We are seeking advice in the context of using semantic mediawiki to document terms, at present as an example the taxpub vocabularly, which ontology to use to document the following: 1. generic part-of relations of concepts. Example: the taxpub nomenclature section is part of the taxpub treatment We consider using DublinCore for this. We found nothing in SKOS for this. 2. part of relations of things both petal and sepal are part of the corolla. Bob Morris recommends the ro ontology: http://obofoundry.org/ro/ However, this one says that it is undergoing strong changes in the near future. 3. When documenting legacy xml ontologies, the xml variant requires documenting a difference between element and attribute children. How to express this in RDF form? ("why would you want do that" -> Bob Morris likes to answer that, but in fact we just want to have an orthogonal form to simplify things). Any help or insight is appreciated. Gregor -- --------------------------------- Dr. G. Hagedorn +49-(0)30-8304 2220 (work) +49-(0)30-831 5785 (private) http://www.linkedin.com/in/gregorhagedorn This communication (including all attachments) is sent on a personal basis. It is intended only for the person(s) to whom it is addressed. Redistributing or publishing it without permission is a violation of privacy rights and copyright.
Yes, but... On Wed, Sep 21, 2011 at 5:11 PM, Gregor Hagedorn <g.m.hagedorn@gmail.com> wrote:
We are seeking advice in the context of using semantic mediawiki to document terms, at present as an example the taxpub vocabularly, which ontology to use to document the following:
1. generic part-of relations of concepts. Example: the taxpub nomenclature section is part of the taxpub treatment We consider using DublinCore for this. We found nothing in SKOS for this.
2. part of relations of things both petal and sepal are part of the corolla.
Bob Morris recommends the ro ontology: http://obofoundry.org/ro/ However, this one says that it is undergoing strong changes in the near future.
Well, I don't know what "near" means for their future, but the discussion of it petered out over 15 months ago. A very brief glance at that indicates that the initial impetus was mainly to refactor RO, although the discussion seems to raise other points also. Hilmar might be able to through some insight on the matter.
3. When documenting legacy xml ontologies, the xml variant requires documenting a difference between element and attribute children.
How to express this in RDF form? ("why would you want do that" -> Bob Morris likes to answer that, but in fact we just want to have an orthogonal form to simplify things).
Ah, you slightly mis-state my underlying point. I (hope I am) famous for often saying that questions that begin with "Why would anybody ever want to") are forbidden to software engineers, and by extension to ontology engineers. But (when prompted) I always point out that this is a mantra to guard against two things, depending on the state of development: (a) When bleated, sheeplike, by programmers whose code crashed gracelessly by users who do something unexpected, it almost always signals that the code contains an implicit, unexamined, assumption that a certain use case would never appear. (b) Before development it often signals that insufficient use-case modelling has not been done. So in this case, I literally mean that one needs a clear requirement for modelling the syntax of XML in a semantic modeling language. Phrased that way, I would worry that the impedance mismatches between a set of XML constraints on a document's structure and a set of RDF descriptions of the content of the document might be very high.
Any help or insight is appreciated.
Gregor
-- --------------------------------- Dr. G. Hagedorn +49-(0)30-8304 2220 (work) +49-(0)30-831 5785 (private) http://www.linkedin.com/in/gregorhagedorn
This communication (including all attachments) is sent on a personal basis. It is intended only for the person(s) to whom it is addressed. Redistributing or publishing it without permission is a violation of privacy rights and copyright.
-- Robert A. Morris Emeritus Professor of Computer Science UMASS-Boston 100 Morrissey Blvd Boston, MA 02125-3390 IT Staff Filtered Push Project Department of Organismal and Evolutionary Biology Harvard University email: morris.bob@gmail.com web: http://efg.cs.umb.edu/ web: http://etaxonomy.org/mw/FilteredPush http://www.cs.umb.edu/~ram phone (+1) 857 222 7992 (mobile)
On Sep 21, 2011, at 5:11 PM, Gregor Hagedorn wrote:
1. generic part-of relations of concepts. Example: the taxpub nomenclature section is part of the taxpub treatment We consider using DublinCore for this. We found nothing in SKOS for this.
I would use DC partof / haspart for this. This, I think, will make you most likely to use the same vocabulary and properties as many others who state relationships between publications, their parts, and things connected to them (such as, for example in the case of Dryad, data packages and their parts).
2. part of relations of things: both petal and sepal are part of the corolla.
I would use RO. This is a well-defined relationship, and will make you use the same as many others who express relationships between anatomical parts, or organisms and their parts.
Bob Morris recommends the ro ontology: http://obofoundry.org/ro/ However, this one says that it is undergoing strong changes in the near future.
I wouldn't worry about this (but see below). The plan is to roll (most of) the relationships currently in RO into BFO (Basic Formal Ontology). They'll continue to exist, but will then receive different URIs. Since there are lots and lots of ontologies and projects that use RO (for example, using it is one of the OBO Foundry criteria for membership), there's already a requirement to make this reasonably painless for users, which is perhaps part of the reason that it still hasn't happened yet. The one caveat is that once these are in BFO and you apply them as BFO properties, by doing so you subscribe to the BFO worldview for semantic purposes, strictly speaking. For most practical purposes and applications that's likely of negligible if any consequences, except if you plan to make processes (in the BFO sense) bear qualities (in the BFO sense). And there's something about 3D immaterial extents in BFO that's the subject of controversy, but I forget the details (which says something already).
