Alex:
As always, your posts are a breath of fresh air!
For the record, FRBR is registered as properties here:
http://metadataregistry.org/schema/show/id/5.html These are the
'official' version, under review by IFLA, expected to be approved this
summer.
RDA is also available here: http://metadataregistry.org/rdabrowse.htm
These are currently under review as well, but the process is a little
less clear and the timetable a bit fuzzier. They are, however, usable
as is by those who wish to use them.
This IS the 'serious attempt' you spoke about--we're not just playin'
around.
Diane
On 6/24/10 5:00 AM, Alexander Johannesen wrote:
> Excuse me, coming through, pardon me! *shove* *shove*
>
> Bernhard Eversberg<ev_at_biblio.tu-bs.de> wrote:
>
>> But exactly "what is needed" is what we still don't know or
>> can't pin down.
>>
> No, that's not true. We can pin down a lot of the things we need, and
> we have, in all sorts of documents and websites and reports and
> whatsnot. We talk about it, talk it to death, write reports about it,
> fill websites with it, mailing-lists with miles and miles of prose
> about our needs and wants and yearnings.
>
> What you *haven't* done is written it down in such a way that we can use it.
>
> Think about it; all those models, formats, needs, wants and cries for
> help, they're all documented, written about, talked to death, and then
> we all return to MARC / AACR2 / RDA and keep going because no one has
> created The One True Thing That Encompasses All What We Need [TM]. You
> people need to get your head out of the format cloud and into the
> model cloud! You're *all* talking about a shared platform for
> librarianship, for sharing resources, meta data and whatever tickles
> your librarian fancy, yes? Well, the truth is that that model doesn't
> exist in a *format*, but in a *model*.
>
> Let's go over just a smidgen of our needs, like that of title. How
> many ways can we slice and dice the concept of "title"? Let's just
> give it an arbitrary number that's not true but will work for this
> example; 100. So we've got 100 ways of conceptually capture this
> "title" thing. What's the next step? It is *not* to create some XML
> for it, nor is it to create any other format record of it, because
> that stuff isn't important, it means nothing, it's just binary fluff
> and technological smoke and mirrors instead of what you actually need.
> You can bitch and argue for the way you build an XML or MARC record to
> represent your thing, but no matter how complex we document how that
> format shall capture it, there will always be some use case it won't
> capture (and we must all wait for the errata, or live with faults, or
> hack around it, or ...), or, probably even more correct, there's
> always a point at which the formats complexity overshadows its
> usefulness. (And I thought we invented the computer to help us out, to
> make our lives easier. Hmm)
>
> Let's make it a bit clearer. Instead of doing ;
>
> <format>
> <datafield tag="245" ind1="1" ind2="0">
> <subfield code="a">Arithmetic /</subfield>
> <subfield code="c">Carl Sandburg ; illustrated as an anamorphic
> adventure by Ted Rand.</subfield>
> </datafield>
> </format>
>
> .... create concepts out of them, document them, and agree to them.
> Just write prose; short, concise prose. If special cases are needed,
> create them and document them in the same place. Put up a darn Wiki,
> for heavens' sake! Join forces, and agree on what the heck you need
> right now, and let's create a model out of it, because it's TONS
> easier to extend a model than to extend an infra-structure!
>
> But let me dig into that of identifiers a bit more, as they are
> terribly important ;
>
> <controlfield tag="001"> 92005291</controlfield>
> <controlfield tag="003">DLC</controlfield>
>
> I'll be blunt and go straight for the jugular; what about that
> identifier? What the heck is that? Some internal identifier, an id
> from an OPAC, or external DB, or some serialisation process, what? You
> see such numbers being tossed around a lot, and some of them, like LOC
> numbers, gets used as persistent identifiers, yet you still got to
> hack your systems to give them meaning. And *this* one, the DLC, when
> we search for it (although this one we all know by heart) gives us ;
>
> 2035 DLC, normalized: dlc, obsolete code: DNLB
> United States, Library of Congress
> Variant name(s): Library of Congress, Congressional Library, United
> States, Congress, Library
>
> What is this supposed to mean? What does it mean that LOC issues this
> item the number "92005291"? It is meaningless at this point! Sure, we
> can, through the magic of "done it before", pop that number into any
> LOC service and possibly create some direct link to it (because we all
> assume it's part of LOCs union catalog, or some such) We still don't
> really know what that number is, we can only guess it.
