jamie

Roadmap of future features: Option to require included type

2008-07-25T03:44:38.0012Z

That's an interesting case. We don't support that in the client currently but perhaps it could be a property hint in schema editor. We'll take a look.

Halloumi: is it just me?

2008-07-25T01:06:51.0012Z

Jon, it's not just you, but it's not everyone either. :) From what I understand, the featured topic that shows up on a logged-in user's home page is dynamically generated, and depends on the user's domain membership. I'm not a Food domain member, so (sadly), Halloumi is never a featured topic on my home page.

Sports: Sports Team - generalize City to Location

2008-07-25T00:43:05.0012Z

alright, this has been changed. See sports team

Person: Relationships

2008-07-25T00:25:57.0012Z

Another long delay - I need to remember to login in more than once a month (or figure out how to get RSS feeds set up).

I guess the real answer to whether the accuracy is needed depends on whether Freebase has aspirations to hold genealogical data. The current scheme isn't adequate for anything genealogical, including medical applications which need genealogical information.

I tend to lean towards the "if you're going to do it, you should do it right" camp, but I recognize that that often leads to over engineering. I don't know anything about how Freebase stores things internally, so it's difficult to evaluate where the "quadrupling" comes in and how big an impact it has in real world terms.

On the other hand, no one's addressed the sibling side of the argument. That could clearly be derived from traversing the parent-child graph as well, yet it seems to have been important enough to record separately (and redundantly).

That's enough lobbying for me though. If anyone ever decides to change the decision, ping me and I'll provide pointers to how things are modeled in the genealogical data world.

Person: Relationships

2008-07-25T00:25:57.0012Z

Another long delay - I need to remember to login in more than once a month (or figure out how to get RSS feeds set up).

That's enough lobbying for me though. If anyone ever decides to change the decision, ping me and I'll provide pointers to how things are modeled in the genealogical data world.

Publishing: Possible revision for Journal Articles

2008-07-24T23:20:13.0005Z

Spurred on by an aside that spatialed made in some discussion post awhile back (I can't locate the post), I'm considering revising the way that we model journal issues. Currently, each issue has its own topic, which links to both the journal and the articles contained in that issue. (See
http://www.freebase.com/type/schema/book/journal_issue) The main problems with this format are that it is cumbersome to enter data, and also that most bibliographic sources are concerned primarily with the article and the journal, relegating the issue to a series of strings (volume, issue, date). This latter issue might make integration with standard bibliographic schemas a bit cumbersome, although it wouldn't be insurmountable.

As an experiment, though, I thought I'd try to see what a model that eliminated the issue type entirely looked like. Here are the results:
http://sandbox.freebase.com/view/guid/9202a8c04000641f8000000008cd1edd.

I've replaced the issue type with a CVT that connects the article and the journal, and includes the standard bibliographic data of Volume, issue, date, issue date extra, and pages ("issue date extra" is something I had to make up for journals that aren't published on a schedule that translates into mm/dd/yyyy). Journal articles have both Scholarly Work and Written Work as included types, although a journal article can also be a review, editorial, letter or other type of writing.

The only real disadvantage that I see to this is that constructing the contents of a given issue will be harder -- users will have to query on a combination of several fields (volume, issue, etc.) to find what they're looking for.

I'd love to hear what people think about this. If it seems to work, I might do the same thing for newspaper issues.

This is also being discussed on the data-modelers mailing list: http://lists.freebase.com/pipermail/data-modeling/2008-July/000989.html

International Affairs: Split

2008-07-24T22:49:16.0010Z

The field of study should be split from the journal.

Roadmap of future features: What links here

2008-07-24T22:07:20.0012Z

I like it too -- there have been a many instances where a topic was flagged for deletion because it had no property values of its own, even though it was being linked to by other properties. Having an easy way to discover these links would be a boon to both flaggers and voters on flagged items.

