- From: Young,Jeff (OR) <jyoung@oclc.org>
- Date: Sun, 15 Mar 2015 22:23:39 +0000
- To: Antoine Isaac <aisaac@few.vu.nl>
- CC: "public-schemabibex@w3.org" <public-schemabibex@w3.org>
Antoine, The daffodils/deprivation was just a lark that someone challenged me with. The more useful examples are where Schema:Role can be combined with WikiData URI to account for countless varieties of creator/contributor roles like bassoonist, trumpeter, and others already mentioned. It would be interesting to see how many cases in the ending credits of a movie could be accounted for that way. For example: http://www.wikidata.org/entity/Q2962047 Jeff > On Mar 15, 2015, at 4:55 PM, Antoine Isaac <aisaac@few.vu.nl> wrote: > > Hi Jeff, > > Belated reply... The mail and example confused me first. Upon returning to it today I see a bit better, but am still a bit lukewarm on having something that coul be a language of its own... Even though I also realize the role trick is a bit akin as what is implied by allowing usage of properties like dc:creator with literal or URIs. (a resource is rarely created by a literal of course, so the semantics of the property differ a bit depending on the usage). > > Cheers, > > Antoine > >> On 3/12/15 10:45 PM, Young,Jeff (OR) wrote: >> It was weird to me at first too and I'm sure it will get abused on the open web. OTOH, it transforms some extremely thorny problems into a simple mechanical indirection. >> >> I don't think in terms of "model" so much anymore. Schema.org feels more like a language now. Here's an example: >> >> https://dl.dropboxusercontent.com/u/68401641/examples/daffodils_and_deprivation.ttl >> >> Jeff >> >>> -----Original Message----- >>> From: Antoine Isaac [mailto:aisaac@few.vu.nl] >>> Sent: Thursday, March 12, 2015 5:26 PM >>> To: public-schemabibex@w3.org >>> Subject: Re: [schemaorg] Vocabulary for comics (#378) >>> >>> Hi Jeff, >>> >>> Thanks for the explanation! >>> I guess I'm still too biased by RDF modeling tradition... >>> I'm willing to accept that schema:member could be used to relate two >>> persons in one dataset, and to link a person to a role in another. >>> But it makes me shiver a bit when I see this sort of semantic difference in >>> two closely related triples. I.e schema:member between a person and a role >>> in one triple, and schema:member between a role and a person in the next >>> one! >>> >>> Antoine >>> >>>> On 3/12/15 10:17 PM, Young,Jeff (OR) wrote: >>>> Antoine, >>>> >>>> As I understand it, this should be understood as "A4 is a member of A1.... >>> and oh, by the way, I snuck in this Role node so I can say something more >>> about the 'member' relationship ". >>>> >>>> Here's the description of schema:Role: >>>> >>>> "Represents additional information about a relationship or property. For >>> example a Role can be used to say that a 'member' role linking some >>> SportsTeam to a player occurred during a particular time period. Or that a >>> Person's 'actor' role in a Movie was for some particular characterName. Such >>> properties can be attached to a Role entity, which is then associated with the >>> main entities using ordinary properties like 'member' or 'actor'." >>>> >>>> It's like a get-out-of-triple-jail-free card without having to jail break RDF >>> parsers. >>>> >>>> Jeff >>>> >>>>> -----Original Message----- >>>>> From: Antoine Isaac [mailto:aisaac@few.vu.nl] >>>>> Sent: Thursday, March 12, 2015 5:09 PM >>>>> To: public-schemabibex@w3.org >>>>> Subject: Re: [schemaorg] Vocabulary for comics (#378) >>>>> >>>>> Hi Dan, everyone, >>>>> >>>>> I wonder how the comics ontology relates to some work on manga seem >>>>> in the past >>>>> http://dx.doi.org/10.1145/2467696.2467731 >>>>> http://dcpapers.dublincore.org/pubs/article/download/979/952 >>>>> >>>>> About role, I have a perhaps stupid question: does anyone remembers >>>>> why schema.org has a "double use" of the property that relates the >>>>> role to the two resources it relates? >>>>> >>>>> Reading Jeff's example: >>>>> >>>>> _:A1 schema:member [ >>>>> a schema:Role; >>>>> schema:roleName entity:Q5371902; # harpsichordist >>>>> schema:member _:A4; >>>>> ]; >>>>> >>>>> This could be understood as "the agent A4 is a member of a role that >>>>> is a member of a group." >>>>> >>>>> If the group had had a "guest" harpsichordist for one concert, then >>>>> we would have >>>>> >>>>> _:A1 schema:member [ >>>>> a schema:Role; >>>>> schema:roleName entity:Q5371902; # harpsichordist >>>>> schema:member _:A4; >>>>> ]; >>>>> _:A1 schema:guest [ >>>>> a schema:Role; >>>>> schema:roleName entity:Q5371902; # harpsichordist >>>>> schema:guest _:A5; >>>>> ]; >>>>> >>>>> Is there interest in having to adapt the pattern in two places, as >>>>> opposed