Re: [Digir-dev] Question on DiGIR Darwin Core and uniqueness of Catalog Numbers

[Robert A. M. <ra...@cs...>]

As you possibly know from talking to Gregor in Oaxaca, SDD and ABCD have 
a somewhat  related problem which will be discussed next week at the 
Berlin meeting
http://efgblade.cs.umb.edu/twiki/bin/view/SDD/SDD2004Berlin

Our issue has been called "TransformationHistory" and is meant to
memorialize changes made to information as it goes through various
providers. It is particularly meant to support managing attribution.

There is not yet a narrative description of the thinking on this, which
Gregor synthesized from a meeting between Walter, Gregor, me,
<an ABCD programmer from Spain> and <?>, but there probably will be
shortly.

Meanwhile, the adventuresome could persuse the current frozen version of 
SDD, downloadable from the SDD Wiki at 
http://efgblade.cs.umb.edu/twiki/pub/SDD/SchemaChangeLog091EarlyBetaVersions/SDD_091beta10.zip
The issues are well laid out in Gregor's usual extensive annotations
in the Types whose names begin with "TransformationHistory".

--Bob


Blum, Stan wrote:
> 
>>-----Original Message-----
>>From: Jim Croft [mailto:jr...@an...] 
>>Sent: Monday, May 10, 2004 6:11 PM
>>To: Blum, Stan
>>Cc: Donald Hobern; DiG...@li...
>>Subject: RE: [Digir-dev] Question on DiGIR Darwin Core and 
>>uniqueness of Catalog Numbers
> 
> 
> [snip]
> 
> 
>>>Having multiple resources under a DiGIR provider creates a 
>>
>>separation 
>>
>>>of
>>>the distributed query technology from the way people group 
>>
>>their records 
>>
>>>into resources; there doesn't need to be a one-to-one correspondence 
>>>between provider and resource.
>>
>>In that one provider can serve several resources, but not the 
>>other way around?
> 
> 
> We spent a fair bit of time in Oaxaca talking about the many cases in which a
> single resource is served by more than one provider.  Actually, these are
> cases where several *copies* (not necessarily identical) of a resource are
> being served by more than one provider.  We don't have a solution to this,
> yet, as it seems to be a problem primarily for people who are interested to
> know how many records we have on line.  (And the people anticipating that
> different copies of the same resource are eventually going to give different
> answers to a question.)  
> 
> Even more to the point, there is nothing that prevents a single record from
> being exposed/served through two or more resources (imagine a taxonomically
> defined resource and a geographically defined resources from a single
> provider).
> 
> 
> 
> 
>>Is the one-to-one correspondence [between provider and resource] something
> 
> we should be 
> 
>>striving for or at 
>>least recommending?
> 
> 
> I don't know.  Does the simple web services approach say that a separate URx
> should be available for every resource?  I think this is an important
> question.
> 
> 
> 
>>>I also want to point out that a provider does not correspond to an
>>>institution.  Many cases exist where an information resource 
>>
>>is "provided" 
>>
>>>(exposed via DiGIR) by a different organization than the one that 
>>>maintains it.
>>
>>This is both a good thing (in terms of providing access to 
>>data that might 
>>not be able to get there by more direct means), and a not so 
>>good thing (in 
>>that the identity and credit of the data maintainers seems to get 
>>downplayed in the DIGIR provision process).   I think there 
>>is still some 
>>work needed in this area as the data passes through several 
>>sets of hands 
>>on its way from source to destination.
> 
> 
> I don't agree that it's a problem with DiGIR, but with the people who are
> configuring DiGIR providers/resources.  Our view is that the technology
> should be made as transparent as possible.  An end user should see a list of
> resources in a DiGIR portal almost the same way they would keep a list of
> addresses for written requests.  Resources should be identified primarily by
> the custodians or owners of those resources.  The technical provider has done
> something they shouldn't, the original resource manager needs to take that up
> with the provider.  
> 
> The more difficult problem is going to come with the Russian Doll model for
> data, where a provider doesn't just make other peoples' resources available,
> but "adds value" to them by doing things like "correcting" taxonomic names,
> adding latitude-longitude, or converting lat-lon to a consistent map
> projection.  This will come.  What problems will it create?
> 
> -Stan
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by Sleepycat Software
> Learn developer strategies Cisco, Motorola, Ericsson & Lucent use to 
> deliver higher performing products faster, at low TCO.
> http://www.sleepycat.com/telcomwpreg.php?From_______________________________________________
> DiGIR-developers mailing list
> DiG...@li...
> https://lists.sourceforge.net/lists/listinfo/digir-developers
> 

-- 
Robert A. Morris
Professor of Computer Science
UMASS-Boston
http://www.cs.umb.edu/~ram
phone (+1)617 287 6466