>
> Why?
>
> For organisations that pride themselves on the bucket-loads of
> documentation, process and strict rules these *essential* questions
> about the meta data are rather absent. What do these numbers mean? How
> can we use these identifiers? Where are they applicable? Who can use
> them? How can we use them?
>
> Why aren't we coming up with an identity management plan that we all
> can use? Even FRBR sucks in this respect (concepts defined in a
> document that isn't addressable is not a good basis for persistent
> identifiers), and don't get me started on AACR2 / RDA. Why on earth
> isn't there better dialog about what concepts can be entities,
> enumerated values, or some such? Lots and lots of meta data is
> shareable, not as text, but as concepts! And, a crazy thing maybe, is
> that concepts are, again, easier to share than infra structure and
> formats!
>
> Here's a concept ;
>
> * book
>
> I'll write up a definition in my Wiki ;
>
> "A collection of sheets of paper bound together to hinge at one
> edge, containing printed or written material, pictures, etc." (And
> don't nitpick on this definition; it should be more specific, better,
> longer, crafted by more, agreed and discussed, blah, blah, blah ...)
>
> I'll link it to a definition ;
>
> http://en.wiktionary.org/wiki/book
>
> What happens here is that the concept "Book" has two identifiers ; one
> to our Wiki, and one to Wiktionary. The prose of both might and will
> change over time, but the two (or more) together make up a thing that
> most people would agree is a "Book."
>
> Libraries have a rich language that needs to be formally defined and
> identified so that we all can use and reuse them. Why hasn't this
> happened? Well, I suspect librarians are happy to talk their language
> and don't see the need for writing this stuff down, they're happy to
> *use* their language to document *other* things, like processes, meta
> data, documentation, systems, and so on, but you've forgotten to
> define your language *itself*. The need for this is *huge*, and let me
> clarify this with the example I usually give ;
>
> * Stacks
>
> No one in the whole world can use this concept except librarians.
> People read "Stacks", and have no understanding of what it means (do
> they mean what is physically sitting in a stack? Is it a synonym to a
> shelf? Or is it their stacks of stuff? A collection? Or all their
> collections in one?), neither orally from a library nor from a library
> system. But look, we can define a new concept ;
>
> * Collections
>
> Now we can edit the prose of that, and create a link ;
>
> Stacks
> is_really_another_word_for
> Collections
>
> We need the library world to be mapped. Not extensively, but enough
> for us to define, model and define that world, and we need to do that
> for ourselves but also for those who wishes to cooperate with us. We
> need these definitions to work as translations between systems. Right
> now, all we've got now is (mostly) MARC records doing the translations
> of computer records between systems, so the *model* we're basing our
> systems on are whatever is defined in, probably, MARC21! We cannot
> escape MARC through FRBR / RDA if these things don't do a better job
> of identifying the stuff they're trying to save us from.
>
> So what I'm really saying is that this discussion shouldn't really be
> about formats, but about a collaboration between all librarians to
> build up a system of identifiable concepts that can be reused in all
> sorts of weird and wonderful formats, systems and applications.
>
> I think Karen Coyle and her posse is doing something lightweight for
> the FRBRizing RDF thingie, right? How about we escalate this into a
> serious attempt by all to actually save the library world from being
> forgotten by the future, like some old ancient language no one speaks
> or understands anymore?
>
>
> Hope I was making sense,
>
>
> Alexander
>
Received on Thu Jun 24 2010 - 07:06:43 EDT