From looking at the key to that map, I'd say they're different since they have different entries in the key. The Census Bureau uses approximately a bazillion different levels in its hierarchy, including those that you see in the key like block, tract, CDP, CCD, MCD, etc, etc.

Do you need to know that it's a CCD, or can you just use the generic statistical region type?

Additionally, some pitchers are also position players. Having all the relevant stats on Baseball Player makes the most sense to me.

I just checked Retrosheet, and it looks like it'd be pretty easy to load all the pitching stats from there. See, for example, http://www.retrosheet.org/boxesetc/W/Pwilsb001.htm

I think you're right. More broadly speaking, however, I suspect that the real problem is that the four topics currently in your saved view are somewhat dubiously typed.

• French Language is not an ethnicity, but it can be a field of study. The "french ethnicity" info needs to be split to French people.
• English (e.g. "English people") is not a field of study but is an ethnicity; due to the ambiguity of the name, the field of study data presumably should be split to English Literature. I renamed the topic to "English people" to reduce the number of English majors who select the first "English" that appears before they remember that they really majored in English lit. (I can make fun of people who do this because, somewhere, deep in the depths of the graph, you can find the historical assertion that I majored in the English language, for similar reasons.)
• Judaism is a religion (and therefore a legitimate field of study); the ethnicity data should be split to Jewish people.
• Sephardic Judaism is arguably not even a religion (although I'm not entirely sure what type it should have); there is a separate topic for the ethnicity, Sephardi Jews

Oops, you're right these shouldn't be merged. I was going by the matching scientific names (and the fact that the genus article said that it contained 8 species and we had 11).

I'll rejigger the structured data to match the article, by removing the scientific name and level (species) for Minke Whale and making the two species subclassifications of it.

I thought you and Jeff agreed that the operational entity (the Port of Call) would and should be different than the supporting city/town/village ? To me they need to be 2 different topics with 2 different types to encapsulate additional metadata and structure and statistics for the Port of Call vs. the sometimes containing city. Just like /airport/serves . It's common for a Port of Call to have a similar sounding name like it's supporting city.

If you see any city/town/villages that are also a port of call, then that's probably NOT correct and the topics ideally need to be split as we had already agreed. I think the very first uploads that I had with Ports of Call and was with the nasty import tool that didn't really allow me to inspect each one easily, so we (I) may have to run back through each Port of Call and split it from the City topic.

I have now added the City_Served property to the Port of Call type that should now help to avoid the confusion.

Anything else ?

Ah, OK. Sorry about about that. The new property should help make things clearer. Would it make sense to have the default name for these things to be "Port of "? I think that's the general convention in English. The plain city name could be kept as an alias.

It'd be nice to include more location information too. Is anything available in the database that you loaded these from? In this case, I think we could say that the port is contained by the city and that would probably be generally true, but there are likely some outliers where the port is a ways from the modern city.

I agree. Skud's schema seems pretty good as a starting point, though it would be great to see things like "affirmed_by", "followed_by", "overruled_by" and so on. It would also be useful to see:

• ways of linking decisions to jurisdiction (maybe you can do this via Court though I'm not entirely sure)
• something more flexible concerning parties (per my comments on Skud's "Legal Case" schema)
• perhaps more generality for disposition (since you don't find in favour of a party).

The Judge schema doesn't really work for my jurisdiction since judges aren't appointed to courts, a judge will have a particular status (district judge, circuit judge, high court judge etc) and may then sit in multiple courts (the rule for circuit judges). I don't think its useful to record a relationship like "may sit in" for judge/court since that gives a potentially huge plurality of options, but I'm not sure you need "has sat in" since you can do that via the Legal Case type (maybe?).

I'd be interested to know where the right place to discuss this might be and really new to the project so not sure what to do that doesn't mess up the existing schema. I'm kind of interested in entering information about cases that interest me (as a copyright lawyer).

Welcome to Freebase! Here's as good a place as any to discuss this since it's right in the context that you're talking about, but if you want a wider audience there's the freebase-discuss mailing list.

Domain expertise is very welcome. A lot of these schemas have been put together by people with an interest in the domain, but not any professional experience.

There are a few ways you could proceed:

1. Experiment with loading data into one of the existing domains. Even though we refer to them as Skud's or Toby's, they're all available for anyone to use.

2. If you are comfortable with modeling, clone the schema that you think is closest to what you need and start modifying it and entering your data.