Roadmap of future features: What links here

2008-07-24T19:58:20.0012Z

I like this idea alot... let's give it some thought.

Hockey: Former team

2008-07-24T19:49:40.0005Z

Awesome! Thanks a lot!

Roadmap of future features: Clone one type from another

2008-07-24T19:45:52.0012Z

So, at long last, I think we've got the feature y'all were looking for. See my blog post on delegated properties.

ilive4music: Clear you HF Inbox!

2008-07-24T19:41:31.0017Z

I can't send you PM. Your HF inbox is full.

Hockey: Former team

2008-07-24T19:16:52.0012Z

No problem - its actually a property both on team and player - I was talking about the same link from the other direction. In any event, its there now. See Steve Yzerman.

Hockey: Former team

2008-07-24T18:37:10.0012Z

I was referring to a property of Hockey Player. There is a "current team" property, but not a "former team" property. Sorry, I should have been more clear.

Hockey: Former team

2008-07-24T18:04:10.0012Z

Do you mean a historical roster property on team?
Its just an oversight. I'll try to add that in the next couple of days.

Hockey: Former team

2008-07-24T15:10:56.0017Z

Other sports have a former team property, but hockey doesn't. Is this a function of there being little data to fill in, or is there another reason?

Location: Contains/Contained By

2008-07-23T15:44:01.0000Z

Can you put a limit on this? I always see more than one level on these items and its frustrating. Can you guys just cap that field at 1?

Roadmap of future features: What links here

2008-07-21T16:42:06.0012Z

Since it's not always appropriate to include all the suggested incoming properties in a schema, would it be possible to have a "What links here" page that shows the topics that link to this topic in one of their properties? It'd be useful for when you see an empty topic and wonder why it exists.

Roadmap of future features: Option to require included type

2008-07-18T23:48:42.0013Z

Beyond the ranking suggestion above, people have tried to add bands and tv shows to the influence domain, and I'm not really watching the additions that carefully.

Is there a way to require an included type (person, in this case)? Tv shows and bands aren't the focus for the domain, but beyond saying that in the description, it's hard to monitor.

Halloumi: is it just me?

2008-07-18T22:33:56.0005Z

or is Halloumi the most frequently featured topic in freebase?

Web Ontology: Ontology Type

2008-07-17T00:40:52.0005Z

Would it be possible to add a property to the Ontology type so that you can define equivalent types the same way that you can define equivalent properties?

For example, you could map the foaf:Person class to the /people/person type.

Location: City Parks/Golf Courses

2008-07-09T20:35:32.0000Z

Cool, thanks for the update.

Location: City Parks/Golf Courses

2008-07-09T18:37:54.0015Z

It looks like the general consensus was in favor of not denormalizing the data, but instead having a synthetic property that could represent current data from the historical data. However, this suggestion is currently not something that is possible, and would require development work. I'll check around to see if there are open tasks for this and update you.

Location: City Parks/Golf Courses

2008-07-09T15:41:51.0000Z

Did we ever come to a conclusion about this? It seems to have died without an answer in the Data Modeling discussions.

Bug / Feature Request: Airport ICAO Code

2008-07-09T14:36:33.0000Z

Thanks for the info. I'll wait to add smaller airports until the codes come in -- it should make it a lot easier to keep track of (some airports(some have multiple names and it gets confusing!).

Freebase: Edit a Topic with Redirect Type?

2008-07-08T22:10:50.0030Z

I've removed the type /proficiency from the developers page. There really isn't a way to do this with the UI, so bringing it up here was a good choice.

Bug / Feature Request: Airport ICAO Code

2008-07-08T21:59:57.0044Z

Hi evening,

See http://www.freebase.com/discuss/threads/authority/icao/ESSX#/guid/9202a8c04000641f80000000089b3836 . The long and short of it is that we've updated the property to be an property enumeration with a more restrictive permission. We hope to be loading some more airport data soon, so stay tuned!

