crism

Jefferson Park: same location, please merge

2008-07-09T11:14:11.0005Z

It's just the same station serving CTA and METRA: http://en.wikipedia.org/wiki/index.html?curid=7275620. Thanks!

Foster-Purple Line: same location, please merge

2008-07-09T09:47:31.0005Z

They are just the same location/transit stop of Chicago Transit Authority.

Know You Rider: Multiple Artists

2008-07-09T02:33:02.0021Z

Is there a way to enter multiple artists? It's a song that's labeled "traditional" but has been recorded by the Grateful Dead, Byrds, Joan Baez, Hot Tuna and others.

Kontious Filosofy: Kontious filosofy

2008-07-08T16:26:54.0000Z

he's a okay

Kontious Filosofy: Kontious filosofy

2008-07-08T16:26:37.0005Z

is he a good lyricist

Michael Collins: Blend and single malt

2008-07-06T19:00:58.0012Z

We need a discrete topic for the single malt product.

Cumberland: same location, please merge

2008-07-04T00:42:04.0011Z

They are the same location or transit stop of Chicago Authority Transit.

69th/Red Line: same location, please merge

2008-07-03T23:41:15.0005Z

They are just the same location or transit stop of Chicago Transit Authority.

United States: Removed "Official Language" and Liberty Dollar as form of Currency

2008-07-03T16:55:01.0005Z

Removed "Official Language" as per Wikipedia entry:

http://en.wikipedia.org/wiki/Languages_of_the_United_States

Removed "Liberty Dollar" as it is not a currency in itself, but rather a unit of currency.

Kingdom songs: Name

2008-07-03T15:34:26.0000Z

This does not seem like a good merge to me. “Sing Praises to Jehovah” is a a musical album. This topic”s Wikipedia article is about, roughly, the songs that Jehovah’s Witnesses sing; that is not the same as a specific album. If “Kingdom songs” is not a good title for that kind of song, then the title should be changed, but don’t merge it with the album. (The corresponding Wikipedia article’s title should probably also be changed.)

Kingdom songs: Name

2008-07-03T12:29:09.0012Z

'kingdom songs' is not an official name of, well, anything. The album is "Sing Praises to Jehovah" which is an 8 disc album of 225 tracks, recorded and released by Watch Tower Bible and Track Society of Pennsylvania

Sing Praises to Jehovah: Name

2008-07-03T12:26:31.0005Z

'kingdom songs' is not an official name of, well, anything. The album is "Sing Praises to Jehovah" which is an 8 disc album of 225 tracks.

Awards: Competitions

2008-07-01T17:18:58.0015Z

Ah... I get it now. I was confused because the EuroCup 2008 was typed as a competition! I think just doing it for architecture makes sense. Later, if we discover that there are other types of competitions that use the same pattern, we can decide whether or not to make the types more abstract, but it's much easier to move a model from specific to abstract than the other way around.

I'd recommend dropping a comment in the architecture domain for more comments, since the architecture fans might not be watching the awards discussions.

Awards: Competitions

2008-07-01T05:42:25.0000Z

Uh oh, opening an entire can of worms here!

I created it with Architectural competitions in mind, i.e. where Architects explicitly compete to have their designs used for a building.
I should probably rename it as 'Architectural Competitions' and keep it focused for the moment.

Awards: Competitions

2008-06-30T19:31:15.0012Z

How are you intending it to be used? As a catch-all for all types of competitions, including sporting events, reality television shows, beauty pageants, spelling bees, poetry slams, etc.? If you're considering including sports (and it looks like you are), there are currently models for sports league championships (http://www.freebase.com/tools/schema/sports/sports_championship) and cricket matches (http://www.freebase.com/view/cricket/cricket_match), which should be taken into account. I'm personally inclined to let sporting events stay in the sports domains, since the semantics may vary between sports.

Awards: Should awards given by a government be seperated?

2008-06-30T19:16:32.0012Z

It seems to me that the current awards model fits military medals and most of the British Empire medals pretty well; I think "military", "civilian", etc. could fit in the "disciplines or subjects" field, although maybe we need to rename it to accommodate that better. I'm not sure that a property for the honorific title that an award imbues is necessary: only a small subset of awards have this attribute; if we need to add it, maybe a new type (that has "Award Category" as an included type) would be best.

In terms of outright knighthoods, I'm not sure how that should be modeled -- in modern times, it's more-or-less an award, but historically it's tied into notions of peerage (IIRC) which are rather a different thing.

