Comments on: Refactoring Bio With Einstein Part 3: Temporal Invariants http://blog.iandavis.com/2006/03/refactoring-bio-with-einstein-part-3-temporal-invariants blog.iandavis.com Wed, 04 Nov 2009 14:19:37 +0000 http://wordpress.org/?v=3.0-alpha hourly 1 By: Chris Mungall http://blog.iandavis.com/2006/03/refactoring-bio-with-einstein-part-3-temporal-invariants/comment-page-1#comment-299 Chris Mungall Thu, 16 Mar 2006 01:30:49 +0000 http://iandavis.com/blog2/?p=918#comment-299 The solutions outlined here seem to be about the best that can be achieved given the constraints, that is the fact that in RDF all relations are two-place. It's still far from simple. There seems to be no way of avoiding the introduction of time slices or bio:Conditions (for example: try to define Mother using intersectionOf without them), then all instance relations must be represented as two triples, transitivity breaks down unless you introduce owl:transitive_over... However, it's also seems clear to me that this is a recipe for trouble for the semantic web. Surely all real-world data that concerns non-trivial applications such as science and electronic health records, or any kind of human activity _must_ take time into account? Which ever hack you make to account for time, it has to propagate through all your ontologies. An ontology that treats the world as time-slices can't interoperate with one that has a standard view of objects and processes. It may be just about workable, but I can't see it being anything other than tremendously complicated. We'll essentially end up with layering 3-place relations on top of RDF in an extremely inelegant way. This is not made clear when people are lured into the semantic web with examples of toy ontologies about pizzas that live floating in some mathematical space untroubled by time. Unless more is done to address these issues (and I commend this article for tackling this) the semantic web will face a huge backlash when people start realising they have to warp their ontologies and refactor their instance data to deal with time in order to represent real entities. Why is there no best practices document on representing instances that vary in time (that is, all real-world instances)? I do find it curious that more people aren't making noise about this problem - I can only conclude that there's a dearth of serious applications using RDF or OWL for instance data. I'm not sure what the solution is. Efforts such as the temporal RDF paper (see above) are welcome, although it's still not satisfactory, and I'm not sure we haven't put all our eggs in one basket with RDF at all. The solutions outlined here seem to be about the best that can be achieved given the constraints, that is the fact that in RDF all relations are two-place. It’s still far from simple. There seems to be no way of avoiding the introduction of time slices or bio:Conditions (for example: try to define Mother using intersectionOf without them), then all instance relations must be represented as two triples, transitivity breaks down unless you introduce owl:transitive_over…

However, it’s also seems clear to me that this is a recipe for trouble for the semantic web. Surely all real-world data that concerns non-trivial applications such as science and electronic health records, or any kind of human activity _must_ take time into account? Which ever hack you make to account for time, it has to propagate through all your ontologies. An ontology that treats the world as time-slices can’t interoperate with one that has a standard view of objects and processes. It may be just about workable, but I can’t see it being anything other than tremendously complicated. We’ll essentially end up with layering 3-place relations on top of RDF in an extremely inelegant way.

This is not made clear when people are lured into the semantic web with examples of toy ontologies about pizzas that live floating in some mathematical space untroubled by time. Unless more is done to address these issues (and I commend this article for tackling this) the semantic web will face a huge backlash when people start realising they have to warp their ontologies and refactor their instance data to deal with time in order to represent real entities. Why is there no best practices document on representing instances that vary in time (that is, all real-world instances)? I do find it curious that more people aren’t making noise about this problem – I can only conclude that there’s a dearth of serious applications using RDF or OWL for instance data.

I’m not sure what the solution is. Efforts such as the temporal RDF paper (see above) are welcome, although it’s still not satisfactory, and I’m not sure we haven’t put all our eggs in one basket with RDF at all.

]]> By: iand http://blog.iandavis.com/2006/03/refactoring-bio-with-einstein-part-3-temporal-invariants/comment-page-1#comment-298 iand Wed, 15 Mar 2006 11:32:33 +0000 http://iandavis.com/blog2/?p=918#comment-298 danbri, it's clear that our traditional view of the nature of human ancestry is changing. As much as I'm trying to be sociologically neutral it's inevitable that I'm going to be influenced by my own world view and experience. But I think that is core to the semantic web effort, in that we accept that there is no one-true-ontology of the world that is true for all people at all times. At best we can reflect a particular consensual view of the world. There's also an element of pragmatism in that we strive to create ontologies that provide useful abstractions of the world. For the domains that Bio is intended it's useful to assume that people only have one mother. It worries me a little that Bio can't be used to completely describe the child in that news article, but then it can't be used to describe cloned humans either. Of course, FOAF isn't immune from these issues :) With the advent of gene therapy and even more established practices such as bone marrow transplants ( http://www.newscientist.com/article.ns?id=mg18825234.600 ), a person can end up with different or mixed DNA which surely affects the infamous foaf:dnaChecksum... danbri, it’s clear that our traditional view of the nature of human ancestry is changing. As much as I’m trying to be sociologically neutral it’s inevitable that I’m going to be influenced by my own world view and experience. But I think that is core to the semantic web effort, in that we accept that there is no one-true-ontology of the world that is true for all people at all times. At best we can reflect a particular consensual view of the world. There’s also an element of pragmatism in that we strive to create ontologies that provide useful abstractions of the world. For the domains that Bio is intended it’s useful to assume that people only have one mother. It worries me a little that Bio can’t be used to completely describe the child in that news article, but then it can’t be used to describe cloned humans either.

