Registry services

[Pierre Fernique]

Wil O'Mullane wrote:
 > Nvo searchable registry is a webservice already.
 > We are working on integrating the latest VOResource.
 > I believe we should mandate that Searchable Registries have to be 
webservices
 > and we should do a Registry Interface document (like the SkyNode 
document)
 > to describe that webservice interface. This would allow indendant
 > interoperable registries.
 > OAI would provide a non searchable non webservices based method for
 > archiove to publish a registry at a lower cost of entry. The searchable
 > registries then would need to harvest OAI and Other Registries.
 > This is the thinking in the NVO prototype.
 >
 >
 >
 >>Can I ask whether we should mandate a format for registry services?
 >>
 >>AstroGrid will provide its registry service as a web service. Is this the
 >>same for other projects? Given that the registry is something new, can we
 >>mandate that all VO registries must be web services? Or at least must
 >>provide a web service interface which meets our minimum standards for
 >>methods served up? This does not stop anyone providing other 
interfaces and
 >>other methods for internal project use
 >>
 >>Cheers,
 >>Tony.
 >>
 >>__
 >>Tony Linde                       Phone:  +44 (0)116 223 1292
 >>AstroGrid Project Manager        Fax:    +44 (0)116 252 3311
 >>Dept of Physics & Astronomy      Mobile: +44 (0)7753 603356
 >>University of Leicester          Email:  ael at star.le.ac.uk
 >>Leicester, UK   LE1 7RH          Web:    http://www.astrogrid.org
 >

GLU registry has both CGI interfaces and Web service interfaces.
Presently the CGI interfaces are more powerful and flexible, the WS
interface just allows a "remote GLU resolution".

I'm not sure that an agreement on Web service technology will be enough
to have interoperable systems. I imagine that we will have several
harvesting or replication WS methods, more or less different for each
registry system. But we feel that it would be much better to define
a standard on the data structures (meta data in this context) rather 
than on the methods to access them. Just an example: compare a WS 
solution to access catalogs to the VOTable standard. Which solution is 
really open, not technology dependent ? Does human readability really 
not matter ?


Cheers,
Pierre.