Hiya,
Bernhard Eversberg <ev_at_biblio.tu-bs.de> wrote:
> Fine ER diagrams at
> http://www.rdatoolkit.org/background
Yes, they don't look bad at all. Well done!
Actually, I do remember seeing some early versions, and especially one
where all of the ERs were together. I think I blogged doom and gloom
about the complexity involved, and indeed there is a great complexity
in RDA that does not come out in these diagrams, but maybe that's a
good thing for getting people onboard.
> I agree with Jim that the RDA/FRBR approach at the data model is
> an exaggeration. It is much too firmly rooted in the 19th century
> traditional craft of bibliography.
Indeed, there's a heavy, heavy bibliographic leaning in there,
probably an attempt to schmooze with AACR/MARC21 and seem somewhat
familiar to the people who'll handle it.
I keep noticing other things that perhaps bug me more, but are more
subtle, for example where an item has "restrictions on access",
"identifier", and "URI" where the web and the future of cataloging
easily mash these into one (not to mention the wording 'restrictions'
rather than, say, 'constraints') and perhaps show an antiquated model
for all three and what they mean. Or we find nagging meta data
hangovers like for manifestations ;
Computer Carriers
Computer Card, Computer Chip Cartridge, Computer Disc
Computer Disc Cartridge, Computer Tape Cartridge
Computer Tape Cassette, Computer Tape Reel, Online Resource
and then we go to ;
Video Carriers
Videocassette, Videodisc, Videotape Reel
I know this isn't an exhaustive list, mind you, but there's a
disconnect between the physical and online world in the way the model
is being used. It's more like mashed together slightly semantically
ambiguous rather than building a real model in taxonomical form. Hmm.
All of these could be simple Containers that are either Physical or
Virtual, with a name attached instead which would be very easy to use
and reuse in search, but no, with the current model we either have to
pre-process our data into a simple indexed model, or write code that
slows down the search to nip out the semantics of the models we've
got. I can only guess this is legacy stuff, but I reckon you should
seriously look into a more modern approach.
> For the manifestation level, I think there's an excess of elements.
Amen. :)
> The most valuable part of RDA will be the vocabularies. They should put
> an end to the horrible inconsistencies we now have in legacy data.
Only if the RDA model match what people are doing. I suspect it is too
complex to fit in people's heads easily (because the RDA model
semantically ambiguous models it becomes so much more complex).
> Presently, it looks like RDA will remain a closed box with no way of
> open linking into its content. Leave it that way, and you can forget about
> it. The LC tests will not result in anything close to a full
> implementation anyway, and that will be the end of it.
What does RDA people say about this, though? Surely there's a good
plan put in place, a way to sway the masses, a list of vendors that
are ready to jump onboard ... Tim, you've done bits of this, no?
Alex
--
Project Wrangler, SOA, Information Alchemist, UX, RESTafarian, Topic Maps
--- http://shelter.nu/blog/ ----------------------------------------------
------------------ http://www.google.com/profiles/alexander.johannesen ---
Received on Thu Jun 24 2010 - 08:44:59 EDT