Another greatly delayed response, sorry....
Quoting Alexander Johannesen <alexander.johannesen_at_GMAIL.COM>:
>
> I agree with you. (I forgot to comment on this part of Karens
> otherwise good post) In the web world there's no such thing as
> "records." That is one of the biggies the library world needs to give
> up on. This is what RDF and Semantic Web technologies are trying to
> do; go from records to a big mesh of recursive key-value property
> pairs (tuples, triplets, etc.), and this clashes *tremendously* with
> records and closed databases and silos. It is its inverse, and you
> will *not* survive it unless this is re-evaluated.
I agree that the record is no longer the primary unit that defines
data, but I do think that we will gather together groups of data
elements (or properties, however you want to call them) for the
convenience of, for example, transferring all of the elements related
to a manifestation from one data store to another. I think the record
will look somewhat like the OAI or METS wrappers around a group of
properties. We also have to think about what administrative data we
will need and where it will be stored -- the idea of independent
triples is nice, but I also want to know who created them, when they
were last updated, etc. Eric Hellman covered this somewhat in this post:
http://go-to-hellman.blogspot.com/2009/06/triple-stores-arent.html
The key thing to me is that the data has to be able to live and be
useful outside of a record, or mashed together with any other data
elements. It's the ability to take apart and reform groupings that I
find exciting.
--
Karen Coyle
kcoyle@kcoyle.net http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet
Received on Thu Oct 29 2009 - 20:20:41 EDT