ontolog-forum
[Top] [All Lists]

## Re: [ontolog-forum] Curation view of Ontologies

 To: "[ontolog-forum]" "John F. Sowa" Tue, 24 Jun 2008 14:03:16 -0400 <486136E4.3010509@xxxxxxxxxxx>
 ```I received an offline note about Ranganathan's colon classification and the reasons why it has not been widely used. I recommended Common Logic as a simpler system, but I also realize that many people consider logic to be "difficult". But whenever anybody tries to develop a "simpler" system, they usually create something that is much harder to use than the so-called "difficult" logic.    (01) To illustrate the principles, I'll use some examples from the slides that I presented at the tutorial on Common Logic at the SemTech'08 conference in San Jose:    (02) http://www.jfsowa.com/talks/cl_sowa.pdf Common Logic: A Framework for a Family of Logic-Based Languages    (03) The first slide has the title "How to say 'A cat is on a mat.'" It shows Frege's original Begriffsschrift, Peirce's algebraic notation, and Peano's variation of Peirce's notation, which has become the modern predicate calculus.    (04) Since people complained that logic was too "hard", the gang that gave us SQL decided to "simplify" the notation. My second slide begins with that so-called "simplification":    (05) SELECT FIRST.ID, SECOND.ID FROM OBJECTS FIRST, OBJECTS SECOND, SUPPORTS WHERE FIRST.TYPE = "Cat" AND SECOND.TYPE = "Mat" AND SUPPORTS.SUPPORTER = SECOND.ID AND SUPPORTS.SUPPORTEE = FIRST.ID    (06) The SELECT line indicates that the ID columns of two tables are being selected. The FROM line says that the first comes from a table named OBJECTS, the second comes from the same OBJECTS table, and the SUPPORTS table is also being used.    (07) The four lines beginning with WHERE specify a statement in a version of first-order logic that specifies how the two IDs are related: The type of the first is "Cat", the type of the second is "Mat", the supporter of the SUPPORTS table is the second ID, and the supportee of the SUPPORTS table is the first ID.    (08) For comparison, following is a statement in the Common Logic Interchange Format (CLIF):    (09) (exists ((x Cat) (y Mat)) (On x y))    (010) This sentence says that there exists an x of type Cat and a y of type Mat where x is on y. For comparison with the SQL statement, the relation (On x y) could be written (Supports y x). The ID and TYPE, which name columns of the OBJECT table in SQL, are represented in CLIF by variables (ID) and restrictions on the variables (TYPE).    (011) A logically equivalent notation is the Conceptual Graph Interchange Format (CGIF):    (012) [Cat *x] [Mat *y] (On ?x ?y)    (013) This statement is shorter because the quantifier "there exists" is the default. The first concept [Cat *x] says "There is a cat x", the concept [Mat *y] says "There is a mat y", and the relation (On ?x ?y) says "x is on y".    (014) At the bottom of that slide is an example from an English-like notation called Common Logic Controlled English (CLCE):    (015) A cat is on a mat.    (016) This looks like English, but it has a formally defined mapping to either CLIF or CGIF notation.    (017) Beginning with slide #15 is an example of medical English, which was used in a conference on medical informations as an exercise for comparing several different notations. What follows is the mapping to CLCE, and from CLCE to both CLIF and CGIF.    (018) The process of doing the mapping sounds complicated, and it takes several slides to explain the details. Finally, slide #21 shows the complete mapping to CLCE, slide #22 shows the CLIF equivalent, and slide #23 shows the CGIF equivalent.    (019) Notice that there are many details, even in the CLCE statement, that require some explanation. However, the person who mapped the same English to RDF and OWL said that he wasn't sure whether he could even do it, but by defining a lot of intermediate forms, he was finally able to express it in OWL. But it wouldn't fit on one slide. He didn't attempt to explain it, but he did scroll through page after page after page of angle brackets.    (020) The irony is that RDF originated from an attempt by R. V. Guha (the former associate director of the Cyc project) to develop a simplified notation that would be easier to use than the version of logic used in Cyc. In the late 1990s, Guha worked with Tim Bray (one of the people who collaborated on the development of XML) to represent simple "triples" in XML. That became RDF.    (021) Unfortunately, the notation Tim Bray defined, which is still used today, was vastly more complex than it should have been. Tim admitted and apologized for his mistake:    (022) http://www.tbray.org/ongoing/When/200x/2003/05/21/RDFNet    (023) As Tim says, "It's the syntax, stupid!", and he adds:    (024) Speaking only for myself, I have never actually managed to write down a chunk of RDF/XML correctly, even when I had the triples laid out quite clearly in my head. Furthermore -- once again speaking for myself -- I find most existing RDF/XML entirely unreadable. And I think I understand the theory reasonably well.    (025) The basic idea of RDF was to represent triples of Resource, Property, Value. Any such triple can be represented in CLIF or CGIF as    (026) (Property Resource Value)    (027) All three URIs as used in RDF are valid name sequences in CLIF or CGIF. Any or all of the other information expressed in RDF can be represented in both CLIF and CGIF, either in structured comments or in additional relations that relate the triples to each other or to other resources.    (028) For anyone who wants to use the XML tools, the ISO standard for Common Logic also defines XCL as a third dialect that can be mapped directly to or from CLIF and CGIF. Furthermore, XCL is far more compact than RDF, and it supports a cleaner, simpler, more expressive, and more efficient logic than the combination of RDF(S), OWL, RuleML, and SPARQL.    (029) Finally, CLIF, CGIF, and XCL support full n-tuples. Therefore, any SQL table can be directly downloaded to or uploaded from Common Logic in any of those three notations. Furthermore, CL in any of those notations can be used to write rules to specify whatever translations are needed to relate any or all of the above notations to one another    (030) For further references, see slide 28 of my tutorial.    (031) John Sowa    (032) _________________________________________________________________ Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/ Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/ Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx Shared Files: http://ontolog.cim3.net/file/ Community Wiki: http://ontolog.cim3.net/wiki/ To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (033) ```
 Current Thread Re: [ontolog-forum] Curation view of Ontologies, (continued) Re: [ontolog-forum] Curation view of Ontologies, Biswanath Dutta Re: [ontolog-forum] Curation view of Ontologies, paola . dimaio Re: [ontolog-forum] Curation view of Ontologies, Biswanath Dutta Re: [ontolog-forum] Curation view of Ontologies, paola . dimaio Re: [ontolog-forum] Curation view of Ontologies, Biswanath Dutta Re: [ontolog-forum] Curation view of Ontologies, Pat Hayes [ontolog-forum] Naming and Semantics of Identity Management for SOA/Web Services, Gary Berg-Cross Re: [ontolog-forum] Naming and Semantics of Identity Management for SOA/Web Services, Duane Nickull Re: [ontolog-forum] Naming and Semantics of Identity Management for SOA/Web Services, Barker, Sean (UK) Re: [ontolog-forum] Curation view of Ontologies, John F. Sowa Re: [ontolog-forum] Curation view of Ontologies, John F. Sowa <= Re: [ontolog-forum] Curation view of Ontologies, ravi sharma