From ksrinivs at gmail.com Mon Oct 1 14:28:59 2007 From: ksrinivs at gmail.com (Kavitha Srinivas) Date: Mon, 1 Oct 2007 10:28:59 -0400 Subject: [Developers] Question about the backend software and data In-Reply-To: <47000610.908@gmail.com> References: <46FCC9CC.8080007@chello.at> <46FD6A50.6080807@metaweb.com> <46FD6CD8.3030206@thefirst.org> <46FD70B3.8080802@metaweb.com> <46FD7C20.3000400@thefirst.org> <46FDAA27.90101@metaweb.com> <02770076-54BC-4D1B-9F0D-DEE024864196@metaweb.com> <47000610.908@gmail.com> Message-ID: Not really, am in interested in modeling issues rather than a specific domain. Thanks! Kavitha On Sep 30, 2007, at 4:24 PM, Shawn Simister wrote: > I'm just putting together a new version of the ontology now which I > will make available on my website. Is there a particular domain > that you want to see? Right now I'm just focusing on the 67 or so > main Freebase domains. > > Shawn > > Kavitha Srinivas wrote: >> Is Shawn's OWL ontology available somewhere? >> Kavitha >> >> On Sep 28, 2007, at 9:52 PM, John Giannandrea wrote: >> >> >>> Kirrily Robert wrote: >>> >>>> To me, RDF sounds like it meets those criteria. >>>> >>> I agree that an RDF dump of the data makes sense. >>> There are important details though, like how to represent >>> the schema. Shawn Simister previously posted about >>> creating an OWL ontology from the freebase types, rather >>> than using RDF Schema, which seemed sensible. >>> >>> The key point was that Shawn was able to extract the data >>> and turn it into those formats, using HTTP, MQL and JSON. >>> >>> -jg >>> >>> _______________________________________________ >>> Developers mailing list >>> Developers at freebase.com >>> http://lists.freebase.com/mailman/listinfo/developers >>> >> _______________________________________________ >> Developers mailing list >> Developers at freebase.com >> http://lists.freebase.com/mailman/listinfo/developers >> >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.freebase.com/pipermail/developers/attachments/20071001/778c39bb/attachment.htm From niallo at metaweb.com Mon Oct 1 23:49:27 2007 From: niallo at metaweb.com (Niall O'Higgins) Date: Mon, 1 Oct 2007 16:49:27 -0700 Subject: [Developers] Weekly sandbox update Message-ID: <20071001234927.GP28540@digdug.inet> Hi, Data refresh and software update of http://sandbox.freebase.com/ happening over the next few minutes. -- Niall O'Higgins Software Engineer Metaweb Technologies, Inc. From niallo at metaweb.com Tue Oct 2 00:16:51 2007 From: niallo at metaweb.com (Niall O'Higgins) Date: Mon, 1 Oct 2007 17:16:51 -0700 Subject: [Developers] Weekly sandbox update In-Reply-To: <20071001234927.GP28540@digdug.inet> References: <20071001234927.GP28540@digdug.inet> Message-ID: <20071002001651.GR28540@digdug.inet> Update complete. On Mon, Oct 01, 2007 at 04:49:27PM -0700, Niall O'Higgins wrote: > Hi, > > Data refresh and software update of http://sandbox.freebase.com/ > happening over the next few minutes. > > -- > Niall O'Higgins > Software Engineer > Metaweb Technologies, Inc. > _______________________________________________ > Developers mailing list > Developers at freebase.com > http://lists.freebase.com/mailman/listinfo/developers -- Niall O'Higgins Software Engineer Metaweb Technologies, Inc. From pingel at cognity.de Tue Oct 2 09:55:37 2007 From: pingel at cognity.de (Christoph Pingel) Date: Tue, 2 Oct 2007 11:55:37 +0200 Subject: [Developers] one-off error in freebase? Message-ID: There seems to be a one-off error in freebase. Happens often. I'm using Camino and Firefox on mac OS X. See for example: http://www.freebase.com/view/simon_penny Simon is born in 1955 (which is what I entered), he left his art schools in 1979 and 1982 respectively. As far as I can tell, this happens with more exact dates as well. If it's a day-precision date, the *next* day is shown. Is this a known issue? best regards, Christoph // Christoph Pingel, MA ? Mediendesign & Semantische Technologien ? // pingel at cognity.de -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.freebase.com/pipermail/developers/attachments/20071002/cf8b4cae/attachment.htm From rictic at gmail.com Tue Oct 2 10:04:10 2007 From: rictic at gmail.com (Peter Burns) Date: Tue, 2 Oct 2007 03:04:10 -0700 Subject: [Developers] one-off error in freebase? In-Reply-To: References: Message-ID: <2f9bef730710020304j5633c14mcd09a93d460db1ed@mail.gmail.com> I see the page exactly as you describe. That is to say, his listed date of birth is 1955, and the end-dates for each of the schools he attended are displayed as 1979 and 1982. I'm not sure I understand what the error you're seeing is. What do you expect to happen, and how is that different from what actually is happening? I'm using Camino nightly build 2007062522 on Mac OS X On 10/2/07, Christoph Pingel wrote: > > There seems to be a one-off error in freebase. > > Happens often. I'm using Camino and Firefox on mac OS X. > See for example: http://www.freebase.com/view/simon_penny > > Simon is born in 1955 (which is what I entered), he left his art schools > in 1979 and 1982 respectively. > As far as I can tell, this happens with more exact dates as well. If it's > a day-precision date, the *next* day is shown. > > Is this a known issue? > > best regards, > Christoph > > > > > > > // Christoph Pingel, MA ? Mediendesign & Semantische Technologien ? > // pingel at cognity.de > > > _______________________________________________ > Developers mailing list > Developers at freebase.com > http://lists.freebase.com/mailman/listinfo/developers > > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.freebase.com/pipermail/developers/attachments/20071002/71720b66/attachment-0001.htm From pingel at cognity.de Tue Oct 2 10:29:02 2007 From: pingel at cognity.de (Christoph Pingel) Date: Tue, 2 Oct 2007 12:29:02 +0200 Subject: [Developers] one-off error in freebase? In-Reply-To: <2f9bef730710020304j5633c14mcd09a93d460db1ed@mail.gmail.com> References: <2f9bef730710020304j5633c14mcd09a93d460db1ed@mail.gmail.com> Message-ID: This is odd. On *all* my Mac browsers (Camino, Firefox, Safari) the date of birth is 1956, and the end-dates 1980 and 1983. June 30th will display as July 1st. And so on. No matter if I'm logged in or not. Here's a snippet of HTML (for the first institution passed): 1980 The datevalue and topicvalue properties are correct, but the actual node says 1980 ... ??? best regards, Christoph Am 02.10.2007 um 12:04 Uhr schrieb Peter Burns: > I see the page exactly as you describe. That is to say, his listed > date of birth is 1955, and the end-dates for each of the schools he > attended are displayed as 1979 and 1982. > > I'm not sure I understand what the error you're seeing is. What do > you expect to happen, and how is that different from what actually > is happening? > > I'm using Camino nightly build 2007062522 on Mac OS X > > On 10/2/07, Christoph Pingel wrote: > There seems to be a one-off error in freebase. > > Happens often. I'm using Camino and Firefox on mac OS X. > See for example: http://www.freebase.com/view/simon_penny > > Simon is born in 1955 (which is what I entered), he left his art > schools in 1979 and 1982 respectively. > As far as I can tell, this happens with more exact dates as well. > If it's a day-precision date, the *next* day is shown. > > Is this a known issue? > > best regards, > Christoph > > > > > > > // Christoph Pingel, MA ? Mediendesign & Semantische Technologien ? > // pingel at cognity.de > > > _______________________________________________ > Developers mailing list > Developers at freebase.com > http://lists.freebase.com/mailman/listinfo/developers > > > _______________________________________________ > Developers mailing list > Developers at freebase.com > http://lists.freebase.com/mailman/listinfo/developers -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.freebase.com/pipermail/developers/attachments/20071002/d883ad56/attachment.htm From skud at infotrope.net Tue Oct 2 11:58:42 2007 From: skud at infotrope.net (Kirrily Robert) Date: Tue, 2 Oct 2007 21:58:42 +1000 Subject: [Developers] one-off error in freebase? In-Reply-To: References: <2f9bef730710020304j5633c14mcd09a93d460db1ed@mail.gmail.com> Message-ID: On 10/2/07, Christoph Pingel wrote: > This is odd. On *all* my Mac browsers (Camino, Firefox, Safari) the date of > birth is 1956, and the end-dates 1980 and 1983. > June 30th will display as July 1st. And so on. No matter if I'm logged in or > not. I'm seeing this too on Firefox on Mac. I think it's relatively recent, though; I was playing with dates on some fields a month or so back and didn't notice anything, but I did notice this off-by-one error yesterday or today, and just thought it was a mis-entry. K. -- Kirrily Robert skud at infotrope.net http://infotrope.net From skud at infotrope.net Tue Oct 2 14:00:03 2007 From: skud at infotrope.net (Kirrily Robert) Date: Wed, 3 Oct 2007 00:00:03 +1000 Subject: [Developers] content of documents Message-ID: Is there any way in MQL to get at the content of a document object? When I look at things like the example at http://www.freebase.com/view/helptopic?id=%239202a8c04000641f800000000544e139#WhatsNew.html, it seems that the only way to get at it is via a separate web query. I'm trying to build an RSS feed for watched discussions -- you can see it at http://tools.freebasing.org/feeds/watched.cgi?user=skud (or substitute your own username) if you don't mind it breaking etc, because I'm still working on it -- but obviously the content of each post would be a useful thing to have. But all the examples I've seen that include document content are in JavaScript and are fetching something like the blurb you find at http://freebase.com/api/trans/blurb/ Obviously I *could* do that, but I'd rather not be spawning 50 HTTP requests every time someone hits the RSS feed. K. -- Kirrily Robert skud at infotrope.net http://infotrope.net From brendan at metaweb.com Tue Oct 2 18:16:06 2007 From: brendan at metaweb.com (brendan neutra) Date: Tue, 02 Oct 2007 11:16:06 -0700 Subject: [Developers] one-off error in freebase? In-Reply-To: References: <2f9bef730710020304j5633c14mcd09a93d460db1ed@mail.gmail.com> Message-ID: <47028AE6.3090006@metaweb.com> I believe this is a known issue with our javascript datetime widget. I don't think it's browser specific. I just tried changing my computers time zone to "down under" and reproduced the issue you're describing in IE7. If you view source you can see that the underlying datapoint for dob is "1955". I think the javascript is trying to adjust for timezone, which would be nice if the datetime were describing when a conference call is starting but no so good in this case (I think it finds 1955 to be ambiguous as a fully qualified point in time). I'll let someone with more understanding of the issue chime in... Brendan Kirrily Robert wrote: > On 10/2/07, Christoph Pingel wrote: >> This is odd. On *all* my Mac browsers (Camino, Firefox, Safari) the date of >> birth is 1956, and the end-dates 1980 and 1983. >> June 30th will display as July 1st. And so on. No matter if I'm logged in or >> not. > > I'm seeing this too on Firefox on Mac. I think it's relatively > recent, though; I was playing with dates on some fields a month or so > back and didn't notice anything, but I did notice this off-by-one > error yesterday or today, and just thought it was a mis-entry. > > K. > From pingel at cognity.de Tue Oct 2 21:37:50 2007 From: pingel at cognity.de (Christoph Pingel) Date: Tue, 2 Oct 2007 23:37:50 +0200 Subject: [Developers] Reasoner? Message-ID: <7DE19E89-C469-4FC2-BFD3-DFFE68A3EC10@cognity.de> Hello, I'm wondering if there is any artificial reasoner involved in freebase? My impression from looking into MQL is that queries only result in the explicit structures, not implicit ones. For example, can I ask for universities in France and get *all* French universities even if "contained by Lyon" or "contained by Toulouse" are the only location values that are there? Another case would be the research for people with a certain expertise that can be expressed in rules like "person knows about topic if person wrote book about topic or person wrote article about topic or person organized workshop about topic or person curated exposition about topic ". My second question is somewhat related. Perhaps it would make sense to have specialisation hierarchies (or inheritance chains) for relationships as well, not just for types. Looking at movies, for example, all the actors, the director etc. are all involved with a certain movie. From time to time, there are cases when it would be convenient to just ask for ?people involved?... Hm, I'm not sure. Perhaps what I'm asking for is to have inheritance between compound value types... best regards, Christoph From narphorium at gmail.com Tue Oct 2 22:08:22 2007 From: narphorium at gmail.com (Shawn Simister) Date: Tue, 02 Oct 2007 18:08:22 -0400 Subject: [Developers] Reasoner? In-Reply-To: <7DE19E89-C469-4FC2-BFD3-DFFE68A3EC10@cognity.de> References: <7DE19E89-C469-4FC2-BFD3-DFFE68A3EC10@cognity.de> Message-ID: <4702C156.9010108@gmail.com> Once the Freebase data is available as RDF it will be possible to reason about it using any of the existing RDF reasoners out there. I'm working on making this happen but I only have a little bit of free time right now. Shawn Christoph Pingel wrote: > Hello, > > I'm wondering if there is any artificial reasoner involved in > freebase? My impression from looking into MQL is that queries only > result in the explicit structures, not implicit ones. > > For example, can I ask for universities in France and get *all* > French universities even if "contained by Lyon" or "contained by > Toulouse" are the only location values that are there? > > Another case would be the research for people with a certain > expertise that can be expressed in rules like > "person knows about topic if > person wrote book about topic > or person wrote article about topic > or person organized workshop about topic > or person curated exposition about topic > ". > > My second question is somewhat related. Perhaps it would make sense > to have specialisation hierarchies (or inheritance chains) for > relationships as well, not just for types. > > Looking at movies, for example, all the actors, the director etc. are > all involved with a certain movie. From time to time, there are cases > when it would be convenient to just ask for ?people involved?... > Hm, I'm not sure. Perhaps what I'm asking for is to have inheritance > between compound value types... > > best regards, > Christoph > _______________________________________________ > Developers mailing list > Developers at freebase.com > http://lists.freebase.com/mailman/listinfo/developers > > From alecf at metaweb.com Tue Oct 2 22:52:38 2007 From: alecf at metaweb.com (Alec Flett) Date: Tue, 02 Oct 2007 15:52:38 -0700 Subject: [Developers] one-off error in freebase? In-Reply-To: <47028AE6.3090006@metaweb.com> References: <2f9bef730710020304j5633c14mcd09a93d460db1ed@mail.gmail.com> <47028AE6.3090006@metaweb.com> Message-ID: <4702CBB6.6020600@metaweb.com> I can explain this a little better: In the graph we can represent the date as 2007, 2007-10, 2007-10-03, etc..technically these dates are UTC. Unfortunately when this gets loaded into a Javascript object in the browser, it gets converted to an absolute time (like Jan 1, 2007 at 0:00UTC, etc) and when we UTC-adjust that to whatever local timezone you're viewing this in, it gets moved to the day before. So users "west" of Greenwich see times adjusted down, and users "east" see it adjusted up by a few hours.. or maybe it's the other way around. This is probably a good lesson for any mashup writers - make sure your local date/time representation properly handles generalized dates, without timezone information. In the graph, all dates/times are really meant to be UTC ultimately, but if a time isn't specified, the behavior isn't well defined. We'll have a fix for this on freebase.com for the specific problems mentioned below in the next few weeks. Alec brendan neutra wrote: > I believe this is a known issue with our javascript datetime widget. I don't > think it's browser specific. I just tried changing my computers time zone to > "down under" and reproduced the issue you're describing in IE7. If you view > source you can see that the underlying datapoint for dob is "1955". I think the > javascript is trying to adjust for timezone, which would be nice if the datetime > were describing when a conference call is starting but no so good in this case > (I think it finds 1955 to be ambiguous as a fully qualified point in time). > I'll let someone with more understanding of the issue chime in... > > Brendan > > Kirrily Robert wrote: > >> On 10/2/07, Christoph Pingel wrote: >> >>> This is odd. On *all* my Mac browsers (Camino, Firefox, Safari) the date of >>> birth is 1956, and the end-dates 1980 and 1983. >>> June 30th will display as July 1st. And so on. No matter if I'm logged in or >>> not. >>> >> I'm seeing this too on Firefox on Mac. I think it's relatively >> recent, though; I was playing with dates on some fields a month or so >> back and didn't notice anything, but I did notice this off-by-one >> error yesterday or today, and just thought it was a mis-entry. >> >> K. >> >> > _______________________________________________ > Developers mailing list > Developers at freebase.com > http://lists.freebase.com/mailman/listinfo/developers > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.freebase.com/pipermail/developers/attachments/20071002/8defa2d9/attachment-0001.htm From jeff at metaweb.com Tue Oct 2 23:04:49 2007 From: jeff at metaweb.com (Jeff Prucher) Date: Tue, 2 Oct 2007 16:04:49 -0700 Subject: [Developers] Major changes to the government domain Message-ID: <004701c80548$b3d7fd00$bc01a8c0@p4> Sorry for the cross-post, but I want to make sure I get everyone who might be affected. We are planning to make some very major changes to the government domain in the near future, which will almost certainly break the code of any application that uses these types. If you have written such an application, please let me know (on- or off-list) and I can tell you more about what we're doing and why and we can talk about what, if any changes, might need to be made. My apologies in advance for any difficulties this causes. Thanks, Jeff Prucher Typelibrarian & Ontologist Metaweb, Inc. From ksrinivs at gmail.com Wed Oct 3 02:21:02 2007 From: ksrinivs at gmail.com (Kavitha Srinivas) Date: Tue, 2 Oct 2007 22:21:02 -0400 Subject: [Developers] Reasoner? In-Reply-To: <4702C156.9010108@gmail.com> References: <7DE19E89-C469-4FC2-BFD3-DFFE68A3EC10@cognity.de> <4702C156.9010108@gmail.com> Message-ID: <825F3C0E-DAC1-4BE9-8B57-67A4678FCE8A@gmail.com> Shawn has a good point. If you want to describe role hierarchies, or transitivity (e.g., locatedIn is transitive), then you can use OWL and use OWL reasoners to do this. Kavitha On Oct 2, 2007, at 6:08 PM, Shawn Simister wrote: > Once the Freebase data is available as RDF it will be possible to > reason > about it using any of the existing RDF reasoners out there. I'm > working > on making this happen but I only have a little bit of free time > right now. > > Shawn > > Christoph Pingel wrote: >> Hello, >> >> I'm wondering if there is any artificial reasoner involved in >> freebase? My impression from looking into MQL is that queries only >> result in the explicit structures, not implicit ones. >> >> For example, can I ask for universities in France and get *all* >> French universities even if "contained by Lyon" or "contained by >> Toulouse" are the only location values that are there? >> >> Another case would be the research for people with a certain >> expertise that can be expressed in rules like >> "person knows about topic if >> person wrote book about topic >> or person wrote article about topic >> or person organized workshop about topic >> or person curated exposition about topic >> ". >> >> My second question is somewhat related. Perhaps it would make sense >> to have specialisation hierarchies (or inheritance chains) for >> relationships as well, not just for types. >> >> Looking at movies, for example, all the actors, the director etc. are >> all involved with a certain movie. From time to time, there are cases >> when it would be convenient to just ask for ?people involved?... >> Hm, I'm not sure. Perhaps what I'm asking for is to have inheritance >> between compound value types... >> >> best regards, >> Christoph >> _______________________________________________ >> Developers mailing list >> Developers at freebase.com >> http://lists.freebase.com/mailman/listinfo/developers >> >> > > _______________________________________________ > Developers mailing list > Developers at freebase.com > http://lists.freebase.com/mailman/listinfo/developers From robert at metaweb.com Wed Oct 3 05:42:13 2007 From: robert at metaweb.com (Robert Cook) Date: Wed, 3 Oct 2007 01:42:13 -0400 (EDT) Subject: [Developers] Reasoner? In-Reply-To: <825F3C0E-DAC1-4BE9-8B57-67A4678FCE8A@gmail.com> Message-ID: <5214993.196051191390133193.JavaMail.root@h00215> Just so you know, it's one of our goals to support transitive closure in particular on geographic containment and do it in a performant way. That is, your French University example below should work with little penalty in execution time. I don't currently have an ETA, but it's something we're serious about. R ----- Original Message ----- From: "Kavitha Srinivas" To: "Shawn Simister" Cc: developers at freebase.com Sent: Tuesday, October 2, 2007 7:21:02 PM (GMT-0800) America/Los_Angeles Subject: Re: [Developers] Reasoner? Shawn has a good point. If you want to describe role hierarchies, or transitivity (e.g., locatedIn is transitive), then you can use OWL and use OWL reasoners to do this. Kavitha On Oct 2, 2007, at 6:08 PM, Shawn Simister wrote: > Once the Freebase data is available as RDF it will be possible to > reason > about it using any of the existing RDF reasoners out there. I'm > working > on making this happen but I only have a little bit of free time > right now. > > Shawn > > Christoph Pingel wrote: >> Hello, >> >> I'm wondering if there is any artificial reasoner involved in >> freebase? My impression from looking into MQL is that queries only >> result in the explicit structures, not implicit ones. >> >> For example, can I ask for universities in France and get *all* >> French universities even if "contained by Lyon" or "contained by >> Toulouse" are the only location values that are there? >> >> Another case would be the research for people with a certain >> expertise that can be expressed in rules like >> "person knows about topic if >> person wrote book about topic >> or person wrote article about topic >> or person organized workshop about topic >> or person curated exposition about topic >> ". >> >> My second question is somewhat related. Perhaps it would make sense >> to have specialisation hierarchies (or inheritance chains) for >> relationships as well, not just for types. >> >> Looking at movies, for example, all the actors, the director etc. are >> all involved with a certain movie. From time to time, there are cases >> when it would be convenient to just ask for ?people involved?... >> Hm, I'm not sure. Perhaps what I'm asking for is to have inheritance >> between compound value types... >> >> best regards, >> Christoph >> _______________________________________________ >> Developers mailing list >> Developers at freebase.com >> http://lists.freebase.com/mailman/listinfo/developers >> >> > > _______________________________________________ > Developers mailing list > Developers at freebase.com > http://lists.freebase.com/mailman/listinfo/developers _______________________________________________ Developers mailing list Developers at freebase.com http://lists.freebase.com/mailman/listinfo/developers From rictic at gmail.com Wed Oct 3 05:51:15 2007 From: rictic at gmail.com (Peter Burns) Date: Tue, 2 Oct 2007 22:51:15 -0700 Subject: [Developers] Reasoner? In-Reply-To: <5214993.196051191390133193.JavaMail.root@h00215> References: <825F3C0E-DAC1-4BE9-8B57-67A4678FCE8A@gmail.com> <5214993.196051191390133193.JavaMail.root@h00215> Message-ID: <2f9bef730710022251k137e56d9l787a9da961d6b6ec@mail.gmail.com> Is the plan to support transitivity in a generalized way? Could someone assert that ancestry is transitive in such a way that a query asking "is Aaron Burr an ancestor of Jody Foster" would cause metaweb to walk up 'parent' links and give an answer? On 10/2/07, Robert Cook wrote: > > Just so you know, it's one of our goals to support transitive closure in > particular on geographic containment and do it in a performant way. That > is, your French University example below should work with little penalty in > execution time. I don't currently have an ETA, but it's something we're > serious about. > > R > > ----- Original Message ----- > From: "Kavitha Srinivas" > To: "Shawn Simister" > Cc: developers at freebase.com > Sent: Tuesday, October 2, 2007 7:21:02 PM (GMT-0800) America/Los_Angeles > Subject: Re: [Developers] Reasoner? > > Shawn has a good point. If you want to describe role hierarchies, or > transitivity (e.g., locatedIn is transitive), then you can use OWL > and use OWL reasoners to do this. > Kavitha > > On Oct 2, 2007, at 6:08 PM, Shawn Simister wrote: > > > Once the Freebase data is available as RDF it will be possible to > > reason > > about it using any of the existing RDF reasoners out there. I'm > > working > > on making this happen but I only have a little bit of free time > > right now. > > > > Shawn > > > > Christoph Pingel wrote: > >> Hello, > >> > >> I'm wondering if there is any artificial reasoner involved in > >> freebase? My impression from looking into MQL is that queries only > >> result in the explicit structures, not implicit ones. > >> > >> For example, can I ask for universities in France and get *all* > >> French universities even if "contained by Lyon" or "contained by > >> Toulouse" are the only location values that are there? > >> > >> Another case would be the research for people with a certain > >> expertise that can be expressed in rules like > >> "person knows about topic if > >> person wrote book about topic > >> or person wrote article about topic > >> or person organized workshop about topic > >> or person curated exposition about topic > >> ". > >> > >> My second question is somewhat related. Perhaps it would make sense > >> to have specialisation hierarchies (or inheritance chains) for > >> relationships as well, not just for types. > >> > >> Looking at movies, for example, all the actors, the director etc. are > >> all involved with a certain movie. From time to time, there are cases > >> when it would be convenient to just ask for ?people involved?... > >> Hm, I'm not sure. Perhaps what I'm asking for is to have inheritance > >> between compound value types... > >> > >> best regards, > >> Christoph > >> _______________________________________________ > >> Developers mailing list > >> Developers at freebase.com > >> http://lists.freebase.com/mailman/listinfo/developers > >> > >> > > > > _______________________________________________ > > Developers mailing list > > Developers at freebase.com > > http://lists.freebase.com/mailman/listinfo/developers > > _______________________________________________ > Developers mailing list > Developers at freebase.com > http://lists.freebase.com/mailman/listinfo/developers > > _______________________________________________ > Developers mailing list > Developers at freebase.com > http://lists.freebase.com/mailman/listinfo/developers > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.freebase.com/pipermail/developers/attachments/20071002/e2d57e3c/attachment.htm From robert at metaweb.com Wed Oct 3 05:56:24 2007 From: robert at metaweb.com (Robert Cook) Date: Wed, 3 Oct 2007 01:56:24 -0400 (EDT) Subject: [Developers] Reasoner? In-Reply-To: <2f9bef730710022251k137e56d9l787a9da961d6b6ec@mail.gmail.com> Message-ID: <4788702.196221191390984684.JavaMail.root@h00215> We know that we need to handle the geographic containment query problem. The solution to that may or may not be generalizable (again, in a performant way.) R ----- Original Message ----- From: "Peter Burns" To: "Robert Cook" , "Freebase Developers" Sent: Tuesday, October 2, 2007 10:51:15 PM (GMT-0800) America/Los_Angeles Subject: Re: [Developers] Reasoner? Is the plan to support transitivity in a generalized way? Could someone assert that ancestry is transitive in such a way that a query asking "is Aaron Burr an ancestor of Jody Foster" would cause metaweb to walk up 'parent' links and give an answer? On 10/2/07 , Robert Cook < robert at metaweb.com > wrote: Just so you know, it's one of our goals to support transitive closure in particular on geographic containment and do it in a performant way. That is, your French University example below should work with little penalty in execution time. I don't currently have an ETA, but it's something we're serious about. R ----- Original Message ----- From: "Kavitha Srinivas" < ksrinivs at gmail.com > To: "Shawn Simister" < narphorium at gmail.com > Cc: developers at freebase.com Sent: Tuesday, October 2, 2007 7:21:02 PM (GMT-0800) America/Los_Angeles Subject: Re: [Developers] Reasoner? Shawn has a good point. If you want to describe role hierarchies, or transitivity (e.g., locatedIn is transitive), then you can use OWL and use OWL reasoners to do this. Kavitha On Oct 2, 2007 , at 6:08 PM, Shawn Simister wrote: > Once the Freebase data is available as RDF it will be possible to > reason > about it using any of the existing RDF reasoners out there. I'm > working > on making this happen but I only have a little bit of free time > right now. > > Shawn > > Christoph Pingel wrote: >> Hello, >> >> I'm wondering if there is any artificial reasoner involved in >> freebase? My impression from looking into MQL is that queries only >> result in the explicit structures, not implicit ones. >> >> For example, can I ask for universities in France and get *all* >> French universities even if "contained by Lyon" or "contained by >> Toulouse" are the only location values that are there? >> >> Another case would be the research for people with a certain >> expertise that can be expressed in rules like >> "person knows about topic if >> person wrote book about topic >> or person wrote article about topic >> or person organized workshop about topic >> or person curated exposition about topic >> ". >> >> My second question is somewhat related. Perhaps it would make sense >> to have specialisation hierarchies (or inheritance chains) for >> relationships as well, not just for types. >> >> Looking at movies, for example, all the actors, the director etc. are >> all involved with a certain movie. From time to time, there are cases >> when it would be convenient to just ask for ?people involved?... >> Hm, I'm not sure. Perhaps what I'm asking for is to have inheritance >> between compound value types... >> >> best regards, >> Christoph >> _______________________________________________ >> Developers mailing list >> Developers at freebase.com >> http://lists.freebase.com/mailman/listinfo/developers >> >> > > _______________________________________________ > Developers mailing list > Developers at freebase.com > http://lists.freebase.com/mailman/listinfo/developers _______________________________________________ Developers mailing list Developers at freebase.com http://lists.freebase.com/mailman/listinfo/developers _______________________________________________ Developers mailing list Developers at freebase.com http://lists.freebase.com/mailman/listinfo/developers From narphorium at gmail.com Wed Oct 3 09:40:06 2007 From: narphorium at gmail.com (Shawn Simister) Date: Wed, 03 Oct 2007 05:40:06 -0400 Subject: [Developers] Question about the backend software and data In-Reply-To: References: <46FCC9CC.8080007@chello.at> <46FD6A50.6080807@metaweb.com> <46FD6CD8.3030206@thefirst.org> <46FD70B3.8080802@metaweb.com> <46FD7C20.3000400@thefirst.org> <46FDAA27.90101@metaweb.com> <02770076-54BC-4D1B-9F0D-DEE024864196@metaweb.com> <47000610.908@gmail.com> Message-ID: <47036376.30701@gmail.com> I've just put together a first attempt at representing the Freebase type system as OWL ontologies. There's still some work to be done but I figured I'd give you guys an update and hopefully get some feedback on how I'm modeling these domains in OWL. You can download it at http://www.simister.ca/temp/freebase-owl-03-10-07.zip. The way that it's set up right now is one OWL ontology per file per Freebase domain. Let me know what you think. Shawn Kavitha Srinivas wrote: > Not really, am in interested in modeling issues rather than a specific > domain. Thanks! > Kavitha > > On Sep 30, 2007, at 4:24 PM, Shawn Simister wrote: > >> I'm just putting together a new version of the ontology now which I >> will make available on my website. Is there a particular domain that >> you want to see? Right now I'm just focusing on the 67 or so main >> Freebase domains. >> >> Shawn >> >> Kavitha Srinivas wrote: >>> Is Shawn's OWL ontology available somewhere? >>> Kavitha >>> >>> On Sep 28, 2007, at 9:52 PM, John Giannandrea wrote: >>> >>> >>>> Kirrily Robert wrote: >>>> >>>>> To me, RDF sounds like it meets those criteria. >>>>> >>>> I agree that an RDF dump of the data makes sense. >>>> There are important details though, like how to represent >>>> the schema. Shawn Simister previously posted about >>>> creating an OWL ontology from the freebase types, rather >>>> than using RDF Schema, which seemed sensible. >>>> >>>> The key point was that Shawn was able to extract the data >>>> and turn it into those formats, using HTTP, MQL and JSON. >>>> >>>> -jg >>>> >>>> _______________________________________________ >>>> Developers mailing list >>>> Developers at freebase.com >>>> http://lists.freebase.com/mailman/listinfo/developers >>>> >>> _______________________________________________ >>> Developers mailing list >>> Developers at freebase.com >>> http://lists.freebase.com/mailman/listinfo/developers >>> >>> >> > From ksrinivs at gmail.com Wed Oct 3 13:05:46 2007 From: ksrinivs at gmail.com (Kavitha Srinivas) Date: Wed, 3 Oct 2007 09:05:46 -0400 Subject: [Developers] Question about the backend software and data In-Reply-To: <47036376.30701@gmail.com> References: <46FCC9CC.8080007@chello.at> <46FD6A50.6080807@metaweb.com> <46FD6CD8.3030206@thefirst.org> <46FD70B3.8080802@metaweb.com> <46FD7C20.3000400@thefirst.org> <46FDAA27.90101@metaweb.com> <02770076-54BC-4D1B-9F0D-DEE024864196@metaweb.com> <47000610.908@gmail.com> <47036376.30701@gmail.com> Message-ID: <9052D9FD-51F8-402C-B00C-942B101754D1@gmail.com> Many thanks! Will take a look and get back to you. Kavitha On Oct 3, 2007, at 5:40 AM, Shawn Simister wrote: > I've just put together a first attempt at representing the Freebase > type system as OWL ontologies. There's still some work to be done > but I figured I'd give you guys an update and hopefully get some > feedback on how I'm modeling these domains in OWL. > > You can download it at http://www.simister.ca/temp/freebase- > owl-03-10-07.zip. The way that it's set up right now is one OWL > ontology per file per Freebase domain. Let me know what you think. > > Shawn > > Kavitha Srinivas wrote: >> Not really, am in interested in modeling issues rather than a >> specific domain. Thanks! >> Kavitha >> >> On Sep 30, 2007, at 4:24 PM, Shawn Simister wrote: >> >>> I'm just putting together a new version of the ontology now which >>> I will make available on my website. Is there a particular domain >>> that you want to see? Right now I'm just focusing on the 67 or so >>> main Freebase domains. >>> >>> Shawn >>> >>> Kavitha Srinivas wrote: >>>> Is Shawn's OWL ontology available somewhere? >>>> Kavitha >>>> >>>> On Sep 28, 2007, at 9:52 PM, John Giannandrea wrote: >>>> >>>> >>>>> Kirrily Robert wrote: >>>>> >>>>>> To me, RDF sounds like it meets those criteria. >>>>>> >>>>> I agree that an RDF dump of the data makes sense. >>>>> There are important details though, like how to represent >>>>> the schema. Shawn Simister previously posted about >>>>> creating an OWL ontology from the freebase types, rather >>>>> than using RDF Schema, which seemed sensible. >>>>> >>>>> The key point was that Shawn was able to extract the data >>>>> and turn it into those formats, using HTTP, MQL and JSON. >>>>> >>>>> -jg >>>>> >>>>> _______________________________________________ >>>>> Developers mailing list >>>>> Developers at freebase.com >>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>> >>>> _______________________________________________ >>>> Developers mailing list >>>> Developers at freebase.com >>>> http://lists.freebase.com/mailman/listinfo/developers >>>> >>>> >>> >> > From tim at metaweb.com Wed Oct 3 21:44:00 2007 From: tim at metaweb.com (Tim Kientzle) Date: Wed, 03 Oct 2007 14:44:00 -0700 Subject: [Developers] Mailing List Configuration Change In-Reply-To: <97808690-5A78-4EAA-99A4-D9E598D9E28D@cognity.de> References: <8ED22C35-AA2C-487C-916E-7B046005631B@cognity.de> <4703AA79.8020508@gmail.com> <8A02B17D-6CDA-47C7-A3A4-9049E596AC7B@metaweb.com> <97808690-5A78-4EAA-99A4-D9E598D9E28D@cognity.de> Message-ID: <47040D20.7030002@metaweb.com> Christoph Pingel wrote: > > PS - I think mails from the list should have a reply-to field in the > header with the list address. We've just changed this. Reply-to now defaults to the entire list, for both data-modeling and developers mailing lists. Tim Kientzle From alecf at metaweb.com Wed Oct 3 21:53:08 2007 From: alecf at metaweb.com (Alec Flett) Date: Wed, 03 Oct 2007 14:53:08 -0700 Subject: [Developers] content of documents In-Reply-To: References: Message-ID: <47040F44.1050803@metaweb.com> Kirrily Robert wrote: > Is there any way in MQL to get at the content of a document object? > When I look at things like the example at > http://www.freebase.com/view/helptopic?id=%239202a8c04000641f800000000544e139#WhatsNew.html, > it seems that the only way to get at it is via a separate web query. > > At the moment, yes that is the only way. It is unfortunate to have to make 50 HTTP requests, but you could do some amount of caching on your end, as blurbs are unlikely to change anytime soon, and are nice and small. Obviously this adds state and complexity to an app like yours. In any case, 50 HTTP requests clearly isn't ideal. I wonder if it would be useful to provide some sort of bulk retrieval where you're retrieving either a list of GUIDs to documents, or doing it based on the results of a mql query? What would be helpful to you? If we batch them up we'd probably have to use some sort of mime enclosure mechanism.. Alec > I'm trying to build an RSS feed for watched discussions -- you can see > it at http://tools.freebasing.org/feeds/watched.cgi?user=skud (or > substitute your own username) if you don't mind it breaking etc, > because I'm still working on it -- but obviously the content of each > post would be a useful thing to have. But all the examples I've seen > that include document content are in JavaScript and are fetching > something like the blurb you find at > http://freebase.com/api/trans/blurb/ Obviously I *could* > do that, but I'd rather not be spawning 50 HTTP requests every time > someone hits the RSS feed. > > K. > > From elh at cs.pdx.edu Thu Oct 4 07:47:51 2007 From: elh at cs.pdx.edu (Eric Hanson) Date: Thu, 04 Oct 2007 00:47:51 -0700 Subject: [Developers] products domain Message-ID: <47049AA7.9080602@cs.pdx.edu> Hi, I do contract web work for a natural foods distributor called Azure Standard, located in Dufur, OR. They distribute over 8000 products, many of which are food, with a focus on sustainability and health. They distribute to stores and bulk consumers in the western US and Canada. We're working on the next version of our website, and are using RDF to describe products and create a product search interface that makes use of RDF's flexible and multi-faceted structure. We're developing an ontology of product characteristics like organic, wheat free, celiac-safe, sugar free, etc, and also a set of classes like fruits, grains, nuts, vegetables, processed foods, beverages, etc. I'm very impressed with Freebase, and would like to align our product description work with whatever is being done here in Freebase. Are there any plans to build a centralized product database? Is there something existing or in the works that we could integrate with? Might we be able to contribute some types and properties? Thanks, Eric From elh at cs.pdx.edu Thu Oct 4 07:58:07 2007 From: elh at cs.pdx.edu (Eric Hanson) Date: Thu, 04 Oct 2007 00:58:07 -0700 Subject: [Developers] [Data-modeling] products domain Message-ID: <47049D0F.9020303@cs.pdx.edu> Oops...forwarding this to the data-modeling list. Eric -------- Original Message -------- Hi, I do contract web work for a natural foods distributor called Azure Standard, located in Dufur, OR. They distribute over 8000 products, many of which are food, with a focus on sustainability and health. They distribute to stores and bulk consumers in the western US and Canada. We're working on the next version of our website, and are using RDF to describe products and create a product search interface that makes use of RDF's flexible and multi-faceted structure. We're developing an ontology of product characteristics like organic, wheat free, celiac-safe, sugar free, etc, and also a set of classes like fruits, grains, nuts, vegetables, processed foods, beverages, etc. I'm very impressed with Freebase, and would like to align our product description work with whatever is being done here in Freebase. Are there any plans to build a centralized product database? Is there something existing or in the works that we could integrate with? Might we be able to contribute some types and properties? Thanks, Eric _______________________________________________ Data-modeling mailing list Data-modeling at freebase.com http://lists.freebase.com/mailman/listinfo/data-modeling From skud at infotrope.net Thu Oct 4 13:54:38 2007 From: skud at infotrope.net (Kirrily Robert) Date: Thu, 4 Oct 2007 23:54:38 +1000 Subject: [Developers] Whining about the # sign Message-ID: Whose bright idea was it to put that # sign in object IDs? URI encoding it to %23 all the time is a PITA. K. -- Kirrily Robert skud at infotrope.net http://infotrope.net From jason at metaweb.com Thu Oct 4 16:41:20 2007 From: jason at metaweb.com (Jason Douglas) Date: Thu, 4 Oct 2007 09:41:20 -0700 Subject: [Developers] Whining about the # sign In-Reply-To: References: Message-ID: Hang in there... a fix is on the works. :-) On Oct 4, 2007, at 6:54 AM, Kirrily Robert wrote: > > Whose bright idea was it to put that # sign in object IDs? URI > encoding it to %23 all the time is a PITA. > > > K. > > -- > Kirrily Robert > skud at infotrope.net > http://infotrope.net > _______________________________________________ > Developers mailing list > Developers at freebase.com > http://lists.freebase.com/mailman/listinfo/developers From jason at metaweb.com Thu Oct 4 16:53:39 2007 From: jason at metaweb.com (Jason Douglas) Date: Thu, 4 Oct 2007 09:53:39 -0700 Subject: [Developers] products domain In-Reply-To: <47049AA7.9080602@cs.pdx.edu> References: <47049AA7.9080602@cs.pdx.edu> Message-ID: We're definitely interested in fleshing out the products domain. We've started thinking about the schemas a little, but honestly having real-world uses of it would really help. By all means, I encourage you to try out types and properties in your user domain and we'd be happy to offer advice and feedback along the way (take advantage of the data-modeling at freebase.com list). For example, one area of the products domain that I'm curious to see how it works out is the whole lumping of products as most people would think of things (Hamlet, iPod) vs. the splitting by individual SKUs (every published edition of Hamlet, every color or branded version of the iPod, etc.). -jason On Oct 4, 2007, at 12:47 AM, Eric Hanson wrote: > Hi, > > I do contract web work for a natural foods distributor called Azure > Standard, located in Dufur, OR. They distribute over 8000 products, > many of which are food, with a focus on sustainability and health. > They > distribute to stores and bulk consumers in the western US and Canada. > > We're working on the next version of our website, and are using RDF to > describe products and create a product search interface that makes use > of RDF's flexible and multi-faceted structure. We're developing an > ontology of product characteristics like organic, wheat free, > celiac-safe, sugar free, etc, and also a set of classes like fruits, > grains, nuts, vegetables, processed foods, beverages, etc. > > I'm very impressed with Freebase, and would like to align our product > description work with whatever is being done here in Freebase. > > Are there any plans to build a centralized product database? Is there > something existing or in the works that we could integrate with? > Might > we be able to contribute some types and properties? > > Thanks, > Eric > _______________________________________________ > Developers mailing list > Developers at freebase.com > http://lists.freebase.com/mailman/listinfo/developers From skud at infotrope.net Thu Oct 4 23:33:45 2007 From: skud at infotrope.net (Kirrily Robert) Date: Fri, 5 Oct 2007 09:33:45 +1000 Subject: [Developers] Whining about the # sign In-Reply-To: References: Message-ID: On 10/5/07, Jason Douglas wrote: > Hang in there... a fix is on the works. :-) Thanks! *happy dance* K. -- Kirrily Robert skud at infotrope.net http://infotrope.net From ksrinivs at gmail.com Fri Oct 5 03:01:44 2007 From: ksrinivs at gmail.com (Kavitha Srinivas) Date: Thu, 4 Oct 2007 23:01:44 -0400 Subject: [Developers] Question about the backend software and data In-Reply-To: <47036376.30701@gmail.com> References: <46FCC9CC.8080007@chello.at> <46FD6A50.6080807@metaweb.com> <46FD6CD8.3030206@thefirst.org> <46FD70B3.8080802@metaweb.com> <46FD7C20.3000400@thefirst.org> <46FDAA27.90101@metaweb.com> <02770076-54BC-4D1B-9F0D-DEE024864196@metaweb.com> <47000610.908@gmail.com> <47036376.30701@gmail.com> Message-ID: Shawn This is really very good. I had a couple of suggestions: (a) Avoid usage of functional properties for integrity constraints and use them instead for when you'd like to discover something. For instance, let's assume that you marriedTo is a functional property, so that if the same person A is married to two people B and C then you want to infer B and C are the same person. (b) Another pattern you've used is to model domain and ranges for all properties of a Person (e.g., Place of Birth, etc). This is fine, but I don't know if you want to capture the fact that any Person always has a Place of Birth (but we may not necessarily know what that place might be). That is, Person can be modeled as an existential restriction (exists.PlaceOfBirth{Location}). This latter form of modeling is a commonly adopted pattern in many large life sciences ontologies. Kavitha On Oct 3, 2007, at 5:40 AM, Shawn Simister wrote: > I've just put together a first attempt at representing the Freebase > type system as OWL ontologies. There's still some work to be done > but I figured I'd give you guys an update and hopefully get some > feedback on how I'm modeling these domains in OWL. > > You can download it at http://www.simister.ca/temp/freebase- > owl-03-10-07.zip. The way that it's set up right now is one OWL > ontology per file per Freebase domain. Let me know what you think. > > Shawn > > Kavitha Srinivas wrote: >> Not really, am in interested in modeling issues rather than a >> specific domain. Thanks! >> Kavitha >> >> On Sep 30, 2007, at 4:24 PM, Shawn Simister wrote: >> >>> I'm just putting together a new version of the ontology now which >>> I will make available on my website. Is there a particular domain >>> that you want to see? Right now I'm just focusing on the 67 or so >>> main Freebase domains. >>> >>> Shawn >>> >>> Kavitha Srinivas wrote: >>>> Is Shawn's OWL ontology available somewhere? >>>> Kavitha >>>> >>>> On Sep 28, 2007, at 9:52 PM, John Giannandrea wrote: >>>> >>>> >>>>> Kirrily Robert wrote: >>>>> >>>>>> To me, RDF sounds like it meets those criteria. >>>>>> >>>>> I agree that an RDF dump of the data makes sense. >>>>> There are important details though, like how to represent >>>>> the schema. Shawn Simister previously posted about >>>>> creating an OWL ontology from the freebase types, rather >>>>> than using RDF Schema, which seemed sensible. >>>>> >>>>> The key point was that Shawn was able to extract the data >>>>> and turn it into those formats, using HTTP, MQL and JSON. >>>>> >>>>> -jg >>>>> >>>>> _______________________________________________ >>>>> Developers mailing list >>>>> Developers at freebase.com >>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>> >>>> _______________________________________________ >>>> Developers mailing list >>>> Developers at freebase.com >>>> http://lists.freebase.com/mailman/listinfo/developers >>>> >>>> >>> >> > From narphorium at gmail.com Fri Oct 5 05:36:17 2007 From: narphorium at gmail.com (Shawn Simister) Date: Fri, 05 Oct 2007 01:36:17 -0400 Subject: [Developers] Question about the backend software and data In-Reply-To: References: <46FCC9CC.8080007@chello.at> <46FD6A50.6080807@metaweb.com> <46FD6CD8.3030206@thefirst.org> <46FD70B3.8080802@metaweb.com> <46FD7C20.3000400@thefirst.org> <46FDAA27.90101@metaweb.com> <02770076-54BC-4D1B-9F0D-DEE024864196@metaweb.com> <47000610.908@gmail.com> <47036376.30701@gmail.com> Message-ID: <4705CD51.8080603@gmail.com> Thanks for the feedback. I can see what you mean about the functional properties so I'll switch those over to cardinality restrictions. With respect to placing existential restrictions on certain properties, I don't think that that will be feasible for a couple of reasons. First of all, I'm not aware of any such restrictions in the Freebase type system (although it might be nice to have them), so if I were to add these sorts of restrictions to ontologies there would be no guarantee that future RDF dumps of the Freebase data would actually conform to those ontologies. Secondly, due to the fact that there are no explicit existential restrictions on any of the Freebase