Awards: Competitions

2008-06-30T08:45:07.0005Z

I have started putting together a domain for competition, which would best fit in as an extension of this domain. Please let me know your comments.

Awards: Should awards given by a government be seperated?

2008-06-29T10:37:05.0000Z

I think a bit of work is needed on this.
For example, in the UK, awards presented by the Queen usually come with an honorific title/initials (Sir, MBE, OBE etc..). Currently I don't think there is a way of properly capturing honorific awards in either this or the people domain.

Awards: Listed Awards

2008-06-29T10:26:56.0000Z

Perhaps Fortune 500 was a poor example and there is perhaps a better place for this (a published lists type?), but for lists with a complex and/or subjective judging criteria - UK University Ratings etc.

jeff, I agree, I think a seperate type such as "Award Finalist" rather than Award Winner would be more appropriate - I will get to work on that.

Political party: Leadership Property

2008-06-27T22:49:24.0012Z

Good suggestion. We probably need a leadership CVT, to hold person, party, title, and dates of office. Alternately, we could add the type "Employer" as an included type, and use that instead, but I'm not sure how accurate that is -- I imagine a lot of political parties don't pay their leadership.

Political party: Leadership Property

2008-06-27T14:17:56.0012Z

Can you add a property for leadership which would point to a person?

Government Agency: Funding?

2008-06-25T20:09:46.0000Z

I think what Jeff is suggesting is right on the money, no pun intended.

Composition: Date Created

2008-06-25T17:11:53.0012Z

Good suggestions. I’ve made these changes over on sandbox for comment; I’ll recreate them here if folks like them.

Awards: Should awards given by a government be seperated?

2008-06-25T15:27:21.0017Z

I don't have strong feelings about this but it would be handy to have a country field and whether the award/knighthood/ribbon is civilian or miltary. Any thoughts?

Composition: Date Created

2008-06-25T11:09:10.0015Z

Could we have a Date Created (type Date/Time) property?

Also a First Played (type Date/Time) property would be good. Possibly extend this to a compound type including a first played (type Location) property?

Government Agency: Funding?

2008-06-24T21:21:11.0012Z

I think that a relatively simple CVT as Jeff proposes would suffice to say things like, “The FY 2007 budget of the Department of Homeland Security was G$44.9.” Sources and breakdowns could be added to that CVT later if there is real demand.

Awards: Listed Awards

2008-06-24T20:36:03.0000Z

The "Listed Award Honor" type includes a rank property, I think is the main difference. A few thoughts:

Rather than using the existing "Award Winner" and "Award-winning work" types, I think new types might be better.

Is "award" the right thing to call these? The Fortune 500 isn't really an award, it's just an ordered list of companies by gross annual income.

Awards: Listed Awards

2008-06-24T20:00:53.0012Z

I may be missing something… in what way is Listed Award Category different from Award Category? I see you have typed a few instances like the Fortune 500. How would you model the actual list of the 2006 Fortune 500 companies?

Government Agency: Funding?

2008-06-24T19:17:55.0000Z

Not to be difficult here, but most agencies (and organizations) receive "funding" from multiple sources. I would suggest "operating budget" or something else more inclusive sounding to describe how much money was available for use during a certain time period. Either that or add a "Funding source" property to the CVT if you want to be specific about funding.

Government Agency: Funding?

2008-06-24T18:18:21.0012Z

They do seem reasonable. For funding, would a simple CVT that had properties for currency, amount, and year be sufficient?

Government Agency: Funding?

2008-06-23T21:21:50.0005Z

I added a comment to the FBI page, but perhaps it is better suited here. Could a "funding" property be added to the agency? Also, how about a "number of employees" property? These both seem reasonable and necessary.

Awards: Listed Awards

2008-06-22T13:59:47.0005Z

I have created a new type, 'Listed Award Category', which allows for awards where award winners are listed in order. This is applicable to the ever popular Top Lists: e.g. Fortune 500 or Times Top 100 Graduate Employers.

Quotation: Quote as an alias for this type?

2008-06-10T18:31:41.0007Z

Good idea. Consider it done.

Quotation: Quote as an alias for this type?

