[DAS2] toy - das2 registry

[Andrew Dalke]

Andreas Prlic wrote:
> A  "toy" das2 registry serving das1 servers,  via das2 responses can 
> be accessed at
>
> http://www.spice-3d.org/dasregistry/das2/sources/
>
> I will work on adding the first das2 servers tomorrow.

There are differences between this and the spec.  These are

"CATEGORY" -> "CAPABILITIES"

Andreas knew that but didn't get it changed before having
to head out for a bit.

"testcode" should be "test_range" - it was added this afternoon
but I changed the name on Andreas.  (He agreed to the change.)

   # this is range string (eg, "Chr1/1:100" or "CloneABC123/500:599")
   # used in an "inside=" feature query.  It is used by the registry
   # server when doing a heartbeat check.
   attribute test_range { text }?,

The underlying problem is that a web server can be up while
the back-end database is down.  While a server should report
that as an error, sadly that's not always the case.  This
test_range is used by Andreas registry server in a periodic
feature query.  It should return a "reasonable" number of
features.

I decided to make it part of the spec for two reasons:
  - it simplifies auto-fill-in during registry discovery
  - clients can also use it to query the server and see if
      it's really alive or if it really means to return
      an empty list of features all the time.

It is optional.


The MAINTAINER "name" was required.  Andreas has examples where
there is only an email address and wants the name to be optional.
So now "name", "email" and "href" are all optional.  I would
like that one must be provided.

Finally, the "taxid" in the COORDINATES is optional.  The
RNG schema thought it was mandatory.

I've updated the schemas and the spec for the last two.  Committed.

Looks like I'll be spending most of tomorrow updating the
rest of the spec document.

I got a copy of Andreas' document and edited it to meet the
current spec and I've checked it in under
   "scratch/registry_sources.xml"
Feel free to test it out with your parsers.

					Andrew
					dalke at dalkescientific.com