Rwp03: RSM changes

[Tony Linde]

> ... While everything is 
> a resource, not all resources are services.

Which is why I think we should refer to Resource Metadata. Text will then
explain that 'resource' applies to service and non-service-oriented items.

> I had assumed that all metadata elements would be "filled in" 

I guess that refers to the mandatory/optional idea? I think we do need to
have mandatory and optional items - where optional items have a 'not filled
in' default and some items can take a 'not applicable' value.

> I am not sure about Supported Metadata Formats.  In looking 
> back over the OAI stuff, it seems to me that OAI has these in 

I only referred to OAI as the place I stole the name/idea from. I am
convinced that we must have the equivalent idea though: subsidiary sets of
metadata that apply only to certain types of resource. 

> .... Perhaps SMFs 
> could be a mechanism for handling UCDs, but it just does not 

Not just UCDs but for any resource-specific metadata. But UCDs and column
names are examples of metadata that apply to data-oriented resources.

> This all 
> seems to me to be requirements creep.

What requirements? :)

> granularity.  Is it really a problem if a registry query 
> returns with both MAST and HST?  Or with both Vizier and a 
> particular catalog in Vizier?  The associated metadata will 
> make it clear which is the more specific resource.  In 

It could be a problem (caller will query both HST and MAST when it only
needs to query HST perhaps). Which metadata will 'make it clear'?

> -----Original Message-----
> From: Robert Hanisch [mailto:hanisch at stsci.edu] 
> Sent: 27 May 2003 17:39
> To: registry at ivoa.net
> Subject: Re: Rwp03: RSM changes
> 
> ...