2008-06-10T01:47:25.0012Z

Seems like 'Quote' might be a good alias for the 'Quotation' type

Government Agency: Suggestions for additional properties for Government Agency Type

2008-06-09T17:04:21.0000Z

Yeah, level of government is complicated. Not all countries are the same, nor are all cities. Some have metro bodies that encompass more than one locality.

I think I've always done 3 levels, and they can mean whatever it means for that country: Federal, Regional, Local:

Federal=country level only
Regional=provinces, states, cantons, administrative divisions (like Spain), or even cross-provincial entities
Local=below state regional (i.e., cities, metro agencies)

I'm not saying that is the best way to go or anything, but just an example.

Government Agency: Government "Agency" vs. "Body"

2008-06-09T16:57:04.0000Z

I disagree. I think at the federal level it would be the country's congress or equiv. It is not an agency (yes it may have employees but I still think it is a "body"), and I think it is a valid distinction. Yes, it is limited in use, but still a distinction. (most govt entities will be an agency)

Government Agency: Government "Agency" vs. "Body"

2008-06-08T17:43:27.0005Z

I am a little anxious about saying anything as a big old newbie, but I don't think the disctinction between "governmental body" and "government agency" is valid, either legally or practically except in limited circumstances. The distinction seems to swing on whether the "body" has employees, in which case it becomes an "agency."

Many government agencies that have employees also have boatloads of politicians who are elected or appointed to leadership or managerial roles in that agency: the FCC, for example.

Second, there are virtually no "government bodies" that are composed of politicians only. Courts, many city councils, legislatures all have employees, even though the people calling the shots in those entities were elected or appointed through the political process. They all have employees / flunkies to carry out the directives of the decision-makers.

The division between government "body" and government "agency" does not take into account the realities of modern government, wherein various "bodies" have employees and "agencies" have politicians on board.

Just my thoughts, but I believe the distinction is artificial. I think that categories based upon the legal status of the entity, such as belonging to the executive, legislative or judicial branches would be a better starting point. The three branches of government are not just a taxonomy, but tell something about the nature and powers of the entities that belong to them.

Thanks for your help! Dennis

Measurement Unit: Dimensions in Floating Point Range and Floating Point Statistic

2008-06-04T16:10:23.0007Z

This is a problem with freebase right now. The fact that dated values must be represented as CVTs (compound value types) means that we can't treat the date as an independent dimension.

This is something that other users have brought up before and we will need to come up with a better long-term solution.

For now, you have three options:

- Use the dated integer as is, and your value will be dimensionless

- Create your own dated value types in your domain. You would have to create one for each dimension of dated values you would like to use

- Use spatialed's approach which is to encode the dimension as an additional property of the CVT. This may give too much flexibility, in some cases, however, but it has worked well for him.

Measurement Unit: Dimensions in Floating Point Range and Floating Point Statistic

2008-06-04T00:56:27.0000Z

If I understand your question correctly, the reason that types like Floating Point Range and Floating Point Statistic have to be unitless (or be tied to a single unit) is that they are types, and units are tied to properties. That is, when you say that a property has an expected type of either floating point or integer, you can specify a unit for that type. But when you say that a property has an expected type of Floating point range, you're saying that the expected type is another type, and you can't assert anything about the properties of one type from another type.

Measurement Unit: Dimensions in Floating Point Range and Floating Point Statistic

2008-06-04T00:49:06.0000Z

I mean, I guess it's academic. Thanks for the replies.

Measurement Unit: Dimensions in Floating Point Range and Floating Point Statistic

2008-06-04T00:48:08.0000Z

Sorry, I don't understand the problem with the fact that different types that would use such a unit-ed floating point range would need different units: how is that different than the various types using a unit-ed floating point number (built-in), all with different units?

And the slick interface for choosing units as in the built-in floating point number wouldn't be able to exist in any interface for a custom range or distribution type, right?

Measurement Unit: Dimensions in Floating Point Range and Floating Point Statistic

2008-06-03T23:09:42.0000Z

Yes. Floating point statistic is intended to be unitless. For example, the property "Clutch size" already means # eggs so there is no reason to specify other units. The "statistic" part of the name refers to the type's property that links to summary statistics (e.g., mean, median, min, max).

On the flipside, I created a user_defined_length type that allows users to specify the length measurement unit (e.g. cm, in, mm) for cases when there are not standard ways to define length and I didn't want to force the user to do the conversion.

These types are in my Summary Statistics domain, not the Freebase Common domain.