3. When documenting legacy xml ontologies, the xml variant requires documenting a difference between element and attribute children.
What's an XML ontology? Isn't that an oxymoron? -hilmar -- =========================================================== : Hilmar Lapp -:- Durham, NC -:- informatics.nescent.org : ===========================================================
Several unintend double negatives and other misuses in my reply reverse my meaning and/or cast doubt on my command of English ... :-( "Hilmar might be able to through some insight on the matter." --> [...] to throw some insight" "signals that insufficient use-case modelling has not been done" --"signals that insufficient [...] has been done" On Wed, Sep 21, 2011 at 5:52 PM, Bob Morris <morris.bob@gmail.com> wrote:
Yes, but...
On Wed, Sep 21, 2011 at 5:11 PM, Gregor Hagedorn <g.m.hagedorn@gmail.com> wrote:
We are seeking advice in the context of using semantic mediawiki to document terms, at present as an example the taxpub vocabularly, which ontology to use to document the following:
1. generic part-of relations of concepts. Example: the taxpub nomenclature section is part of the taxpub treatment We consider using DublinCore for this. We found nothing in SKOS for this.
2. part of relations of things both petal and sepal are part of the corolla.
Bob Morris recommends the ro ontology: http://obofoundry.org/ro/ However, this one says that it is undergoing strong changes in the near future.
Well, I don't know what "near" means for their future, but the discussion of it petered out over 15 months ago. A very brief glance at that indicates that the initial impetus was mainly to refactor RO, although the discussion seems to raise other points also. Hilmar might be able to through some insight on the matter.
3. When documenting legacy xml ontologies, the xml variant requires documenting a difference between element and attribute children.
How to express this in RDF form? ("why would you want do that" -> Bob Morris likes to answer that, but in fact we just want to have an orthogonal form to simplify things).
Ah, you slightly mis-state my underlying point. I (hope I am) famous for often saying that questions that begin with "Why would anybody ever want to") are forbidden to software engineers, and by extension to ontology engineers. But (when prompted) I always point out that this is a mantra to guard against two things, depending on the state of development:
(a) When bleated, sheeplike, by programmers whose code crashed gracelessly by users who do something unexpected, it almost always signals that the code contains an implicit, unexamined, assumption that a certain use case would never appear. (b) Before development it often signals that insufficient use-case modelling has not been done. So in this case, I literally mean that one needs a clear requirement for modelling the syntax of XML in a semantic modeling language. Phrased that way, I would worry that the impedance mismatches between a set of XML constraints on a document's structure and a set of RDF descriptions of the content of the document might be very high.
Any help or insight is appreciated.
Gregor
-- --------------------------------- Dr. G. Hagedorn +49-(0)30-8304 2220 (work) +49-(0)30-831 5785 (private) http://www.linkedin.com/in/gregorhagedorn
This communication (including all attachments) is sent on a personal basis. It is intended only for the person(s) to whom it is addressed. Redistributing or publishing it without permission is a violation of privacy rights and copyright.
-- Robert A. Morris
Emeritus Professor of Computer Science UMASS-Boston 100 Morrissey Blvd Boston, MA 02125-3390 IT Staff Filtered Push Project Department of Organismal and Evolutionary Biology Harvard University
email: morris.bob@gmail.com web: http://efg.cs.umb.edu/ web: http://etaxonomy.org/mw/FilteredPush http://www.cs.umb.edu/~ram phone (+1) 857 222 7992 (mobile)
-- Robert A. Morris Emeritus Professor of Computer Science UMASS-Boston 100 Morrissey Blvd Boston, MA 02125-3390 IT Staff Filtered Push Project Department of Organismal and Evolutionary Biology Harvard University email: morris.bob@gmail.com web: http://efg.cs.umb.edu/ web: http://etaxonomy.org/mw/FilteredPush http://www.cs.umb.edu/~ram phone (+1) 857 222 7992 (mobile)
On Sep 21, 2011, at 6:00 PM, Bob Morris wrote:
"signals that insufficient use-case modelling has not been done"
Where I live this is considered fine English :-) -- =========================================================== : Hilmar Lapp -:- Durham, NC -:- informatics.nescent.org : ===========================================================
Thanks to Paul, Hilmar and Bob, the advice is useful to us.
3. When documenting legacy xml ontologies, the xml variant requires documenting a difference between element and attribute children.
What's an XML ontology? Isn't that an oxymoron?
I personally think an xml schema represents an ontology in a generic sense. It has only a subset of functionalities, just like a thesaurus from the 1960 is a useful subset of the functionality of current ontologies. But Owl has only a subset of Owl2, and I suspect this will be a subset of the future. But our task is actually: just read "documenting a publishing dtd or xml on a semantic wiki". The goal is to keep as much orthogonal as possible, whether an RDF vocab or an xml schema is documented. That does not mean that everything has to match. To my limited understanding, both attributes and child elements in xml schema/DTDs are a class-membership relation. Is a membership, say in any kind of UML model a part-of relation? I suspect so. Incorrect? And then we want to express something about the kind of relation, which in RDF requires a reification, which in semantic mediawiki we don't (at least not easily). So I wonder what to do, to keep it simple. The easy way is to define an xml schema class with our own special properties, but we would like to re-use existing vocabularies where possible. Gregor
participants (3)
-
Bob Morris -
Gregor Hagedorn -
Hilmar Lapp