Harvesting

[Ray Plante]

Hi,

> Firstly, does everyone agree that harvesting is the next big issue? 

Yes, I agree that this is the next big issue.  The issue after that is the 
interface for querying registries.  

Of course, we're not finished with the previous issues, identifiers and 
the resource metadata.  Following on Bob's discussion of identifiers, we 
are starting to see some drivers for addressing the issue of persistant, 
location-independent identifiers, so we will need to address the question 
of mirrors that Bob and Pierre have discussed.  The biggest driver is the 
NASA ADEC effort to set up persistent dataset identifiers to be cited in
journals.  This brings up several issues that we should discuss elsewhere.  

Another aspect of identifiers that needs clarifying is the mechanism for 
securing namespaces and checking uniqueness.  Clearly, registries are 
needed for this: we know that our standard querying interface must support 
a simple "by-identifier" resource query.

> And, if
> so, what work has been done so far in Rwp04 (Keith?) and what needs adding
> to create a proposal?

A major question we need to answer is, can we use an existing standard 
(e.g. OAI or something from the Grid world), or is there a compelling 
reason to roll our own.

I'm partial to OAI given our success with it, but it does have a few 
disadvantages.  I'll try to put together a few sentences about this.

Meanwhile, I might point out for folks who might want to take a closer 
peak at OAI, you try out the prototype publishing registry from NCSA.  You 
can either visit our own installation at 
http://nvo.ncsa.uiuc.edu/nvoregistration.html, or you can install your own 
with our VORegistry-in-a-Box package
(http://nvo.ncsa.uiuc.edu/VO/software/VORegistry_in_a_box.tar.gz).  It 
comes with an OAI interface ready to go.  

cheers,
Ray