[Top] [All Lists]

Re: [ontolog-forum] Practical onomastics...

To: <edbark@xxxxxxxx>, "[ontolog-forum] " <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "Godfrey Rust" <godfrey.rust@xxxxxxxxxxxxx>
Date: Fri, 21 May 2010 19:55:49 +0100
Message-id: <5327550D077F411FB36AAFB800165224@GodfreyPC>
> Godfrey Rust wrote:
>> the
>> issue that John describes reflects the real world problem of 
>> "meta-metadata"
>> being much larger than the metadata it is describing, which I am finding 
>> is
>> increasingly a practical concern as the need for identity becomes more 
>> and
>> more granular.
>>    (01)

Ed B wrote:    (02)

> I would argue that this version of 'turtles all the way down' is
> fictitious.  There are indeed cases in which there is value to a
> possibly very long source tree, like an etymology or a genealogy, but
> they are rare, and they all stop where some practical consideration
> intervenes, the usual one being the total lack of value to further
> pursuit.  (In the particular case of genealogy, the practical
> consideration is the absence of any reliable source, but that didn't
> stop etymologists. ;-) )
>    (03)

I wasnt talking about endless recursion (which I agree can normally be 
stopped simply enough).  The issue of meta-metadata I was referring to (to 
quote a real world example) is that in a system where data is being 
submitted by multiple sources, and may be conflicting (for example, in a IP 
database used in collective rights management, conflicts over the titles, 
contributors, contents or rights in a particular resource) then it can 
become necessary to record the provenance of each element of data at the 
level of a single attribute: both the creator/changer of the data and the 
authority on which it was maintained (including links to source 
documentation/message files), so that a single RDF triple of metadata may be 
supported by a dozen of meta-metadata. One consequence of this in an RDF 
database (as I'm not sophisticated enough to use CSV as John recommends) is 
that every individual attribute and link can become an fully-fledged entity 
with a URI (or your identifier of choice in another formalism) to support 
its meta-metadata, as well as any attributes and links which it may have of 
its own, of which - once you starting looking for them - there are often a 
surprising number. That's what I meant by the increasing granularity of 
identity, albeit "under the hood" in many cases. The more sources you have, 
the more important such granularity of provenance becomes if you require 
your data to be authoritative, which is a challenge for "linked data" 
initiatives at any level.    (04)

Godfrey     (05)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (06)

<Prev in Thread] Current Thread [Next in Thread>