New version of VO Support Interfaces: v0.26

[Ray Plante]

Hi,

On Mon, 7 May 2007, Roy Williams wrote:
> (4) If somebody makes a service that has fabulous science data but NONE OF 
> THE ABOVE, should the IVOA reject it?

I'm not sure exactly what was intended by the current version of the 
VOSI document.  I believe the wording suggests that the VOSI standard 
would be integrated into every standard protocol, with the MUSTs and 
SHOULDs applying.

There is another way to approach this document: as an additional 
standard a provider MAY support on top of the specific protocol 
specification in order to better integrate a service with the VO.  The 
"MAY", of course, means optional.  In fact, this is what I thought we were 
planning to do, at least initially (but maybe I was alone in this notion).

The motivation to support this interface is in capturing the highest 
validationLevel rating assigned to it by our registries.  That is, if it 
is supported, it will be easier for a registry to curate the record--the 
registry can keep the information up to date and can monitor its 
availability.

At some time in the future, a future protocol spec may say "you MUST
support the VOSI interface".  This might happen once we have (a) 
demonstrated that VOSI support is *necessary* in order to integrate the 
service, and (b) made it easy to suppport through the wide availablility 
of service toolkits.  I do not think VOSI should be required in advance of 
either of these.

cheers,
Ray