Of course, FOAF isn’t immune from these issues :) With the advent of gene therapy and even more established practices such as bone marrow transplants ( http://www.newscientist.com/article.ns?id=mg18825234.600 ), a person can end up with different or mixed DNA which surely affects the infamous foaf:dnaChecksum…

]]> By: iand http://blog.iandavis.com/2006/03/refactoring-bio-with-einstein-part-3-temporal-invariants/comment-page-1#comment-297 iand Wed, 15 Mar 2006 11:14:15 +0000 http://iandavis.com/blog2/?p=918#comment-297 DanC, 1^^xsd:nonNegativeInteger is just a thinko on my part - I'll fix the post. Thanks for the subAbstractions pointer which looks very interesting. At first glance it seems like there's lots to learn from that approach. DanC, 1^^xsd:nonNegativeInteger is just a thinko on my part – I’ll fix the post. Thanks for the subAbstractions pointer which looks very interesting. At first glance it seems like there’s lots to learn from that approach.

]]> By: Dan Brickley http://blog.iandavis.com/2006/03/refactoring-bio-with-einstein-part-3-temporal-invariants/comment-page-1#comment-296 Dan Brickley Wed, 15 Mar 2006 08:54:23 +0000 http://iandavis.com/blog2/?p=918#comment-296 Is it necessarily so that one can't have two "biological" mothers? Consider http://news.bbc.co.uk/1/hi/health/4225564.stm for example... Is it necessarily so that one can’t have two “biological” mothers?

Consider http://news.bbc.co.uk/1/hi/health/4225564.stm for example…

]]> By: Danny http://blog.iandavis.com/2006/03/refactoring-bio-with-einstein-part-3-temporal-invariants/comment-page-1#comment-295 Danny Wed, 15 Mar 2006 08:49:40 +0000 http://iandavis.com/blog2/?p=918#comment-295 Wonderful material, thanks Ian! Wonderful material, thanks Ian!

]]> By: Dan Connolly http://blog.iandavis.com/2006/03/refactoring-bio-with-einstein-part-3-temporal-invariants/comment-page-1#comment-294 Dan Connolly Wed, 15 Mar 2006 06:02:37 +0000 http://iandavis.com/blog2/?p=918#comment-294 see also <a href="http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Apr/0005.html" rel="nofollow">a message I wrote to www-rdf-calendar about cyc subabstractions in apr 2003</a> see also a message I wrote to www-rdf-calendar about cyc subabstractions in apr 2003

]]> By: Dan Connolly http://blog.iandavis.com/2006/03/refactoring-bio-with-einstein-part-3-temporal-invariants/comment-page-1#comment-293 Dan Connolly Wed, 15 Mar 2006 06:00:48 +0000 http://iandavis.com/blog2/?p=918#comment-293 See also the cyc time vocabulary, especially <a href="http://www.cyc.com/cycdoc/vocab/time-vocab.html#subAbstractions" rel="nofollow">subAbstractions</a>: `AlbertEinsteinWhileAtPrinceton' is a #$subAbstractions of `AlbertEinsteinAsAnAdult', which in turn is a #$subAbstractions of 'AlbertEinstein', which in turn is a #$subAbstractions only of itself (hence 'AlbertEinstein' is an instance of #$Entity (q.v.)). See also the cyc time vocabulary, especially subAbstractions: `AlbertEinsteinWhileAtPrinceton’ is a #$subAbstractions of `AlbertEinsteinAsAnAdult’, which in turn is a #$subAbstractions of ‘AlbertEinstein’, which in turn is a #$subAbstractions only of itself (hence ‘AlbertEinstein’ is an instance of #$Entity (q.v.)).

]]> By: Dan Connolly http://blog.iandavis.com/2006/03/refactoring-bio-with-einstein-part-3-temporal-invariants/comment-page-1#comment-292 Dan Connolly Wed, 15 Mar 2006 05:57:28 +0000 http://iandavis.com/blog2/?p=918#comment-292 er... 1^^xsd:nonNegativeInteger is overkill... and buggy. You can just write 1. If you want to use the long form, it's "1:^^xsd:nonNegativeInteger , with the quotes. er… 1^^xsd:nonNegativeInteger is overkill… and buggy.
You can just write 1.
If you want to use the long form, it’s “1:^^xsd:nonNegativeInteger ,
with the quotes.

]]> By: Seth Ladd http://blog.iandavis.com/2006/03/refactoring-bio-with-einstein-part-3-temporal-invariants/comment-page-1#comment-291 Seth Ladd Wed, 15 Mar 2006 02:53:38 +0000 http://iandavis.com/blog2/?p=918#comment-291 Keep 'em coming! These practical examples help fill in many gaps when reading all of the academic literature. If you have any pointers or recommendations for other sources that go into this deeper, I'd love to check them out. Keep ‘em coming! These practical examples help fill in many gaps when reading all of the academic literature. If you have any pointers or recommendations for other sources that go into this deeper, I’d love to check them out.

]]> By: iand http://blog.iandavis.com/2006/03/refactoring-bio-with-einstein-part-3-temporal-invariants/comment-page-1#comment-290 iand Tue, 14 Mar 2006 23:25:26 +0000 http://iandavis.com/blog2/?p=918#comment-290 Thanks. I'm learning a lot as I think through these posts too and using concrete examples to explain the ideas really helps. Thanks. I’m learning a lot as I think through these posts too and using concrete examples to explain the ideas really helps.

]]>