properties I would have to enter these restrictions by hand which is not easy given the growing number of types within Freebase. Even if someone where to model all these restrictions by hand, maintaining them over time as the types change would be a challenge. One other issue that I have just noticed is that some domains have multiple types which have properties with the same names. For example, In the /measurement_unit domain, both DatedInteger and CountAsOfDate have a property named number. This causes a naming collision in some of the OWL ontologies. My first instinct was to simply merge the two properties into a single property which shares the domains and ranges of both properties. The down side to this is that we would be making a naive assumption about the similarity of the properties being merged. An alternate approach would be to model each Freebase type as a separate OWL file so that each type has a unique URI. I think that this approach is probably safer in the long run but I'd appreciate any input on what people think is a better design. Shawn Kavitha Srinivas wrote: > Shawn > This is really very good. I had a couple of suggestions: > (a) Avoid usage of functional properties for integrity constraints and > use them instead for when you'd like to discover something. For > instance, let's assume that you marriedTo is a functional property, so > that if the same person A is married to two people B and C then > you want to infer B and C are the same person. > (b) Another pattern you've used is to model domain and ranges for all > properties of a Person (e.g., Place of Birth, etc). This is fine, but > I don't know if you want to capture the fact that any Person always > has a Place of Birth (but we may not necessarily know what that place > might be). That is, Person can be modeled as an existential > restriction (exists.PlaceOfBirth{Location}). This latter form of > modeling is a commonly adopted pattern in many large life sciences > ontologies. > > Kavitha > > On Oct 3, 2007, at 5:40 AM, Shawn Simister wrote: > >> I've just put together a first attempt at representing the Freebase >> type system as OWL ontologies. There's still some work to be done but >> I figured I'd give you guys an update and hopefully get some feedback >> on how I'm modeling these domains in OWL. >> >> You can download it at >> http://www.simister.ca/temp/freebase-owl-03-10-07.zip. The way that >> it's set up right now is one OWL ontology per file per Freebase >> domain. Let me know what you think. >> >> Shawn >> >> Kavitha Srinivas wrote: >>> Not really, am in interested in modeling issues rather than a >>> specific domain. Thanks! >>> Kavitha >>> >>> On Sep 30, 2007, at 4:24 PM, Shawn Simister wrote: >>> >>>> I'm just putting together a new version of the ontology now which I >>>> will make available on my website. Is there a particular domain >>>> that you want to see? Right now I'm just focusing on the 67 or so >>>> main Freebase domains. >>>> >>>> Shawn >>>> >>>> Kavitha Srinivas wrote: >>>>> Is Shawn's OWL ontology available somewhere? >>>>> Kavitha >>>>> >>>>> On Sep 28, 2007, at 9:52 PM, John Giannandrea wrote: >>>>> >>>>> >>>>>> Kirrily Robert wrote: >>>>>> >>>>>>> To me, RDF sounds like it meets those criteria. >>>>>>> >>>>>> I agree that an RDF dump of the data makes sense. >>>>>> There are important details though, like how to represent >>>>>> the schema. Shawn Simister previously posted about >>>>>> creating an OWL ontology from the freebase types, rather >>>>>> than using RDF Schema, which seemed sensible. >>>>>> >>>>>> The key point was that Shawn was able to extract the data >>>>>> and turn it into those formats, using HTTP, MQL and JSON. >>>>>> >>>>>> -jg >>>>>> >>>>>> _______________________________________________ >>>>>> Developers mailing list >>>>>> Developers at freebase.com >>>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>>> >>>>> _______________________________________________ >>>>> Developers mailing list >>>>> Developers at freebase.com >>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>> >>>>> >>>> >>> >> > > From ksrinivs at gmail.com Fri Oct 5 13:19:20 2007 From: ksrinivs at gmail.com (Kavitha Srinivas) Date: Fri, 5 Oct 2007 09:19:20 -0400 Subject: [Developers] Question about the backend software and data In-Reply-To: <4705CD51.8080603@gmail.com> References: <46FCC9CC.8080007@chello.at> <46FD6A50.6080807@metaweb.com> <46FD6CD8.3030206@thefirst.org> <46FD70B3.8080802@metaweb.com> <46FD7C20.3000400@thefirst.org> <46FDAA27.90101@metaweb.com> <02770076-54BC-4D1B-9F0D-DEE024864196@metaweb.com> <47000610.908@gmail.com> <47036376.30701@gmail.com> <4705CD51.8080603@gmail.com> Message-ID: <461F146C-0859-451D-A743-980D59550346@gmail.com> Hi Shawn I don't think I was clear about what I was proposing in my previous message. Here's what I was saying: 1. Avoid functional properties and cardinality restrictions (cardinality restrictions are a generalized form of functional properties), unless you want to be able to make the sorts of inferences I described earlier. 2. I was proposing removing the domain/range restrictions because they are global restrictions (precisely for the reasons you describe, that is, many types may share the same property, so domain and range applied globally is inappropriate). Instead, I was suggesting modeling the same link between properties and types as existential restrictions. Concretely as an example from your people.owl file, we have a type of modeling [A]: What I am proposing is an alternative type of modeling [B]: Person is a subclass of \exists.place_of_birth{Location} AND \exists.nationality{Country} Logically, [B] is saying that if you are a person, you must have a place of birth property that has a type of Location and nationality property that has a type of Country (although the specific instance may not actually be specified in the data). Using local constraints like those in [B], one can avoid the problem in [A] of specifying the constraints globally. As you noted, [A] poses a problem when several types share the same property. This is why a lot of life sciences OWL models choose [B]. Note, that [B] could be automatically generated from freebase, just the same way you are generating [A]. Your point about automatically generating models from Freebase is well taken. I agree with that completely. Kavitha On Oct 5, 2007, at 1:36 AM, Shawn Simister wrote: > Thanks for the feedback. I can see what you mean about the > functional properties so I'll switch those over to cardinality > restrictions. With respect to placing existential restrictions on > certain properties, I don't think that that will be feasible for a > couple of reasons. > > First of all, I'm not aware of any such restrictions in the > Freebase type system (although it might be nice to have them), so > if I were to add these sorts of restrictions to ontologies there > would be no guarantee that future RDF dumps of the Freebase data > would actually conform to those ontologies. > > Secondly, due to the fact that there are no explicit existential > restrictions on any of the Freebase properties I would have to > enter these restrictions by hand which is not easy given the > growing number of types within Freebase. Even if someone where to > model all these restrictions by hand, maintaining them over time as > the types change would be a challenge. > > One other issue that I have just noticed is that some domains have > multiple types which have properties with the same names. For > example, In the /measurement_unit domain, both DatedInteger and > CountAsOfDate have a property named number. This causes a naming > collision in some of the OWL ontologies. My first instinct was to > simply merge the two properties into a single property which shares > the domains and ranges of both properties. The down side to this > is that we would be making a naive assumption about the similarity > of the properties being merged. > > An alternate approach would be to model each Freebase type as a > separate OWL file so that each type has a unique URI. I think that > this approach is probably safer in the long run but I'd appreciate > any input on what people think is a better design. > > Shawn > > Kavitha Srinivas wrote: >> Shawn >> This is really very good. I had a couple of suggestions: >> (a) Avoid usage of functional properties for integrity constraints >> and use them instead for when you'd like to discover something. >> For instance, let's assume that you marriedTo is a functional >> property, so that if the same person A is married to two people B >> and C then >> you want to infer B and C are the same person. >> (b) Another pattern you've used is to model domain and ranges for >> all properties of a Person (e.g., Place of Birth, etc). This is >> fine, but I don't know if you want to capture the fact that any >> Person always has a Place of Birth (but we may not necessarily >> know what that place might be). That is, Person can be modeled as >> an existential restriction (exists.PlaceOfBirth{Location}). This >> latter form of modeling is a commonly adopted pattern in many >> large life sciences ontologies. >> >> Kavitha >> >> On Oct 3, 2007, at 5:40 AM, Shawn Simister wrote: >> >>> I've just put together a first attempt at representing the >>> Freebase type system as OWL ontologies. There's still some work >>> to be done but I figured I'd give you guys an update and >>> hopefully get some feedback on how I'm modeling these domains in >>> OWL. >>> >>> You can download it at http://www.simister.ca/temp/freebase- >>> owl-03-10-07.zip. The way that it's set up right now is one OWL >>> ontology per file per Freebase domain. Let me know what you think. >>> >>> Shawn >>> >>> Kavitha Srinivas wrote: >>>> Not really, am in interested in modeling issues rather than a >>>> specific domain. Thanks! >>>> Kavitha >>>> >>>> On Sep 30, 2007, at 4:24 PM, Shawn Simister wrote: >>>> >>>>> I'm just putting together a new version of the ontology now >>>>> which I will make available on my website. Is there a >>>>> particular domain that you want to see? Right now I'm just >>>>> focusing on the 67 or so main Freebase domains. >>>>> >>>>> Shawn >>>>> >>>>> Kavitha Srinivas wrote: >>>>>> Is Shawn's OWL ontology available somewhere? >>>>>> Kavitha >>>>>> >>>>>> On Sep 28, 2007, at 9:52 PM, John Giannandrea wrote: >>>>>> >>>>>> >>>>>>> Kirrily Robert wrote: >>>>>>> >>>>>>>> To me, RDF sounds like it meets those criteria. >>>>>>>> >>>>>>> I agree that an RDF dump of the data makes sense. >>>>>>> There are important details though, like how to represent >>>>>>> the schema. Shawn Simister previously posted about >>>>>>> creating an OWL ontology from the freebase types, rather >>>>>>> than using RDF Schema, which seemed sensible. >>>>>>> >>>>>>> The key point was that Shawn was able to extract the data >>>>>>> and turn it into those formats, using HTTP, MQL and JSON. >>>>>>> >>>>>>> -jg >>>>>>> >>>>>>> _______________________________________________ >>>>>>> Developers mailing list >>>>>>> Developers at freebase.com >>>>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>>>> >>>>>> _______________________________________________ >>>>>> Developers mailing list >>>>>> Developers at freebase.com >>>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>>> >>>>>> >>>>> >>>> >>> >> >> > From narphorium at gmail.com Fri Oct 5 15:59:10 2007 From: narphorium at gmail.com (Shawn Simister) Date: Fri, 05 Oct 2007 11:59:10 -0400 Subject: [Developers] Question about the backend software and data In-Reply-To: <461F146C-0859-451D-A743-980D59550346@gmail.com> References: <46FCC9CC.8080007@chello.at> <46FD6A50.6080807@metaweb.com> <46FD6CD8.3030206@thefirst.org> <46FD70B3.8080802@metaweb.com> <46FD7C20.3000400@thefirst.org> <46FDAA27.90101@metaweb.com> <02770076-54BC-4D1B-9F0D-DEE024864196@metaweb.com> <47000610.908@gmail.com> <47036376.30701@gmail.com> <4705CD51.8080603@gmail.com> <461F146C-0859-451D-A743-980D59550346@gmail.com> Message-ID: <47065F4E.9020003@gmail.com> So just to clarify what you are suggesting. Does this refactored person ontology illustrate the changes that you have described below? Kavitha Srinivas wrote: > Hi Shawn > I don't think I was clear about what I was proposing in my previous > message. Here's what I was saying: > > 1. Avoid functional properties and cardinality restrictions > (cardinality restrictions are a generalized form of functional > properties), unless you want to be able to make the sorts of > inferences I described earlier. > 2. I was proposing removing the domain/range restrictions because > they are global restrictions (precisely for the reasons you describe, > that is, many types may share the same property, so domain and range > applied globally is inappropriate). Instead, I was suggesting > modeling the same link between properties and types as existential > restrictions. > > Concretely as an example from your people.owl file, we have a type of > modeling [A]: > > > > rdf:resource="http://www.freebase.com/owl/location#Location"/> > > > > > rdf:resource="http://www.freebase.com/owl/location#Country"/> > > > What I am proposing is an alternative type of modeling [B]: > Person is a subclass of \exists.place_of_birth{Location} AND > \exists.nationality{Country} > > Logically, [B] is saying that if you are a person, you must have a > place of birth property that has a type of Location and nationality > property that has a type of Country (although the specific instance > may not actually be specified in the data). Using local constraints > like those in [B], one can avoid the problem in [A] of specifying the > constraints globally. As you noted, [A] poses a problem when several > types share the same property. This is why a lot of life sciences OWL > models choose [B]. Note, that [B] could be automatically generated > from freebase, just the same way you are generating [A]. Your point > about automatically generating models from Freebase is well taken. I > agree with that completely. > > Kavitha > > > On Oct 5, 2007, at 1:36 AM, Shawn Simister wrote: > >> Thanks for the feedback. I can see what you mean about the functional >> properties so I'll switch those over to cardinality restrictions. >> With respect to placing existential restrictions on certain >> properties, I don't think that that will be feasible for a couple of >> reasons. >> >> First of all, I'm not aware of any such restrictions in the Freebase >> type system (although it might be nice to have them), so if I were to >> add these sorts of restrictions to ontologies there would be no >> guarantee that future RDF dumps of the Freebase data would actually >> conform to those ontologies. >> >> Secondly, due to the fact that there are no explicit existential >> restrictions on any of the Freebase properties I would have to enter >> these restrictions by hand which is not easy given the growing number >> of types within Freebase. Even if someone where to model all these >> restrictions by hand, maintaining them over time as the types change >> would be a challenge. >> >> One other issue that I have just noticed is that some domains have >> multiple types which have properties with the same names. For >> example, In the /measurement_unit domain, both DatedInteger and >> CountAsOfDate have a property named number. This causes a naming >> collision in some of the OWL ontologies. My first instinct was to >> simply merge the two properties into a single property which shares >> the domains and ranges of both properties. The down side to this is >> that we would be making a naive assumption about the similarity of >> the properties being merged. >> >> An alternate approach would be to model each Freebase type as a >> separate OWL file so that each type has a unique URI. I think that >> this approach is probably safer in the long run but I'd appreciate >> any input on what people think is a better design. >> >> Shawn >> >> Kavitha Srinivas wrote: >>> Shawn >>> This is really very good. I had a couple of suggestions: >>> (a) Avoid usage of functional properties for integrity constraints >>> and use them instead for when you'd like to discover something. For >>> instance, let's assume that you marriedTo is a functional property, >>> so that if the same person A is married to two people B and C then >>> you want to infer B and C are the same person. >>> (b) Another pattern you've used is to model domain and ranges for >>> all properties of a Person (e.g., Place of Birth, etc). This is >>> fine, but I don't know if you want to capture the fact that any >>> Person always has a Place of Birth (but we may not necessarily know >>> what that place might be). That is, Person can be modeled as an >>> existential restriction (exists.PlaceOfBirth{Location}). This >>> latter form of modeling is a commonly adopted pattern in many large >>> life sciences ontologies. >>> >>> Kavitha >>> >>> On Oct 3, 2007, at 5:40 AM, Shawn Simister wrote: >>> >>>> I've just put together a first attempt at representing the Freebase >>>> type system as OWL ontologies. There's still some work to be done >>>> but I figured I'd give you guys an update and hopefully get some >>>> feedback on how I'm modeling these domains in OWL. >>>> >>>> You can download it at >>>> http://www.simister.ca/temp/freebase-owl-03-10-07.zip. The way that >>>> it's set up right now is one OWL ontology per file per Freebase >>>> domain. Let me know what you think. >>>> >>>> Shawn >>>> >>>> Kavitha Srinivas wrote: >>>>> Not really, am in interested in modeling issues rather than a >>>>> specific domain. Thanks! >>>>> Kavitha >>>>> >>>>> On Sep 30, 2007, at 4:24 PM, Shawn Simister wrote: >>>>> >>>>>> I'm just putting together a new version of the ontology now which >>>>>> I will make available on my website. Is there a particular domain >>>>>> that you want to see? Right now I'm just focusing on the 67 or so >>>>>> main Freebase domains. >>>>>> >>>>>> Shawn >>>>>> >>>>>> Kavitha Srinivas wrote: >>>>>>> Is Shawn's OWL ontology available somewhere? >>>>>>> Kavitha >>>>>>> >>>>>>> On Sep 28, 2007, at 9:52 PM, John Giannandrea wrote: >>>>>>> >>>>>>> >>>>>>>> Kirrily Robert wrote: >>>>>>>> >>>>>>>>> To me, RDF sounds like it meets those criteria. >>>>>>>>> >>>>>>>> I agree that an RDF dump of the data makes sense. >>>>>>>> There are important details though, like how to represent >>>>>>>> the schema. Shawn Simister previously posted about >>>>>>>> creating an OWL ontology from the freebase types, rather >>>>>>>> than using RDF Schema, which seemed sensible. >>>>>>>> >>>>>>>> The key point was that Shawn was able to extract the data >>>>>>>> and turn it into those formats, using HTTP, MQL and JSON. >>>>>>>> >>>>>>>> -jg >>>>>>>> >>>>>>>> _______________________________________________ >>>>>>>> Developers mailing list >>>>>>>> Developers at freebase.com >>>>>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>>>>> >>>>>>> _______________________________________________ >>>>>>> Developers mailing list >>>>>>> Developers at freebase.com >>>>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>>>> >>>>>>> >>>>>> >>>>> >>>> >>> >>> >> > > -------------- next part -------------- A non-text attachment was scrubbed... Name: people.owl Type: text/xml Size: 7388 bytes Desc: not available Url : http://lists.freebase.com/pipermail/developers/attachments/20071005/c2dad6af/attachment-0001.bin From ksrinivs at gmail.com Fri Oct 5 17:17:50 2007 From: ksrinivs at gmail.com (Kavitha Srinivas) Date: Fri, 5 Oct 2007 13:17:50 -0400 Subject: [Developers] Question about the backend software and data In-Reply-To: <47065F4E.9020003@gmail.com> References: <46FCC9CC.8080007@chello.at> <46FD6A50.6080807@metaweb.com> <46FD6CD8.3030206@thefirst.org> <46FD70B3.8080802@metaweb.com> <46FD7C20.3000400@thefirst.org> <46FDAA27.90101@metaweb.com> <02770076-54BC-4D1B-9F0D-DEE024864196@metaweb.com> <47000610.908@gmail.com> <47036376.30701@gmail.com> <4705CD51.8080603@gmail.com> <461F146C-0859-451D-A743-980D59550346@gmail.com> <47065F4E.9020003@gmail.com> Message-ID: <38EDA097-FC97-4C3C-BFB8-41DF725583F2@gmail.com> Yes it does! Kavitha On Oct 5, 2007, at 11:59 AM, Shawn Simister wrote: > So just to clarify what you are suggesting. Does this refactored > person ontology illustrate the changes that you have described below? > > Kavitha Srinivas wrote: >> Hi Shawn >> I don't think I was clear about what I was proposing in my >> previous message. Here's what I was saying: >> >> 1. Avoid functional properties and cardinality restrictions >> (cardinality restrictions are a generalized form of functional >> properties), unless you want to be able to make the sorts of >> inferences I described earlier. >> 2. I was proposing removing the domain/range restrictions because >> they are global restrictions (precisely for the reasons you >> describe, that is, many types may share the same property, so >> domain and range applied globally is inappropriate). Instead, I >> was suggesting modeling the same link between properties and types >> as existential restrictions. >> >> Concretely as an example from your people.owl file, we have a type >> of modeling [A]: >> >> >> >> >> >> >> >> >> >> >> >> What I am proposing is an alternative type of modeling [B]: >> Person is a subclass of \exists.place_of_birth{Location} AND >> \exists.nationality{Country} >> >> Logically, [B] is saying that if you are a person, you must have a >> place of birth property that has a type of Location and >> nationality property that has a type of Country (although the >> specific instance may not actually be specified in the data). >> Using local constraints like those in [B], one can avoid the >> problem in [A] of specifying the constraints globally. As you >> noted, [A] poses a problem when several types share the same >> property. This is why a lot of life sciences OWL models choose >> [B]. Note, that [B] could be automatically generated from >> freebase, just the same way you are generating [A]. Your point >> about automatically generating models from Freebase is well >> taken. I agree with that completely. >> >> Kavitha >> >> >> On Oct 5, 2007, at 1:36 AM, Shawn Simister wrote: >> >>> Thanks for the feedback. I can see what you mean about the >>> functional properties so I'll switch those over to cardinality >>> restrictions. With respect to placing existential restrictions on >>> certain properties, I don't think that that will be feasible for >>> a couple of reasons. >>> >>> First of all, I'm not aware of any such restrictions in the >>> Freebase type system (although it might be nice to have them), so >>> if I were to add these sorts of restrictions to ontologies there >>> would be no guarantee that future RDF dumps of the Freebase data >>> would actually conform to those ontologies. >>> >>> Secondly, due to the fact that there are no explicit existential >>> restrictions on any of the Freebase properties I would have to >>> enter these restrictions by hand which is not easy given the >>> growing number of types within Freebase. Even if someone where to >>> model all these restrictions by hand, maintaining them over time >>> as the types change would be a challenge. >>> >>> One other issue that I have just noticed is that some domains >>> have multiple types which have properties with the same names. >>> For example, In the /measurement_unit domain, both DatedInteger >>> and CountAsOfDate have a property named number. This causes a >>> naming collision in some of the OWL ontologies. My first instinct >>> was to simply merge the two properties into a single property >>> which shares the domains and ranges of both properties. The down >>> side to this is that we would be making a naive assumption about >>> the similarity of the properties being merged. >>> >>> An alternate approach would be to model each Freebase type as a >>> separate OWL file so that each type has a unique URI. I think >>> that this approach is probably safer in the long run but I'd >>> appreciate any input on what people think is a better design. >>> >>> Shawn >>> >>> Kavitha Srinivas wrote: >>>> Shawn >>>> This is really very good. I had a couple of suggestions: >>>> (a) Avoid usage of functional properties for integrity >>>> constraints and use them instead for when you'd like to discover >>>> something. For instance, let's assume that you marriedTo is a >>>> functional property, so that if the same person A is married to >>>> two people B and C then >>>> you want to infer B and C are the same person. >>>> (b) Another pattern you've used is to model domain and ranges >>>> for all properties of a Person (e.g., Place of Birth, etc). >>>> This is fine, but I don't know if you want to capture the fact >>>> that any Person always has a Place of Birth (but we may not >>>> necessarily know what that place might be). That is, Person can >>>> be modeled as an existential restriction (exists.PlaceOfBirth >>>> {Location}). This latter form of modeling is a commonly adopted >>>> pattern in many large life sciences ontologies. >>>> >>>> Kavitha >>>> >>>> On Oct 3, 2007, at 5:40 AM, Shawn Simister wrote: >>>> >>>>> I've just put together a first attempt at representing the >>>>> Freebase type system as OWL ontologies. There's still some work >>>>> to