- From: ZENG, MARCIA <mzeng@kent.edu>
- Date: Sun, 10 Nov 2013 00:14:08 +0000
- To: Johan De Smedt <johan.de-smedt@tenforce.com>
- CC: "ZENG, MARCIA" <mzeng@kent.edu>, "vladimir.alexiev@ontotext.com" <vladimir.alexiev@ontotext.com>, "public-esw-thes@w3.org" <public-esw-thes@w3.org>, Stella Dextre Clarke <stella@lukehouse.org>, Joan Cobb <JCobb@getty.edu>
Johan, Notations for AAT might not be so straight forward since it does not have a notation system. MeSH does. Let's think further and discuss more later. Marcia Sent from my iPad > On Nov 9, 2013, at 5:44 PM, "Johan De Smedt" <johan.de-smedt@tenforce.com> wrote: > > Hi Marcia > > Thanks a lot for the explanation! > > I am wondering if and how we should advice on ordering. > - notation is foreseen on both Thesaurus Concept, on Thesaurus Array and on Thesaurus Group. > - notation is translated to skos:notation > So that can do the job if it is properly encoded (typically with a convention or with custom data types). > - The negative remark that could be made is that it is not a RDF or OWL only typing. > So it requires Specific know-haw about the notation typing. > (Compared to e.g. the rdf:List) > My personnel evaluation of this criticism would be that it an application knowing to make mixed orders of concepts and arrays needs to understand (distinguish and order) notation types. > > So the ISO 25964 compliant answer to Vladimir would be (my understanding) to use notation/skos:notation and document the used typing and ordering. > A TMS system should be configurable so > - the ordering properties can be defined (e.g. language specific pref-label, or notation, or rdfs:label or a combination with priorities) > - the ordering rules can be applied using the specified ordering (alphabetic, notation conversion + ordering, numeric, date-time, ...). > > Thanks to let me know if my understanding is correct. > > Kind Regards, > > Johan De Smedt >> -----Original Message----- >> From: ZENG, MARCIA [mailto:mzeng@kent.edu] >> Sent: Saturday, 09 November, 2013 15:31 >> To: Johan De Smedt; vladimir.alexiev@ontotext.com; public-esw-thes@w3.org >> Cc: 'Stella Dextre Clarke'; Joan Cobb >> Subject: RE: how to: ordered collection of a Concept >> >> Hi, all, >> I may be answering out of the bigger context here. Just my two cents. >> >> 1. Orders of children issue is actually the issue of orders of siblings (the related concepts at the same >> level). We have raised this issue before to SKOS. >> >> Orders of siblings is necessary in a hierarchical classification system. Usually a notation scheme takes >> care of it. >> Any notation has both a semantic value and an ordinal value: >> The semantic value of a classification number is the subject or concept it stands for, e.g., 610 Medicine >> (DDC). >> The ordinal value of a number of code places the subject into its determined rank in the scheme. >> There are about 10 common ways used to decide the orders of siblings (coordinate classes). >> >> In thesaurus' multilevel formats, the siblings are often only displayed by viewing the broader term. The >> order of the siblings are not as critical as that in a classification or taxonomy. >> However in some thesauri, the hierarchies are much deeper than others and the whole thesaurus is >> highly-structured, such as AAT and MeSH. >> >> 2. A related situation is the not-trully-poly-hierarchical cases. In AAT, one can see a concept always has >> a preferred parent, while in many cases also an 'additional parent'. So, it is not simply indicating one >> concept has two parents. It has to indicate which parent is preferred. >> See example in AAT: ID: 300265026 loutrophoros-hydriae. >> In TGN: ID: 7010273 Saint Petersburg (inhabited place). >> (The 'preferred' situation of parents is all coded clearly in the Getty vocabularies. ) >> >> Marcia >> ________________________________________ >> From: Johan De Smedt [johan.de-smedt@tenforce.com] >> Sent: Saturday, November 09, 2013 4:50 AM >> To: vladimir.alexiev@ontotext.com; public-esw-thes@w3.org >> Cc: 'Stella Dextre Clarke'; ZENG, MARCIA >> Subject: RE: how to: ordered collection of a Concept >> >> Hello Vladimir, >> >> Can you give an example illustrating the problem and approach you make with equivalentArray? >> >> Thanks, >> >> Kind Regards, >> >> Johan De Smedt >>> -----Original Message----- >>> From: Vladimir Alexiev [mailto:vladimir.alexiev@ontotext.com] >>> Sent: Friday, 08 November, 2013 16:43 >>> To: public-esw-thes@w3.org >>> Cc: 'Stella Dextre Clarke'; 'Marcia Zeng' >>> Subject: how to: ordered collection of a Concept >>> >>> Some AAT *concepts* have ordered children (narrower concepts). >>> How can one represent this? >>> - I guess one could put the children in a skos:OrderedCollection that is "free floating" i.e. one that¡¯s >> not >>> connected to anything. But TMSes will have a hard time figuring what this pattern means >>> - iso:ConceptArray allows you to put ordered children *under* a concept by using >>> iso:subordinateArray >>> - I propose an extension iso:equivalentArray that puts the array "next" to the concept. >>> >>> What do you think? > > >
Received on Sunday, 10 November 2013 00:14:39 UTC