- From: Corey Harper <corey.harper@gmail.com>
- Date: Mon, 14 Jan 2013 11:49:09 -0500
- To: kcoyle@kcoyle.net
- Cc: Richard Wallis <richard.wallis@oclc.org>, public-schemabibex@w3.org
I agree with all of this about holdings, in principle, but I also worry that there are some aspects of our data that will be hard to capture without treating "holding" as an entity rather than--or in addition to--a relationship. My instincts tell me that we have an "instance" or "item" type, which then has a relationship to a library or other institution. This works fine for classifications, physical condition & preservation notes, etc, but falls short with serials, right? Unless we define our "items" as being either individuals or arbitrary sets/groups. Then, we could say that the serial "Nature" has a print "instance" related to (held by) NYU with such & such a call # and this *summary* of holdings, and an electronic instance with a different summary of holdings as well as relationships to NYU (by subscription), and a provider of, say, Proquest. (And in NYU's case, there are half a dozen such instances, (Gale, Ebsco, Ovid, Nature itself, ...) all with different "holdings". Are we comfortable defining the "physical" items as aggregations as well as individuals, and with lumping E- into that same model? -Corey On Mon, Jan 14, 2013 at 9:17 AM, Karen Coyle <kcoyle@kcoyle.net> wrote: > Richard, I'm grabbing some screen shots of library holdings displays so we > can look at them together. I'll just pop them onto a web page. > > sku is a product code, not a code for an individual item. ISBN is a sku. > Libraries do add a barcode to items for checkout, and it acts like a sku, > only it's at an item level. The call number is something else, because it > can contain location information. ("Fiction") I think the examples will make > it all more clear. > > kc > > > On 1/14/13 5:50 AM, Richard Wallis wrote: >> >> Would not 'sku' in http://schema.org/Offer serve this purpose. >> >> I'm thinking of a CreativeWork for the bib stuff, and SomeProducts for the >> manifestation level stuff linked using Offer to a Library - the offer sku >> providing somewhere for [what us library folks call] the call number and >> businessFunction for the types of things you can do with the item in >> question. >> >> ~Richard. >> >> On 14/01/2013 13:23, "Karen Coyle" <kcoyle@kcoyle.net> wrote: >> >>> Richard, I agree that "holdings" is hard to model as a thing. "Name" for >>> the organization is already included, as is location of the >>> organization. So far there is no "location" for the individual product >>> ("call number" in libraries) since other businesses do not provide >>> locations for individual items. In fact, there is very little item-level >>> information so far in schema. Libraries will need that for call number >>> (which is supposed to be unique within the library and to provide a >>> single place for each book -- although this isn't followed 100% in all >>> libraries), and the availability of individual items. So perhaps what we >>> should model is an item level. >>> >>> kc >>> >>> On 1/14/13 2:06 AM, Richard Wallis wrote: >>>> >>>> Hi Karen, >>>> >>>> Holdings is a difficult one. I have trouble in justifying, in data >>>> modelling terms, its existence as an an entity. Here is most of an >>>> email, in another thread I am in, on the subject. >>>> >>>> I still remain to be convinced that a Holding is a thing to be >>>> modelled as an entity in its own right. >>>> >>>> Surely the realisation of a holding is just the relationship >>>> between >>>> a thing (Book, Journal, License to access) and a location (Shelf, >>>> Library, Institution). Its not a thing or a concept. >>>> >>>> Schema, which would probably best describe an item the union >>>> between >>>> a CreativeWork and a Product. The SomeProducts[1] subtype of >>>> Product has the inventoryLevel property. That's what holdings are, >>>> a >>>> count of the number of items at a location. >>>> >>>> Trying to model, the phantom echo of performance enabling RDBMS >>>> denormalization in to a table called Holdings, is definitely a bad >>>> idea. >>>> >>>> My couple of cents.. >>>> >>>> >>>> I believe that those outside of the library domain have equal difficulty >>>> in understanding too. I know this might be a radical suggestion as >>>> holdings have been key to water-cooler discussions in libraries for >>>> decades. However my linked data background has taught me to model the >>>> real things in the real world, and I am yet to meet or pick up a >>>> holding. >>>> >>>> ~Richard. >>>> >>>> >>>> >>>> On 13/01/2013 15:02, "Karen Coyle" <kcoyle@kcoyle.net> wrote: >>>> >>>>> I wasn't quite sure where this went, but I added two objects to the >>>>> object-type page [1]: >>>>> >>>>> - the "library" object that is under localBusiness >>>>> - a new "library holdings" type. >>>>> >>>>> In each I put in some text about some new properties that might be >>>>> needed. >>>>> >>>>> I also have beefed up the commonEndeavor HTML example. [2] If you wrap >>>>> <html> around it is does actually display, although it's not very >>>>> attractive. Just pretend that there's some nice CSS involved that fixes >>>>> that. >>>>> >>>>> kc >>>>> >>>>> >>>>> [1]http://www.w3.org/community/schemabibex/wiki/Object_Types >>>>> [2]http://www.w3.org/community/schemabibex/wiki/CommonEndeavor >> >> >> >> > > -- > Karen Coyle > kcoyle@kcoyle.net http://kcoyle.net > ph: 1-510-540-7596 > m: 1-510-435-8234 > skype: kcoylenet >
Received on Monday, 14 January 2013 16:49:42 UTC