be done but I figured I'd give you guys an update and >>>>> hopefully get some feedback on how I'm modeling these domains >>>>> in OWL. >>>>> >>>>> You can download it at http://www.simister.ca/temp/freebase- >>>>> owl-03-10-07.zip. The way that it's set up right now is one OWL >>>>> ontology per file per Freebase domain. Let me know what you think. >>>>> >>>>> Shawn >>>>> >>>>> Kavitha Srinivas wrote: >>>>>> Not really, am in interested in modeling issues rather than a >>>>>> specific domain. Thanks! >>>>>> Kavitha >>>>>> >>>>>> On Sep 30, 2007, at 4:24 PM, Shawn Simister wrote: >>>>>> >>>>>>> I'm just putting together a new version of the ontology now >>>>>>> which I will make available on my website. Is there a >>>>>>> particular domain that you want to see? Right now I'm just >>>>>>> focusing on the 67 or so main Freebase domains. >>>>>>> >>>>>>> Shawn >>>>>>> >>>>>>> Kavitha Srinivas wrote: >>>>>>>> Is Shawn's OWL ontology available somewhere? >>>>>>>> Kavitha >>>>>>>> >>>>>>>> On Sep 28, 2007, at 9:52 PM, John Giannandrea wrote: >>>>>>>> >>>>>>>> >>>>>>>>> Kirrily Robert wrote: >>>>>>>>> >>>>>>>>>> To me, RDF sounds like it meets those criteria. >>>>>>>>>> >>>>>>>>> I agree that an RDF dump of the data makes sense. >>>>>>>>> There are important details though, like how to represent >>>>>>>>> the schema. Shawn Simister previously posted about >>>>>>>>> creating an OWL ontology from the freebase types, rather >>>>>>>>> than using RDF Schema, which seemed sensible. >>>>>>>>> >>>>>>>>> The key point was that Shawn was able to extract the data >>>>>>>>> and turn it into those formats, using HTTP, MQL and JSON. >>>>>>>>> >>>>>>>>> -jg >>>>>>>>> >>>>>>>>> _______________________________________________ >>>>>>>>> Developers mailing list >>>>>>>>> Developers at freebase.com >>>>>>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>>>>>> >>>>>>>> _______________________________________________ >>>>>>>> Developers mailing list >>>>>>>> Developers at freebase.com >>>>>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>>>>> >>>>>>>> >>>>>>> >>>>>> >>>>> >>>> >>>> >>> >> >> > > > > > > > ]> > xmlns="http://www.freebase.com/owl/people/person#" > xmlns:owl="http://www.w3.org/2002/07/owl#" > xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" > xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#" > xmlns:xsd="http://www.w3.org/2001/XMLSchema#"> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > From alecf at metaweb.com Fri Oct 5 22:57:29 2007 From: alecf at metaweb.com (Alec Flett) Date: Fri, 05 Oct 2007 15:57:29 -0700 Subject: [Developers] deprecating /community/discussion_thread/updated Message-ID: <4706C159.3010904@metaweb.com> Hey folks - This change is a slight tweak to the discussion data model. If you're not writing an app that queries /community/discussion_thread, then you can ignore this. In the next week or two, the discuss pages on www.freebase.com will stop updating the /community/discussion_thread/updated when someone posts, and we will probably garden out all the values from the graph. This property is just residual cruft from an earlier data model. The proper thing to do now is to sort all posts in a discussion thread, and pick the 'updated' property for the most recent post. With mql you can do this like this: If you queries looked like this before: {"id": "..." "type": "/community/discussion_thread", "updated": null } And the updated time was in result["updated"] Now it looks like this: {"id": "..." "type": "/community/discussion_thread", "post": {"updated": null, "sort": "-updated", "limit": 1}} And the the result["post"]["updated"] will have the equivalent semantic meaning. Alec From narphorium at gmail.com Fri Oct 5 23:44:48 2007 From: narphorium at gmail.com (Shawn Simister) Date: Fri, 05 Oct 2007 19:44:48 -0400 Subject: [Developers] Question about the backend software and data In-Reply-To: <38EDA097-FC97-4C3C-BFB8-41DF725583F2@gmail.com> References: <46FCC9CC.8080007@chello.at> <46FD6A50.6080807@metaweb.com> <46FD6CD8.3030206@thefirst.org> <46FD70B3.8080802@metaweb.com> <46FD7C20.3000400@thefirst.org> <46FDAA27.90101@metaweb.com> <02770076-54BC-4D1B-9F0D-DEE024864196@metaweb.com> <47000610.908@gmail.com> <47036376.30701@gmail.com> <4705CD51.8080603@gmail.com> <461F146C-0859-451D-A743-980D59550346@gmail.com> <47065F4E.9020003@gmail.com> <38EDA097-FC97-4C3C-BFB8-41DF725583F2@gmail.com> Message-ID: <4706CC70.6000304@gmail.com> Maybe I'm misunderstanding how an OWL reasoner would interpret these assertions but to me there seems to be a significant difference in the way this ontology models the data and the way that it is represented in Freebase. For example, in Freebase a Person object can have one or no place_of_birth properties and that property must be of the type Location. In the ontology that you have proposed I would interpret that same relationship as: a Person can have any number of place_of_birth properties as long as one of them is of type Location. I don't think that these two definitions are compatible. Shawn Kavitha Srinivas wrote: > Yes it does! > Kavitha > > On Oct 5, 2007, at 11:59 AM, Shawn Simister wrote: > >> So just to clarify what you are suggesting. Does this refactored >> person ontology illustrate the changes that you have described below? >> >> Kavitha Srinivas wrote: >>> Hi Shawn >>> I don't think I was clear about what I was proposing in my >>> previous message. Here's what I was saying: >>> >>> 1. Avoid functional properties and cardinality restrictions >>> (cardinality restrictions are a generalized form of functional >>> properties), unless you want to be able to make the sorts of >>> inferences I described earlier. >>> 2. I was proposing removing the domain/range restrictions because >>> they are global restrictions (precisely for the reasons you >>> describe, that is, many types may share the same property, so domain >>> and range applied globally is inappropriate). Instead, I was >>> suggesting modeling the same link between properties and types as >>> existential restrictions. >>> >>> Concretely as an example from your people.owl file, we have a type >>> of modeling [A]: >>> >>> >>> >>> >> rdf:resource="http://www.freebase.com/owl/location#Location"/> >>> >>> >>> >>> >>> >> rdf:resource="http://www.freebase.com/owl/location#Country"/> >>> >>> >>> What I am proposing is an alternative type of modeling [B]: >>> Person is a subclass of \exists.place_of_birth{Location} AND >>> \exists.nationality{Country} >>> >>> Logically, [B] is saying that if you are a person, you must have a >>> place of birth property that has a type of Location and nationality >>> property that has a type of Country (although the specific instance >>> may not actually be specified in the data). Using local constraints >>> like those in [B], one can avoid the problem in [A] of specifying >>> the constraints globally. As you noted, [A] poses a problem when >>> several types share the same property. This is why a lot of life >>> sciences OWL models choose [B]. Note, that [B] could be >>> automatically generated from freebase, just the same way you are >>> generating [A]. Your point about automatically generating models >>> from Freebase is well taken. I agree with that completely. >>> >>> Kavitha >>> >>> >>> On Oct 5, 2007, at 1:36 AM, Shawn Simister wrote: >>> >>>> Thanks for the feedback. I can see what you mean about the >>>> functional properties so I'll switch those over to cardinality >>>> restrictions. With respect to placing existential restrictions on >>>> certain properties, I don't think that that will be feasible for a >>>> couple of reasons. >>>> >>>> First of all, I'm not aware of any such restrictions in the >>>> Freebase type system (although it might be nice to have them), so >>>> if I were to add these sorts of restrictions to ontologies there >>>> would be no guarantee that future RDF dumps of the Freebase data >>>> would actually conform to those ontologies. >>>> >>>> Secondly, due to the fact that there are no explicit existential >>>> restrictions on any of the Freebase properties I would have to >>>> enter these restrictions by hand which is not easy given the >>>> growing number of types within Freebase. Even if someone where to >>>> model all these restrictions by hand, maintaining them over time as >>>> the types change would be a challenge. >>>> >>>> One other issue that I have just noticed is that some domains have >>>> multiple types which have properties with the same names. For >>>> example, In the /measurement_unit domain, both DatedInteger and >>>> CountAsOfDate have a property named number. This causes a naming >>>> collision in some of the OWL ontologies. My first instinct was to >>>> simply merge the two properties into a single property which shares >>>> the domains and ranges of both properties. The down side to this >>>> is that we would be making a naive assumption about the similarity >>>> of the properties being merged. >>>> >>>> An alternate approach would be to model each Freebase type as a >>>> separate OWL file so that each type has a unique URI. I think that >>>> this approach is probably safer in the long run but I'd appreciate >>>> any input on what people think is a better design. >>>> >>>> Shawn >>>> >>>> Kavitha Srinivas wrote: >>>>> Shawn >>>>> This is really very good. I had a couple of suggestions: >>>>> (a) Avoid usage of functional properties for integrity constraints >>>>> and use them instead for when you'd like to discover something. >>>>> For instance, let's assume that you marriedTo is a functional >>>>> property, so that if the same person A is married to two people B >>>>> and C then >>>>> you want to infer B and C are the same person. >>>>> (b) Another pattern you've used is to model domain and ranges for >>>>> all properties of a Person (e.g., Place of Birth, etc). This is >>>>> fine, but I don't know if you want to capture the fact that any >>>>> Person always has a Place of Birth (but we may not necessarily >>>>> know what that place might be). That is, Person can be modeled as >>>>> an existential restriction (exists.PlaceOfBirth{Location}). This >>>>> latter form of modeling is a commonly adopted pattern in many >>>>> large life sciences ontologies. >>>>> >>>>> Kavitha >>>>> >>>>> On Oct 3, 2007, at 5:40 AM, Shawn Simister wrote: >>>>> >>>>>> I've just put together a first attempt at representing the >>>>>> Freebase type system as OWL ontologies. There's still some work >>>>>> to be done but I figured I'd give you guys an update and >>>>>> hopefully get some feedback on how I'm modeling these domains in >>>>>> OWL. >>>>>> >>>>>> You can download it at >>>>>> http://www.simister.ca/temp/freebase-owl-03-10-07.zip. The way >>>>>> that it's set up right now is one OWL ontology per file per >>>>>> Freebase domain. Let me know what you think. >>>>>> >>>>>> Shawn >>>>>> >>>>>> Kavitha Srinivas wrote: >>>>>>> Not really, am in interested in modeling issues rather than a >>>>>>> specific domain. Thanks! >>>>>>> Kavitha >>>>>>> >>>>>>> On Sep 30, 2007, at 4:24 PM, Shawn Simister wrote: >>>>>>> >>>>>>>> I'm just putting together a new version of the ontology now >>>>>>>> which I will make available on my website. Is there a >>>>>>>> particular domain that you want to see? Right now I'm just >>>>>>>> focusing on the 67 or so main Freebase domains. >>>>>>>> >>>>>>>> Shawn >>>>>>>> >>>>>>>> Kavitha Srinivas wrote: >>>>>>>>> Is Shawn's OWL ontology available somewhere? >>>>>>>>> Kavitha >>>>>>>>> >>>>>>>>> On Sep 28, 2007, at 9:52 PM, John Giannandrea wrote: >>>>>>>>> >>>>>>>>> >>>>>>>>>> Kirrily Robert wrote: >>>>>>>>>> >>>>>>>>>>> To me, RDF sounds like it meets those criteria. >>>>>>>>>>> >>>>>>>>>> I agree that an RDF dump of the data makes sense. >>>>>>>>>> There are important details though, like how to represent >>>>>>>>>> the schema. Shawn Simister previously posted about >>>>>>>>>> creating an OWL ontology from the freebase types, rather >>>>>>>>>> than using RDF Schema, which seemed sensible. >>>>>>>>>> >>>>>>>>>> The key point was that Shawn was able to extract the data >>>>>>>>>> and turn it into those formats, using HTTP, MQL and JSON. >>>>>>>>>> >>>>>>>>>> -jg >>>>>>>>>> >>>>>>>>>> _______________________________________________ >>>>>>>>>> Developers mailing list >>>>>>>>>> Developers at freebase.com >>>>>>>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>>>>>>> >>>>>>>>> _______________________________________________ >>>>>>>>> Developers mailing list >>>>>>>>> Developers at freebase.com >>>>>>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>> >>>>>> >>>>> >>>>> >>>> >>> >>> >> >> >> > >> >> >> >> ]> >> > xmlns="http://www.freebase.com/owl/people/person#" >> xmlns:owl="http://www.w3.org/2002/07/owl#" >> xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" >> xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#" >> xmlns:xsd="http://www.w3.org/2001/XMLSchema#"> >> >> >> > rdf:resource="http://www.freebase.com/owl/common/topic#Topic" /> >> >> >> >> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/location/location#Location" /> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/location/country#Country" /> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/religion/religion#Religion" /> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/people/gender#Gender" /> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/freebase/user_profile#UserProfile" >> /> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/business/employment_tenure#EmploymentTenure" >> /> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/common/image#Image" /> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/people/marriage#Marriage" /> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/people/sibling_relationship#SiblingRelationship" >> /> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/education/education#Education" >> /> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/people/profession#Profession" >> /> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/media_common#Quotation" /> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/people/place_lived#PlaceLived" >> /> >> >> >> >> >> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/freebase/person#person" /> >> >> >> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/business/employment_tenure#person" >> /> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/people/marriag#spouse" /> >> >> >> > rdf:resource="http://www.freebase.com/owl/people/sibling_relationship#sibling" >> /> >> >> >> >> >> >> >> >> >> > rdf:resource="http://www.freebase.com/owl/education#student" /> >> >> >> > rdf:resource="http://www.freebase.com/owl/people/profession#people_with_this_profession" >> /> >> >> >> > rdf:resource="http://www.freebase.com/owl/media_common/quotation#author" >> /> >> >> >> >> >> > > From ksrinivs at gmail.com Mon Oct 8 01:52:33 2007 From: ksrinivs at gmail.com (Kavitha Srinivas) Date: Sun, 7 Oct 2007 21:52:33 -0400 Subject: [Developers] Question about the backend software and data In-Reply-To: <4706CC70.6000304@gmail.com> References: <46FCC9CC.8080007@chello.at> <46FD6A50.6080807@metaweb.com> <46FD6CD8.3030206@thefirst.org> <46FD70B3.8080802@metaweb.com> <46FD7C20.3000400@thefirst.org> <46FDAA27.90101@metaweb.com> <02770076-54BC-4D1B-9F0D-DEE024864196@metaweb.com> <47000610.908@gmail.com> <47036376.30701@gmail.com> <4705CD51.8080603@gmail.com> <461F146C-0859-451D-A743-980D59550346@gmail.com> <47065F4E.9020003@gmail.com> <38EDA097-FC97-4C3C-BFB8-41DF725583F2@gmail.com> <4706CC70.6000304@gmail.com> Message-ID: <15E94B4A-4983-4128-801A-ACD29BA6E93E@gmail.com> All an existential restriction says is that a Person has a place of birth whose value is Location. If a person is specified with no actual place of birth value (e.g., that slot is not filled in Freebase), its fine, because OWL reasoners operate under the open world assumption (i.e., every Person has a place of birth, but it may or may not be known yet). Kavitha On Oct 5, 2007, at 7:44 PM, Shawn Simister wrote: > Maybe I'm misunderstanding how an OWL reasoner would interpret > these assertions but to me there seems to be a significant > difference in the way this ontology models the data and the way > that it is represented in Freebase. For example, in Freebase a > Person object can have one or no place_of_birth properties and that > property must be of the type Location. In the ontology that you > have proposed I would interpret that same relationship as: a Person > can have any number of place_of_birth properties as long as one of > them is of type Location. I don't think that these two definitions > are compatible. > > Shawn > > Kavitha Srinivas wrote: >> Yes it does! >> Kavitha >> >> On Oct 5, 2007, at 11:59 AM, Shawn Simister wrote: >> >>> So just to clarify what you are suggesting. Does this refactored >>> person ontology illustrate the changes that you have described >>> below? >>> >>> Kavitha Srinivas wrote: >>>> Hi Shawn >>>> I don't think I was clear about what I was proposing in my >>>> previous message. Here's what I was saying: >>>> >>>> 1. Avoid functional properties and cardinality restrictions >>>> (cardinality restrictions are a generalized form of functional >>>> properties), unless you want to be able to make the sorts of >>>> inferences I described earlier. >>>> 2. I was proposing removing the domain/range restrictions >>>> because they are global restrictions (precisely for the reasons >>>> you describe, that is, many types may share the same property, >>>> so domain and range applied globally is inappropriate). >>>> Instead, I was suggesting modeling the same link between >>>> properties and types as existential restrictions. >>>> >>>> Concretely as an example from your people.owl file, we have a >>>> type of modeling [A]: >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> What I am proposing is an alternative type of modeling [B]: >>>> Person is a subclass of \exists.place_of_birth{Location} AND >>>> \exists.nationality{Country} >>>> >>>> Logically, [B] is saying that if you are a person, you must have >>>> a place of birth property that has a type of Location and >>>> nationality property that has a type of Country (although the >>>> specific instance may not actually be specified in the data). >>>> Using local constraints like those in [B], one can avoid the >>>> problem in [A] of specifying the constraints globally. As you >>>> noted, [A] poses a problem when several types share the same >>>> property. This is why a lot of life sciences OWL models choose >>>> [B]. Note, that [B] could be automatically generated from >>>> freebase, just the same way you are generating [A]. Your point >>>> about automatically generating models from Freebase is well >>>> taken. I agree with that completely. >>>> >>>> Kavitha >>>> >>>> >>>> On Oct 5, 2007, at 1:36 AM, Shawn Simister wrote: >>>> >>>>> Thanks for the feedback. I can see what you mean about the >>>>> functional properties so I'll switch those over to cardinality >>>>> restrictions. With respect to placing existential restrictions >>>>> on certain properties, I don't think that that will be feasible >>>>> for a couple of reasons. >>>>> >>>>> First of all, I'm not aware of any such restrictions in the >>>>> Freebase type system (although it might be nice to have them), >>>>> so if I were to add these sorts of restrictions to ontologies >>>>> there would be no guarantee that future RDF dumps of the >>>>> Freebase data would actually conform to those ontologies. >>>>> >>>>> Secondly, due to the fact that there are no explicit >>>>> existential restrictions on any of the Freebase properties I >>>>> would have to enter these restrictions by hand which is not >>>>> easy given the growing number of types within Freebase. Even if >>>>> someone where to model all these restrictions by hand, >>>>> maintaining them over time as the types change would be a >>>>> challenge. >>>>> >>>>> One other issue that I have just noticed is that some domains >>>>> have multiple types which have properties with the same names. >>>>> For example, In the /measurement_unit domain, both DatedInteger >>>>> and CountAsOfDate have a property named number. This causes a >>>>> naming collision in some of the OWL ontologies. My first >>>>> instinct was to simply merge the two properties into a single >>>>> property which shares the domains and ranges of both >>>>> properties. The down side to this is that we would be making a >>>>> naive assumption about the similarity of the properties being >>>>> merged. >>>>> >>>>> An alternate approach would be to model each Freebase type as a >>>>> separate OWL file so that each type has a unique URI. I think >>>>> that this approach is probably safer in the long run but I'd >>>>> appreciate any input on what people think is a better design. >>>>> >>>>> Shawn >>>>> >>>>> Kavitha Srinivas wrote: >>>>>> Shawn >>>>>> This is really very good. I had a couple of suggestions: >>>>>> (a) Avoid usage of functional properties for integrity >>>>>> constraints and use them instead for when you'd like to >>>>>> discover something. For instance, let's assume that you >>>>>> marriedTo is a functional property, so that if the same person >>>>>> A is married to two people B and C then >>>>>> you want to infer B and C are the same person. >>>>>> (b) Another pattern you've used is to model domain and ranges >>>>>> for all properties of a Person (e.g., Place of Birth, etc). >>>>>> This is fine, but I don't know if you want to capture the fact >>>>>> that any Person always has a Place of Birth (but we may not >>>>>> necessarily know what that place might be). That is, Person >>>>>> can be modeled as an existential restriction >>>>>> (exists.PlaceOfBirth{Location}). This latter form of modeling >>>>>> is a commonly adopted pattern in many large life sciences >>>>>> ontologies. >>>>>> >>>>>> Kavitha >>>>>> >>>>>> On Oct 3, 2007, at 5:40 AM, Shawn Simister wrote: >>>>>> >>>>>>> I've just put together a first attempt at representing the >>>>>>> Freebase type system as OWL ontologies. There's still some >>>>>>> work to be done but I figured I'd give you guys an update and >>>>>>> hopefully get some feedback on how I'm modeling these domains >>>>>>> in OWL. >>>>>>> >>>>>>> You can download it at http://www.simister.ca/temp/freebase- >>>>>>> owl-03-10-07.zip. The way that it's set up right now is one >>>>>>> OWL ontology per file per Freebase domain. Let me know what >>>>>>> you think. >>>>>>> >>>>>>> Shawn >>>>>>> >>>>>>> Kavitha Srinivas wrote: >>>>>>>> Not really, am in interested in modeling issues rather than >>>>>>>> a specific domain. Thanks! >>>>>>>> Kavitha >>>>>>>> >>>>>>>> On Sep 30, 2007, at 4:24 PM, Shawn Simister wrote: >>>>>>>> >>>>>>>>> I'm just putting together a new version of the ontology now >>>>>>>>> which I will make available on my website. Is there a >>>>>>>>> particular domain that you want to see? Right now I'm just >>>>>>>>> focusing on the 67 or so main Freebase domains. >>>>>>>>> >>>>>>>>> Shawn >>>>>>>>> >>>>>>>>> Kavitha Srinivas wrote: >>>>>>>>>> Is Shawn's OWL ontology available somewhere? >>>>>>>>>> Kavitha >>>>>>>>>> >>>>>>>>>> On Sep 28, 2007, at 9:52 PM, John Giannandrea wrote: >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> Kirrily Robert wrote: >>>>>>>>>>> >>>>>>>>>>>> To me, RDF sounds like it meets those criteria. >>>>>>>>>>>> >>>>>>>>>>> I agree that an RDF dump of the data makes sense. >>>>>>>>>>> There are important details though, like how to represent >>>>>>>>>>> the schema. Shawn Simister previously posted about >>>>>>>>>>> creating an OWL ontology from the freebase types, rather >>>>>>>>>>> than using RDF Schema, which seemed sensible. >>>>>>>>>>> >>>>>>>>>>> The key point was that Shawn was able to extract the data >>>>>>>>>>> and turn it into those formats, using HTTP, MQL and JSON. >>>>>>>>>>> >>>>>>>>>>> -jg >>>>>>>>>>> >>>>>>>>>>> _______________________________________________ >>>>>>>>>>> Developers mailing list >>>>>>>>>>> Developers at freebase.com >>>>>>>>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>>>>>>>> >>>>>>>>>> _______________________________________________ >>>>>>>>>> Developers mailing list >>>>>>>>>> Developers at freebase.com >>>>>>>>>> http://lists.freebase.com/mailman/listinfo/developers >>>>>>>>>> >>>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>> >>>>>> >>>>>> >>>>> >>>> >>>> >>> >>> >>> >> >>> >>> >>> >>> ]> >>> >> xmlns="http://www.freebase.com/owl/people/person#" >>> xmlns:owl="http://www.w3.org/2002/07/owl#" >>> xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" >>> xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#" >>> xmlns:xsd="http://www.w3.org/2001/XMLSchema#"> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >> >> > From narphorium at gmail.com Mon Oct 8 06:18:56 2007 From: narphorium at gmail.com (Shawn Simister) Date: Mon, 08 Oct 2007 02:18:56 -0400 Subject: [Developers] Question about the backend software and data In-Reply-To: <15E94B4A-4983-4128-801A-ACD29BA6E93E@gmail.com> References: <46FCC9CC.8080007@chello.at> <46FD6A50.6080807@metaweb.com> <46FD6CD8.3030206@thefirst.org> <46FD70B3.8080802@metaweb.com> <46FD7C20.3000400@thefirst.org> <46FDAA27.90101@metaweb.com> <02770076-54BC-4D1B-9F0D-DEE024864196@metaweb.com>