Bug / Feature Request: Airport ICAO Code

2008-07-08T20:49:35.0012Z

I can't enter a value for the ICAO field under Airports. When I try to save the value I get a message saying I don't have authorization. Is this a bug, or by design?

an example code is FMMI for Ivato Airport in Madagascar

Freebase: Edit a Topic with Redirect Type?

2008-07-08T20:47:50.0012Z

I accidentally assigned a co-type to the Freebase developers page, and not even on the sandbox :(

However, even trying to access the new instance result in a redirect, so I have no way of removing the type, at least through the GUI.

Freebase: Edit a Topic with Redirect Type?

2008-07-08T19:05:28.0044Z

What are you trying to do? The Redirect type (/freebase/redirect), like most types in the Freebase domain, is a special type, used in the UI, and its instances shouldn't be editable by most users. But there are only four instances of the type, and I assume that you're not trying to edit the home page, so I'm probably missing something.

Freebase: Edit a Topic with Redirect Type?

2008-07-08T02:45:09.0012Z

It doesn't seem like there's any way to edit a topic with the redirect co-type.

I'm guessing there may be some kind of URL param I can enter to avoid the redirect?

TV Episode: Dubbers

2008-07-07T22:27:21.0013Z

Use the "special performance type" column in the "Performances" property on TV Episode. You should be able to select "Voice" for these types of performances.

Human Language: Language family trees

2008-07-07T22:24:18.0028Z

This is a common pattern in Freebase, also seen with location containment (i.e. the "contains" property on the "location" type). Does Australia contain Melbourne University, or does Australia contain Victoria which contains Greater Melbourne which contains Carlton which contains Melbourne University?

The answer is... there is no one true answer. We don't have a one true way, and even if we did people would mess it up because they didn't know our guidelines.

I believe there are people working on ways to query down trees like that so you can say things like "Show me all the locations in Australia" or "all the Afro-Asiatic languages" and have it walk down as deeply as necessary. Here's a link from one of our mailing lists that discusses it: http://freebase.markmail.org/search/?q=transitive#query:transitive%20from%3A%22Robert%20Cook%22+page:1+mid:chub2wx52cep3x4d+state:results

Sports: Sports Team - generalize City to Location

2008-07-06T21:51:58.0012Z

Either generalize the field or add more location fields for those types of teams.

TV Episode: Dubbers

2008-07-06T04:31:18.0013Z

How should I model a specific person who voices a specific other person in a specific TV episode? e.g. Kent Frick voicing Chairman Kaga in the Iron Chef blue crab battle?

Human Language: Language family trees

2008-07-04T17:57:26.0012Z

My question is probably best explained with an example. The Rosetta Project page for the Shekkacho language has this classification for it:

Afro-Asiatic > Omotic > North > Gonga-Gimojan > Gonga > South > Shekkacho

What I'd like to know is whether Shekkacho should have all those things as its Language Family, or whether it should just have South Gonga. Also, should South Gonga itself be a member of Gonga, Gonga-Gimojan, North Omotic, Omotic, and Afro-Asiatic, or just of Gonga?

Educational Institution: Field for founder?

2008-07-03T22:55:31.0005Z

We have a field for the date of foundation, but there isn't one for the founder. e.g. Sir George Downing for Downing College

Would this be a useful addition?

TV: TV filming location

2008-07-02T18:39:23.0012Z

I think we should have a type, "TV filming location", for things like "Survivor: Vanuatu" which was filmed in Vanuatu. Thoughts?

