- From: Karen Coyle <kcoyle@kcoyle.net>
- Date: Tue, 26 Mar 2013 18:59:48 -0700
- To: public-schemabibex@w3.org
Thanks, Richard. I looked at your examples on the vocabulary page, and what is there is - one CreativeWork (abstract) - 2 books (published) - 1 movie (published/released) The books and the movie all declare to be instanceOf the CreativeWork. The CW here seems to be the (by now classic) abstract bibliographic Work entity, such as the Work entry in a name/title authority file. Interestingly, this doesn't show, for example, a relationship between two books, or a book and a movie, that have any physical description. So the question is: does this apply to any relations between any CW's (including sub-classes of CW) at any level of abstraction/concreteness? Or only in the case where there is an abstract CW to point to? (The screenshot of your diagram seemed to imply a linearity from an abstraction.) If it can be applied to any CW/sub-CW, it would be good to have examples that don't hinge on an abstract CW entity like the one here. That, to me, tests the hypothesis that one must have an abstract entity in order to have "instances of." If an abstract entity isn't necessary, that should be made clear. However, without an abstract entity, it is less clear to me how one establishes (or explains, or even makes use of) precedence between bibliographic items. Perhaps a few use cases will help clarify? At the moment I can think of: - two versions of the book in two different languages (e.g. War and Peace in English and French) - a book and a movie (and you don't know which came first, cf. Star Trek) - two editions (e.g. 2nd and 3rd editions) There are undoubtedly a gazillion others, but we only need a start. A couple of examples of these would be helpful in understanding the use of this property. kc On 3/26/13 1:15 PM, Wallis,Richard wrote: > OK - I apologise for being a touch too vague in my description in an > attempt to avoid using words that could be construed to be something that > is already defined in FRBR or BIBFRAME. > > Let me try another approach. > > There is a spectrum of concrete-ness (from the abstract to the concrete) > that could be applied to CreativeWord descriptions. In frbr there are > four, in BIBFRAME there are currently two, out on the wild web there could > be anywhere between one and lots. > > Assuming it is a horizontal spectrum with most abstract to the left and > most concrete to the right - what I believe we are trying to define is > that the relationship between a CreativeWork description and another, > anywhere right of it on the spectrum, would be hasInstance. And the > relationship between one CreativeWork and another, anywhere to the left of > it, would be isInstanceOf. > > Better? > > ~Richard. > > > > > On 26/03/2013 15:32, "Karen Coyle" <kcoyle@kcoyle.net> wrote: > >> >> >> On 3/25/13 4:00 PM, Wallis,Richard wrote: >> >>> >>> >>> I am not suggesting that we use Schema.org to align terms in other >>> vocabularies. I believe that we should be striving for a set of >>> properties that would enable us to describe the relationships between >>> entities that can be described as a schema: CreativeWork. >> >> To me, that does not translate to "instanceOf." As we can see with other >> comments here, folks are taking instanceOf to be a relationship between >> an abstraction and something concrete (the instance). If I am not >> mistaken, using your definition "describe the relationships between >> entities that can be described as a schema: CreativeWork," instanceOf >> could be used between any two (or more) things that are described with >> schema.org/CreativeWork properties (or any of the sub-classes of that). >> As I've said before, this implies that there is not necessarily a >> dominant and a subordinate thing in the relationship. These could be >> what FRBR calls work/work or expression/expression relationships -- as >> opposed to the layered relationships of WEMI. >> >> If it is truly a relationship (of undetermined type) between entities >> that