Measurement Unit: Dimensions in Floating Point Range and Floating Point Statistic

2008-06-03T21:29:43.0007Z

Units can't usefully be attached to the generic "floating point range" type because the different types that use it might need different units; it's probably only appropriate to use for types where units are not needed. If you need a range with specific unit, you should either create a new type that uses that unit, or use two properties (e.g., "minimum" and "maximum"). Generally, if you can have a multitude of ranges, using a new type is best; if there is only going to be one value each for min/max (or whatever you call them), using two properties is probably more efficient.

Floating point statistic is in a user domain, and I'm pretty sure he intended it to be unitless.

Measurement Unit: Dimensions in Floating Point Range and Floating Point Statistic

2008-06-03T19:00:58.0011Z

I believe there should be the ability to attach units to these types as well as just simple Floating Point Numbers.

Comments?

crism: Test

2007-11-20T02:13:29.0129Z

Is this thing on?

Government Agency: Suggestions for additional properties for Government Agency Type

2007-11-13T00:25:27.0007Z

Sounds like we're in agreement about the predecessor/successor model. Level of government is an interesting problem. Part of the problem is nomenclature -- "national" and "local" are fairly straightforward, although there can be several levels of government that are arguably "local" (in the US, city/town/village and township are pretty clearly local, and I would guess county is as well). But for the in-between level(s), I'm not sure -- at what point does an administrative division stop being local and start being "state"-level? Are Italian provinces "local" or "regional" (they sit below "region" but above "comune" in terms of hierarchy)? I also wonder if this could just be inferred from the jurisdiction itself? (I.e., this agency has a jurisdiction of something that is a city, therefore it is local?)

Government Agency: Suggestions for additional properties for Government Agency Type

2007-11-10T22:27:06.0007Z

Scratch my comment about a single government agency splitting in two, as I realized after I posted this that that defines a predecessor/successor relationship! I guess Jeff's suggestion is the best, that a split would involve two or more agencies being separated and that merger would involve two or more agencies being brought together.

On the issue Jeff raised of a "an agency spawning off a section of itself into a new agency, but without dissolving the parent agency," I think you're still looking there at a predecessor relationship. The new agency's predecessor is the parent agency that remains behind.

Government Agency: Suggestions for additional properties for Government Agency Type

2007-11-10T20:09:59.0007Z

Hello Crism and Jeff - sorry for the lengthy delay in getting back to your replies for more information/comments on my suggestion. It looks like you've acted on some of them, which is great. I looked at the current instances of Governmental Jurisdiction and it looks like I used the wrong term for what I was trying to impart, though I thought from the examples I used -- "(federal, state or provincial, regional municipal or local)" -- that what I meant was a property that embodied that aspect of government. I think this is a very important element that can be used for meaningful queries. That being said, I'll leave it up to you to decide whether the "level" of government, for want of a better term, should be a property of the Governmental Jurisdiction Type, the Government Body Type or some other Type.

Regarding the Predecessor/Successor properties, I believe the Split/Merged properties are required and could work the way Jeff indicates, except a that split would also have to allow for the last example he cites of an agency spawning off a section of itself into a new agency without dissolving the parent agency. In other words, a single government agency can simply split in two, which frequently happens at the state and provincial levels. On the other hand, mergers can encompass any number of agencies being amalgamated. In my part of the world, it would be very unusual for four agencies or departments to be merged into one. More typical is a merger of up to three, with two departments being merged into one representing, I think, the norm.

Government Agency: Suggestions for additional properties for Government Agency Type

2007-10-30T19:24:36.0007Z

I'm working on revising this model right now. One thought I had is that we could use predecessor/successor properties to represent split/merged as well: If an agency has two (or more) successor agencies, it would indicate a split; if it has two or more predecessors, it would indicate a merge. The only thing this doesn't cover, I suppose, is the case of an agency spawning off a section of itself into a new agency, but without dissolving the parent agency.

Government Agency: Suggestions for additional properties for Government Agency Type

2007-10-16T01:57:37.0007Z

Thanks for the suggestions, David. The establishment and dissolution dates seem pretty straightforward.

For country and jurisdiction: would it be sufficient if the associated government had more location-based information? Or are there a significant number of agencies whose jurisdictions are different from that of their governments?

Along similar lines, are both predecessor/successor and split/merged needed? Would just predecessor/successor suffice? And what proportion of agencies have this? The answer will influence priorities, I think.

Thanks again for looking at this.