Location: City Parks/Golf Courses

2008-06-24T21:38:56.0012Z

No, no, you're not being obnoxious at all -- this is a persistent pattern. I'm guessing that you're looking for the current head of state of each country, as opposed to the office that the head of state holds. A property for "current head of state" on the country type seems reasonable, but I'm going to repost this to the data-modeling mailing list to get more input because I can see a lot of ways for this to snowball. (If you're not on the mailing list, you can join here: http://lists.freebase.com/mailman/listinfo/data-modeling, or just watch the outcome here: http://lists.freebase.com/pipermail/data-modeling/)

Location: City Parks/Golf Courses

2008-06-24T17:55:30.0012Z

Right, I had seen that type, but when trying to query against the database, it seems more helpful if heads of state were given their own type. For instance, if I wanted to see who the heads of state were in the UK and the US, I currently have to sift through the results to find "President of the United States" and "Prime Minister of the UK". If it were in a separate property, I could just ask for the heads of state of each country and receive only them.

Sorry, not trying to be obnoxious.

Location: City Parks/Golf Courses

2008-06-23T23:42:55.0012Z

Heads of state (and all other people who hold government offices) can be listed on the "governmental jurisdiction" type. Most countries should have this type already, but if they don't, feel free to add it. There's some documentation for entering data about government office-holders here: http://www.freebase.com/view/guid/9202a8c04000641f800000000739446b.

Location: City Parks/Golf Courses

2008-06-23T21:36:06.0012Z

Because Country is a type within Location, I figured I'd ask here: Would it be possible to add "current leader" or "head of state" as a property?

Person: Relationships

2008-06-23T20:03:38.0012Z

It’s less a decision to obscure the facts as a lack of a compelling use case for a more complicated model. The current model is simple and handles 95% of the use cases. To complicate the parent-child relationship—quadrupling the amount of information needed to represent it, in even the simplest cases—really needs a compelling use case. The blurb can tell human consumers that Gerald Ford was adopted. Is there a need for API-based applications to be able to make that distinction?

Person: Relationships

2008-06-23T20:03:38.0012Z

Person: Relationships

2008-06-20T19:52:15.0001Z

I just discovered this reply after many months. Sorry for the delay!

In my opinion the privacy concerns are orthogonal and are a red herring here. The participants can believe whatever they want, but biological parents and adoptive parents are undeniably different relationships.

The decision to obscure the facts which our displayed for living people who aren't public personas shouldn't prevent getting the data accurate in the first place. That's not possible with the current schema.

Take a look a Gerry Ford http://freebase.com/view/en/gerald_ford?pid=%2Fpeople%2Fperson%2Fparents. Leaving aside the fact that his grandfather is also listed as a "parent" (apparently the work of mw_template_bot), how would you accurately model his family using the current schema? How would you tell his adoptive parents from his biological parents? How would you tell which female parent went with which male parent? Would all the half-brothers and half-sisters just get linked together in one big undifferentiated pile of "siblings?"

Person: Relationships

2008-06-20T19:52:15.0001Z

Location: Contains/Contained By

2008-06-20T19:33:14.0000Z

One level is sufficient, Tyler. One of our planned features is transitive property comprehension (like containment and organism classification hierarchy), so there is no need to denormalize the hierarchy completely.

Location: Contains/Contained By

2008-06-20T18:48:46.0005Z

How deep should I go? For instance, I'm working on the Empire State Building. Generally, it is contained by New York. If one clicks to see New York, they'll see that the city is contained the state, and the state by the country, then continent, then earth.

In this sense, should I list that the Emprie State Building is contained by: New York, New York, United States, North American, and Earth?

I haven't been doing this because I think its cleaner to just list the most immediate contained by and followed the path if needed. I guess I'm asking is my way correct or not?

Thanks!

jamie: BR_ESTATE

2008-03-19T20:51:36.0000Z

The sidebar can only display 50 types. To view the other types, click the symbol to the left of the word "Types" at the top of the list, and select "view all".

jamie: BR_ESTATE

2008-03-19T19:39:00.0005Z

Hi Jamie,

Could you add BR_ESTATE on side menu bar at Locations?

http://www.freebase.com/view/location/br_state

I working on it and ask for some friends to help, but without link the job is more complicate.

Tnks

Paulo Colacino