can be described using schema/CreativeWork then it would be best >> not to have directionality or subordination (one is an instance of the >> other) but a term expressing commonality (these two have something in >> common). >> >> I believe that abstract/concrete can be revealed by the actual >> properties used. A description with only an author and a title is >> necessarily abstract; one that includes the publisher and date is more >> concrete. The description itself is the indicator of the degree of >> abstraction. >> >> If this is the intention of the property -- to make a non-specific >> connection between any two or more CreativeWorks -- then it in fact >> negates BIBFRAME's Work/Instance and FRBR's WEMI because it is not a >> relationship between parts of a single description, as those two are, >> but it is a relationship between CW descriptions. That aspect *is* >> covered in FRBR in the entity/entity relationships but so far is not >> evident in BIBFRAME -- although it may come under the Annotation model. >> >> I feel we need to clarify this: is this a CW/CW relationship, without >> any concept of "broader" or "narrower"? >> >> kc >> >>> >>> As we have identified in previous conversations bibframe:Work & >>> bibframe:Instance and frbr:Work, frbr:Expression, frbr:Manifestation, >>> frbr:Item (and the entity types in that model only existing in my head) >>> all can be described as a schema:CreativeWork. >>> >>> So my model, the BIBFRAME model, and frbr are there to test how good, >>> what >>> we come up with, is at providing those descriptions. >>> >>> > "Instance/realization/derivation of the concept of this creative >>> work. >>> > eg. The paperback edition." >>> >>> > This is very specific, and has within it some strong assumptions >>> about >>> > CW. Is this what we want? >>> >>> It is only draft wording - in retrospect it would be simpler if we >>> dropped >>> 'the concept of'. >>> >>> >>> > Can we explore how it would be used with CW to >>> > interact with other models? >>> >>> >>> That is sort of what I am suggesting above we are/should be doing. >>> However I would use 'usefully be used to describe' instead of 'interact >>> with'. >>> >>> ~Richard. >>> >>> >>> On 25/03/2013 21:01, "Antoine Isaac" <aisaac@few.vu.nl> wrote: >>> >>>> Hi all, >>>> >>>> I don't like "instance" either. >>>> Reading the thread I was at one point tempted by "incarnation" but if >>>> you >>>> think it's too "physical" then we're screwed. >>>> Anyway: how about giving us one week to write all the terms on a wiki >>>> page and then give us one week to vote? >>>> I'm afraid otherwise the situation won't really evolve, unless a >>>> miracle >>>> happen and someone finds a perfect term. >>>> >>>> Antoine >>>> >>>> >>>>> I'm not sure where inversion fits in here, so at least in my mind that >>>>> isn't an issue. What I think Richard is aiming at is a generic way to >>>>> link information about creative works, regardless of their level of >>>>> specificity. If that is the case, then "instanceOf" is, IMO, >>>>> inappropriate in a number of ways >>>>> - "instance" implies a single *thing* not a concept or abstraction or >>>>> class, and there can definitely be relationships between creative >>>>> works >>>>> at all of their levels >>>>> - "instance" is being used in BIBFRAME for something much more >>>>> concrete >>>>> and therefore folks will assume (consciously or not) a BIBFRAME >>>>> definition >>>>> - instance seems to have subordination of one thing to another, at >>>>> least in the common usage of the term. However, in many cases there is >>>>> no "primary thing" for another to be an instance of. You just have two >>>>> things that are related. >>>>> >>>>> I'm back to "commonEndeavor" (or something that means approximately >>>>> the >>>>> same thing) at this point, which is broad and vague. However, I am at >>>>> this point unclear on the use case for this relationship, without >>>>> which >>>>> all of this is just theoretical. Here are some cases that I can >>>>> imagine: >>>>> >>>>> - xISBN: the relationship of all of the things in an xISBN cluster to >>>>> each other and/or to the cluster >>>>> - a digital copy of the text and a record in a library database, not >>>>> necessarily the same manifestation but "close enough" >>>>> - a translation of a text >>>>> - a performance of a musical work >>>>> - a reprint of a book >>>>> - a journal article in a journal and the pre- or post-print online >>>>> >>>>> To what extent are these immediate needs for bibliographic data in >>>>> schema.org? I'd say they are not a #1 priority, but that's why I'm >>>>> asking about use cases. >>>>> >>>>> I can't figure out what we would do with WEMI or BIBFRAME since those >>>>> are "directional" -- that is, they have specific relationships with a >>>>> specific order. (I have vague hopes that we never have to model FRBR, >>>>> but that's just me.) >>>>> >>>>> kc >>>>> >>>>> On 3/25/13 7:53 AM, Young,Jeff (OR) wrote: >>>>>> One of the things that may be messing with our heads is the desire to >>>>>> name inverse properties. I think there are plenty of times in natural >>>>>> language where we have a perfectly sensible way to express a >>>>>> relationship in one direction that seems awkward in the inverse. Is >>>>>> that >>>>>> part of the problem? If so, we should keep in mind that inverse >>>>>> relationships are more of a convenience than a necessity in RDF. >>>>>> >>>>>> Jeff >>>>>> >>>>>>> -----Original Message----- >>>>>>> From: Karen Coyle [mailto:kcoyle@kcoyle.net] >>>>>>> Sent: Monday, March 25, 2013 10:45 AM >>>>>>> To: public-schemabibex@w3.org >>>>>>> Subject: Re: InstanceOf/derivativeOf >>>>>>> >>>>>>> Richard, the first part of your message: >>>>>>> >>>>>>> On 3/25/13 3:42 AM, Wallis,Richard wrote: >>>>>>>> In my model I am thinking of several strict definitions of work, >>>>>>>> and >>>>>>>> none >>>>>>>> - that is the way of the broad generic world that Schema is trying >>>>>> to >>>>>>>> serve. >>>>>>>> >>>>>>>> So one set of 'rules' or school of though may say that Story & >>>>>>>> Story-in-English are the same Work others may not. 'We' can define >>>>>>>> what a Work and an Instance are but we are not in the position to >>>>>>>> impose that on the whole web. >>>>>>>> >>>>>>>> In answer to your implied 'how are you defining Work and Instance' >>>>>>>> question - I say 'however you like'. The Schema vocabulary should >>>>>> be >>>>>>>> able to describe both the BIBFRAME and Alan Renear's view equally >>>>>>> well. >>>>>>>> >>>>>>> >>>>>>> defies this second part. If we don't have definitions for Work and >>>>>>> Instance then there is no way to do what you say below -- you can't >>>>>>> align things with other things that are not defined. Yes, it is a >>>>>>> working definition, but without a working definition we have nothing >>>>>> to >>>>>>> propose. And, in fact, every property in schema.org has a >>>>>>> definition. >>>>>>> >>>>>>> The definition there now, and the example, leads to certain >>>>>>> conclusions: >>>>>>> >>>>>>> "Instance/realization/derivation of the concept of this creative >>>>>>> work. >>>>>>> eg. The paperback edition." >>>>>>> >>>>>>> This is very specific, and has within it some strong assumptions >>>>>>> about >>>>>>> CW. Is this what we want? Can we explore how it would be used with >>>>>>> CW >>>>>>> to interact with other models? (btw, "isOneOf" is making a lot of >>>>>> sense >>>>>>> to me now). >>>>>>> >>>>>>> kc >>>>>>> >>>>>>> >>>>>>>> Once we have proposed a generic way to describe relationships >>>>>> between >>>>>>>> things that can be described as Creative works, I believe we have a >>>>>>> place >>>>>>>> to identify good practice in how we wold describe FRBR >>>>>>>> Works/Expressions/Manifestations/Items, BIBFRAME Works/Instances, >>>>>> and >>>>>>>> other well used domain specific entities using this generic >>>>>>> vocabulary. >>>>>>>> >>>>>>>> ~Richard. >>>>>>>> >>>>>>>> On 25/03/2013 01:14, "Karen Coyle" <kcoyle@kcoyle.net> wrote: >>>>>>>> >>>>>>>>> Richard, regarding your model, I think it depends much on how Work >>>>>>> is >>>>>>>>> defined. If Work is defined the way it is in BIBFRAME, then: >>>>>>>>> >>>>>>>>> Story >>>>>>>>> Story in English >>>>>>>>> >>>>>>>>> are part of bibframe:Work, and the story in English is not an >>>>>>> instance >>>>>>>>> of the story. Instance comes into use only when the Work comes >>>>>>>>> into >>>>>>>>> being (in the "realization" sense). As I understand it, BIBFRAME >>>>>>>>> separates the abstract from the concrete. [1] So maybe we should >>>>>>> define >>>>>>>>> what we mean by Work and Instance, and then look again at the >>>>>>>>> terms >>>>>>> we >>>>>>>>> use for them. >>>>>>>>> >>>>>>>>> kc >>>>>>>>> [1] However, if you read Alan Renear's work on FRBR, you may be of >>>>>>> the >>>>>>>>> school that only frbr:Item has physicality, the others are >>>>>>> abstractions. >>>>>>>>> Neither FRBR nor BIBFRAME feel entirely satisfactory, I must say, >>>>>>> but do >>>>>>>>> I have something better? Nope. >>>>>>>>> >>>>>>>>> On 3/24/13 5:34 PM, Wallis,Richard wrote: >>>>>>>>>> I am not a massive fan of instanceOf and hasInstance either. >>>>>>>>>> >>>>>>>>>> But applying my test to creativeInstanceOf we get: >>>>>>>>>> >>>>>>>>>> * Story-in-English is a creativeInstanceOf Story - That >>>>>> sort >>>>>>> of >>>>>>>>>> works >>>>>>>>>> * Story-in-book-in-library is a creativeInstanceOf >>>>>>>>>> Story-in-pbk-book - That doesn't really work. Just stocking in a >>>>>>>>>> library is not really a creative act. >>>>>>>>>> >>>>>>>>>> The works themselves are creative, not the relationships between >>>>>>> them. >>>>>>>>>> >>>>>>>>>> ~Richard. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> From: "Young,Jeff (OR)" <jyoung@oclc.org<mailto:jyoung@oclc.org>> >>>>>>>>>> Date: Sun, 24 Mar 2013 19:47:58 -0400 >>>>>>>>>> To: Richard Wallis >>>>>>>>>> <richard.wallis@oclc.org<mailto:richard.wallis@oclc.org>> >>>>>>>>>> Cc: Antoine Isaac <aisaac@few.vu.nl<mailto:aisaac@few.vu.nl>>, >>>>>>>>>> <public-schemabibex@w3.org<mailto:public-schemabibex@w3.org>> >>>>>>>>>> Subject: Re: InstanceOf/derivativeOf >>>>>>>>>> >>>>>>>>>> I could have beec clearer, but"isRecordOf" was intended as a >>>>>>>>>> joke. >>>>>>>>>> >>>>>>>>>> Regarding, "isInstanceOf", I'm reminded that GoodRelations has >>>>>>>>>> gr:Individual, which is disorienting for reasons similar to >>>>>>> "instance". >>>>>>>>>> When GoodRelations integrated with Schema.org<http://Schema.org>, >>>>>>> this >>>>>>>>>> got translated to schema:IndividualProduct, which is less >>>>>>> offensive. >>>>>>>>>> Perhaps we should consider a similar hair split in this case with >>>>>>>>>> schemap:creativeInstanceOf. >>>>>>>>>> >>>>>>>>>> I have to say I absolutely hate instanceOf. >>>>>>>>>> >>>>>>>>>> Jeff >>>>>>>>>> >>>>>>>>>> Sent from my iPad >>>>>>>>>> >>>>>>>>>> On Mar 24, 2013, at 7:28 PM, "Wallis,Richard" >>>>>>>>>> <Richard.Wallis@oclc.org<mailto:Richard.Wallis@oclc.org>> wrote: >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> My formatting got screwed by the email system, so I attach a >>>>>>> screenshot >>>>>>>>>> of >>>>>>>>>> what I intended. >>>>>>>>>> >>>>>>>>>> ~Richard. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> On 24/03/2013 23:14, "Richard Wallis" <richard.wallis@oclc.org> >>>>>>> wrote: >>>>>>>>>> >>>>>>>>>>> I tend to hold the same suspicions as Antoine as to the content >>>>>> of >>>>>>>>>>> those >>>>>>>>>>> 'few drinks'. I believe your wife was nearer with oneOf. >>>>>> However, >>>>>>> I'm >>>>>>>>>>> not >>>>>>>>>>> sure either convey the meaning of the generic relationship we >>>>>>>>>>> are >>>>>>>>>>> trying >>>>>>>>>>> to achieve. >>>>>>>>>>> >>>>>>>>>>> Personally the test I apply to these is to imagine a set of 3 or >>>>>>> more >>>>>>>>>>> CreativeWorks thus: >>>>>>>>>>> >>>>>>>>>>>> hasInstance >hasInstance >hasInstance >>>>>>>>>>>> hasInstance >>>>>>>>>>> / \ / \ / \ >>>>>>>>>>> / \ >>>>>>>>>>> Story Story-in-English Story-in-Book >>>>>>>>>>> Story-in-pbk-book story-in-book-in-library >>>>>>>>>>> \ / \ / \ / >>>>>>>>>>> \ / >>>>>>>>>>> isInstanceOf< isInstanceOf< isInstanceOf< >>>>>>>>>>> isInstanceOf< >>>>>>>>>>> >>>>>>>>>>> I know this is stretching it a bit, but doing this tends to >>>>>>> highlight >>>>>>>>>>> where focussing in on individual use-cases hides where things >>>>>>>>>>> are >>>>>>> not >>>>>>>>>>> appropriate elsewhere. In the above example I believe 'instance' >>>>>>>>>>> works as >>>>>>>>>>> a broad compromise, where as 'record', 'derivation', >>>>>> 'expression', >>>>>>>>>>> 'realisation', and others seem to possibly work better in one >>>>>> area >>>>>>> but >>>>>>>>>>> much worse in others. >>>>>>>>>>> >>>>>>>>>>> ~Richard. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On 24/03/2013 12:25, "Young,Jeff (OR)" <jyoung@oclc.org> wrote: >>>>>>>>>>> >>>>>>>>>>>> The thing I like about UNIMARC Authorities is that they have >>>>>>>>>>>> the >>>>>>>>>>>> notion >>>>>>>>>>>> of a "primary entity" which is the thing the record represents. >>>>>>> If you >>>>>>>>>>>> look in the same places in MARC21 Authorities you'll find a >>>>>>> tautology. >>>>>>>>>>>> :-/ >>>>>>>>>>>> >>>>>>>>>>>> Sent from my iPad >>>>>>>>>>>> >>>>>>>>>>>> On Mar 24, 2013, at 7:58 AM, "Antoine Isaac" <aisaac@few.vu.nl> >>>>>>> wrote: >>>>>>>>>>>> >>>>>>>>>>>>> Not sure I prefer these ones... >>>>>>>>>>>>> >>>>>>>>>>>>> PS: "record", really? Did your glasses contain MARC brandy? >>>>>>>>>>>>> ;-) >>>>>>>>>>>>> (https://fr.wikipedia.org/wiki/Marc_%28eau-de-vie%29) >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>>> I described the general situation to my wife and she >>>>>>>>>>>>>> suggested >>>>>>> the >>>>>>>>>>>>>> alternative: "oneOf". Hmm. >>>>>>>>>>>>>> >>>>>>>>>>>>>> After a few more drinks, we finally agreed on "isRecordOf". >>>>>> ;-) >>>>>>>>>>>>>> >>>>>>>>>>>>>> Sent from my iPad >>>>>>>>>>>>>> >>>>>>>>>>>>>> On Mar 22, 2013, at 8:26 AM, >>>>>>>>>>>>>> >>>>>>> >>>>>>> "Wallis,Richard"<Richard.Wallis@oclc.org<mailto:Richard.Wallis@oclc.o >>>>>>>>>>>>>> rg>> >>>>>>>>>>>>>> wrote: >>>>>>>>>>>>>> >>>>>>>>>>>>>>> I have renamed the Work-Instance proposal to a more generic >>>>>>>>>>>>>>> CreativeWork >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>> Relationships<http://www.w3.org/community/schemabibex/wiki/CreativeW >>>>>>>>>>>>>>> or >>>>>>>>>>>>>>> k >>>>>>>>>>>>>>> _Relationships> to remove the associations with those words >>>>>> in >>>>>>>>>>>>>>> FRBR, >>>>>>>>>>>>>>> BIBFRAME etc. >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> In yesterday's meeting we suggested that instanceOf& >>>>>>> hasInstance >>>>>>>>>>>>>>> should be renamed to derivativeOf& hasDerivative. However >>>>>>>>>>>>>>> discussion >>>>>>>>>>>>>>> on list has moved away from that idea so I have left it as >>>>>>>>>>>>>>> is >>>>>>> for >>>>>>>>>>>>>>> the >>>>>>>>>>>>>>> moment. >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> I suggest we try some more examples and look at the wording. >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> I think we have general agreement about the need for these >>>>>>>>>>>>>>> properties. It is the names we need to settle, and >>>>>> appropriate >>>>>>>>>>>>>>> examples to test them against and use for explanation in the >>>>>>>>>>>>>>> proposal. >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> ~Richard >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>> >>>>>>>>> -- >>>>>>>>> Karen Coyle >>>>>>>>> kcoyle@kcoyle.net http://kcoyle.net >>>>>>>>> ph: 1-510-540-7596 >>>>>>>>> m: 1-510-435-8234 >>>>>>>>> skype: kcoylenet >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> >>>>>>> -- >>>>>>> Karen Coyle >>>>>>> kcoyle@kcoyle.net http://kcoyle.net >>>>>>> ph: 1-510-540-7596 >>>>>>> m: 1-510-435-8234 >>>>>>> skype: kcoylenet >>>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>> >>>> >>>> >>>> >>> >>> >>> >>> >> >> -- >> Karen Coyle >> kcoyle@kcoyle.net http://kcoyle.net >> ph: 1-510-540-7596 >> m: 1-510-435-8234 >> skype: kcoylenet >> >> > > > > -- Karen Coyle kcoyle@kcoyle.net http://kcoyle.net ph: 1-510-540-7596 m: 1-510-435-8234 skype: kcoylenet
Received on Wednesday, 27 March 2013 02:00:26 UTC