to have a same property (say, "rolePlayer") for every link >>>>> between a role to the entity that plays it? >>>>> >>>>> Thanks, >>>>> >>>>> Antoine >>>>> >>>>>> On 3/12/15 8:51 PM, Wallis,Richard wrote: >>>>>> Yes Dan, I agree - I was over complicating things by inventing the >>>>>> need for a >>>>> Role subType she Role on its own would be sufficient. >>>>>> >>>>>> Good example Jeff. >>>>>> >>>>>> ~Richard >>>>>> >>>>>>> On 12 Mar 2015, at 17:39, Young,Jeff (OR) <jyoung@oclc.org >>>>>> <mailto:jyoung@oclc.org>> wrote: >>>>>> >>>>>>> I agree that coordination on roleNames (especially using URIs) >>>>>>> would be a >>>>> great. >>>>>>> Here's a mockup I did recently to account for the instruments that >>>>> individual musicians played on a music album. It was while I was >>>>> mocking this up that I realized how many were covered by WikiData: >>> https://dl.dropboxusercontent.com/u/68401641/Devon/TextExtraction.ttl >>>>>>> Jeff >>>>>>> *From:*Dan Scott [mailto:denials@gmail.com] *Sent:*Thursday, >>> March >>>>>>> 12, 2015 1:33 PM *To:*Wallis,Richard; Sean Petiya *Cc:*Young,Jeff >>>>>>> (OR); public-schemabibex@w3.org <mailto:public- >>> schemabibex@w3.org> >>>>>>> *Subject:*Re: [schemaorg] Vocabulary for comics (#378) On Thu, 12 >>>>>>> Mar >>>>>>> 2015 at 12:42 Wallis,Richard <Richard.Wallis@oclc.org >>>>> <mailto:Richard.Wallis@oclc.org>> wrote: >>>>>>> >>>>>>> Hi Sean, >>>>>>> My personal opinion is that the work you and the previously >>>>>>> referenced >>>>> draft on the Wiki >>> <http://www.w3.org/community/schemabibex/wiki/Periodicals_and_Comic >>>>> s_synthesis> are within the scope of this group to discuss. >>>>>>> As Jeff indicated, there is some overlap and/or mismatch >>>>>>> between your >>>>> discussions of Role and similar concepts from the Library of Congress >>>>> Relator Codes and WikiData. How these terms are defined/referenced >>>>> in the vocabulary is then a question. I am always sceptical of >>>>> statements such as "set that covers the major...", because it is very >>>>> difficult to a)get agreement on what is major and b) what do you do about >>> defying the minor ones. >>>>>>> Your use of the term name 'role' conflicts with the Role >>>>> <http://schema.org/Role> type inSchma.org <http://schma.org/>, which >>>>> in itself is not a problem (you could use creativeRole for example). >>>>> However in covering off this need, I think it would be worth >>>>> considering the creation of a ContributionRole subtype of Role which >>>>> would allow the qualification of the contributor relationship between >>> CreativeWork and Person or Organization. >>>>> Then using the roleName attribute the type of contribution could be >>>>> qualified either by a URL to the Library of Congress Relators, or WikiData, >>> etc. >>>>> definitions, or, if not available, in plain text. >>>>>>> >>>>>>> Erm. I thought the agreed-upon pattern for using Role (first >>>>>>> proposed by >>>>> danbri athttp://lists.w3.org/Archives/Public/public- >>>>> vocabs/2014Sep/0009.html) would be to apply the external vocabulary >>>>> property in combination with schema:contributor (e.g. lcrel:clr) and >>>>> apply schema:roleName for those consumers that might, for whatever >>>>> reason, want to limit themselves to justschema.org >>> <http://schema.org/>. E.g.: >>>>>>> <dl vocab="http://schema.org/" >>>>> prefix="lcrel:http://id.loc.gov/vocabulary/relators/"> >>>>>>> <span property="contributor" typeof="Role"> >>>>>>> <dt><span property="roleName">Pencils</span>:</dt> >>>>>>> <dd><span property="contributor">Ron Lim</span></dd> >>>>>>> </span> >>>>>>> <span property="contributor" typeof="Role"> >>>>>>> <dt><meta property="roleName" content="colorist">Colors:</dt> >>>>>>> <dd><span property="contributor lcrel:clr">Chris >>>>> Sotomayor</span></dd> >>>>>>> </span> >>>>>>> </dl> >>>>>>> ... which generates something like: >>>>>>> ns1:contributor [ a ns1:Role ; >>>>>>> ns1:contributor "Ron Lim" ; >>>>>>> ns1:roleName "Pencils" ], >>>>>>> [ a ns1:Role ; >>>>>>> lcrel:clr "Chris Sotomayor" ; >>>>>>> ns1:contributor "Chris Sotomayor" ; >>>>>>> ns1:roleName "colorist" ]; This was the direction I >>>>>>> was taking things with my preconference at SWIB, which even >>>>>>> includes a Comic example: >>> https://coffeecode.net/swib14/preconference/rdfa_exercises/6_comic_bo >>>>>>> ok/ We could certainly update guidance and examples to use >>>>>>> contributor types from wikidata and other vocabularies, but I would >>>>>>> like to ensure we're starting from a common understanding. And >>>>>>> having put a fair amount of effort into the last iteration of >>>>>>> Periodicals & Comics, I have some interest in Comics going forward >>>>>>> :) >
Received on Sunday, 15 March 2015 22:24:09 UTC