have popped a very prelimimary draft of my AnIML Schema Documentation on
the Sourceforge CVS for AnIML.  It should appear here when the servers
decide to publish it:
http://cvs.sourceforge.net/viewcvs.py/animl/documents/ and is called "AnIML
Core Schema Documentation.doc" even though it should also include the
Technique Schema.

What I would like you to consider is whether this draft is headed in the
right direction (it is a LOT of work).  I took the opposite tack to Maren
and decided that the term definitions were fundamental and that the
hierarchical relationship between the elements was already described in the
schema themselves, so I included some XMLSpy snapshots in an Appendix
(nonmandatory).  Even those are incomplete and I think clearly indicate how
newcomers find AnIML overwhelming.

As an illustration of what I have meant by a restructuring of AnIML:  I have
also sent up to the CVS an incomplete picture of another approach in which
XML Schema perfectly mimick database tables and their relationships and
constraints. I would attach it if I could, but go look here (AnIML 2.xsd or
AnIML2.gif if you prefer):
http://cvs.sourceforge.net/viewcvs.py/animl/schema/

It looks really different, doesn't it!

  a.. It is really clear where your data goes
  b.. There are no empty elements (at least not yet as I have not wrapped my
head around digital signatures, sorry!)
  c.. There is provision from the start for incorporation of database keys
  d.. Relationships are all spelled out in the constraint section instead of
by parsing endless Russian dolls (and even recursive!) hierarchies.
  e.. The data can slide into and out of databases amazingly easily.  An
AnIML database might look a lot like this.  Some programmatic tools allow
SQL queries of structures like this
  f.. It is strongly typed.
  g.. Extensions are possible by adding tables as long as they reference in
and don't demand any references out from the core.  I talked about this at
our first ASTM meeting and am not sure all reference requirements are yet in
place.
I am sure other intriguing possibiities are out there.

regards, Mark