[Top] [All Lists]

Re: [oor-forum] Defining "Ontology Repository" (maybe "OntologyRegistry"

To: "'OpenOntologyRepository-discussion'" <oor-forum@xxxxxxxxxxxxxxxx>
From: "Dennis Nicholson" <d.m.nicholson@xxxxxxxxxxxx>
Date: Mon, 4 Feb 2008 11:45:27 -0000
Message-id: <003c01c86723$6fdd9cd0$0dbb9f82@LAPDENNIS>

Hi Denise
For what it may be worth, I take the view that:
- There will ultimately be many (geographically distributed) 'repositories' each holding and serving up at least one KOS/Ontology/terminology. Even

   if they only hold one, they will need  some kind of local index describing what they hold, what its features are, and how it can be accessed

   (HILT, for example, is offering SRW/U access). Those running these repositories will not necessarily be directly collaborating and often won't

   know about the repositories run by others. Some, like HILT, may also hold data to support cross-KOS interoperability.

- There will be a need for at least one registry that allows the various (often unknown) repositories to be found and that describes them in a

   way that allows whoever or whatever (it will often be another service) is searching for them to determine whether they are useful for particular

   purposes (do they cover a particular subject area, are they designed for a particular use or user, do they offer access protocols in use by the

   requestor etc). Some of this data could - and probably should - be held by the repositories themselves, but there is an efficiency and

   convenience case for holding the data in a central registry also
- There is a question as far as I am concerned as to whether we need registries specific to terminology services/repositories or whether

   generic service registries such as IESR in the UK (http://iesr.ac.uk/) will do the job just as well or better, but, knowing how these things go, I

   suppose it is inevitable that we will end up with some generic service registries including terminology repositories and some terminology   

   service registries as well. It is also, it seems to me, inevitable that there will be many registries and that they will 'know about' each other, but

   won't individually contain information about all repostories (so that it will be necessary to find all the registries via a local one first before we

   can find all the repositories)
- At the moment, there is also a possibility that there will also be terminology registries that are not intended to be terminology services

  registries. I don't fully understand the function of these yet. The JISC project I mentioned earlier is looking at this (as I understand it). For me,

  these are just like federated 'repositories' according to the definition on this list and not terminology services registries. But I don't claim to

  be (a) an expert on the thinking behind them (b) sure I am right in any case. It's a new area, so these things are still topics for discussion.

  Again though, given the way these things go, I expect these terminology registries will also come into the architecture, and that some of them

  will end up being more like terminology repositories and others more like terminology service registries.


From: oor-forum-bounces@xxxxxxxxxxxxxxxx [mailto:oor-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of dbedford@xxxxxxxxxxxxx

Sent: 04 February 2008 11:07
To: OpenOntologyRepository-discussion
Subject: Re: [oor-forum] Defining "Ontology Repository" (maybe "OntologyRegistry" too) for the OOR Initiative



Thank you for contributing these ideas.  I have been monitoring the discussions in background.  I think it is important to start out the discussion with a context of the range of architectures that are possible (ranging from the practical/possible to the theoretical/idea).   Your ideas raise a lot of questions in my mind regarding how people will contribute, communicate about, leverage and map ontologies.  For a given architecture, are there a minimum set of requirements?   I may have missed this point in the discussion, but have we walked through any use cases?  What are the drawbacks and benefits of each architecture from a 'user' perspective (where user may be a person or an agent)?


My apologies if I am asking questions that have already been addressed and I've missed the discussion. 


Best regards,


Message Archives: http://ontolog.cim3.net/forum/oor-forum/  
Subscribe: mailto:oor-forum-join@xxxxxxxxxxxxxxxx 
Config/Unsubscribe: http://ontolog.cim3.net/mailman/listinfo/oor-forum/  
Shared Files: http://ontolog.cim3.net/file/work/OOR/ 
Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OpenOntologyRepository     (01)
<Prev in Thread] Current Thread [Next in Thread>