3. Start a domain/schema of your own from scratch

During the initial stages, you might want to do your modeling and data entry on the sandbox (www.sandbox-freebase.com) which gets refreshed every week automatically wiping out any mistakes without effort on your part. The downside is that any work that you want to keep needs to be replicated on the production server, so you want to balance risk vs effort.

From your comment about the Crown, it sounds like you're in either Britain or a British colony. If you're familiar with differences between the jurisdictions that you work in and other legal systems, it might be useful to point them out as you go to help people fill in the holes and generalize things for other parts of the world.

Thanks for offering to contribute! Let us know if you need any help.

Hi, thanks. I'll have a play around and see what I can do.

How does one start domains or schemas from scratch? There's nothing in the help documentation to tell you how exactly. Do I need to do it via the API (eg using a python library) or is there a way to do it from the website I haven't found?

Although playing with it myself is fine, it would be nice to know that there were others to chew some of these questions over with. Also there's a few things that I don't know what their intention is, for example under "Law Case" there is a property "Citations" - does that mean ways of citing reports of the case, or other cases which cite the case (actually maybe its obvious that the former is meant).

I am interested in trying to make things not too jurisdiction specific. I would be interested to know what other people think about my problem with "Judge" being assigned to a court. How does one fix that?

Here's another example (with an appellant role): Dolan v. United States Postal Service. I'm starting to really favor this model over the two property plaintiff/defendant model.

I also suggest deleting the "found in favor of" property, and tabling the model for now, in order to move this process along; this seems like too much of a simplification of the complexities of legal cases. I think it's an issue that could be continued to be worked out in a base, and later added to the (presumed) commons legal case type.

Any thoughts?

Thanks for this list! I'll check them out for the next round.

Slaughterhouse I agree isn't ready for prime time. "Infrastructural municipality" is not the most sonorous type name ever (although it's better than some I've created!), but unless we just went for /location/location and therefore didn't reciprocate the properties, I think a specific type is needed. I'll see if I can think of a better name.

Wait... is NGA not stand for National Geospatial-Intelligence Agency ? or is this some other kind of identifying number ?

The reason I brought it up is that I have seen some NGA numbers before and they appeared in this format "110-18360"

An Example: http://commons.wikimedia.org/wiki/Category:Focinho_do_Cabo_Lighthouse

I feel like I'm playing twenty questions.

So your point is that because it's a structured number containing a hyphen (which Freebase sometimes drops for cases like this, btw), it should be a machine readable string instead of an integer.

Since you've researched this and the rest of us (or at least I) haven't, can you provide a pointer to the spec for the format of the identifier?

Changing this to a machine readable string is easy from a data migration point of view, but nominally a breaking change for applications if they are depending on the fact that it's an integer, so it would need to be advertised ahead of time.

I have NO IDEA what NGA number stands for, Tom. So...I did some more digging.

NGA is in fact the National Geospatial-intelligence Agency in this case it appears. With the November 24, 2003 signing of the Fiscal Year 2004 Defense Authorization Act, the National Imagery and Mapping Agency (NIMA) officially became the National Geospatial-Intelligence Agency (NGA). So they were previously NIMA. Ok, finally I'm getting somewhere.

Now, apparently, they produce the List of Lights publications. LLPUB110, List of Lights - NGA Maritime 02010

Apparently, this map here shows what those prefixes really mean, that I saw around the Internet... like 110- which means the number is found in publication 110 for those lighthouses in Mexico / South America. Ah ok, didn't know that. And looking at the 110 publication linked above, it appears that they have started to use decimals in areas where they have run out of 5 digit integer patterns.

Mexico 15636.4 Height 59 ft , Round concrete tower; 52 ft.

And this Screenshot of the query against their database shows the same use of

(pant pant pant - I'm out of breath just in the last 2 hours researching and typing this !) Hopefully the above will help you make a good decision either way.

Thanks Thad. To summarize, the NGA has publications #110-116 which are light lists for different geographic regions. The official light lists have the lights serially numbered starting at #1, so they overlap, but by convention the publication number is used as a prefix when referring to a light. Decimal numbers are used to insert lights at the appropriate point in the list when there's no room available in the numbering sequence.

The decimal numbers are reason enough to switch to a machine readable string, so you're right, this needs to be changed. I guess the remaining question is whether there's any value to treating pub number + light number as two separate properties or to just stick with the convention of concatenating them together.

Note that the NGA list only includes non-U.S. lights (with a few exceptions). The U.S. list is maintained separately by the USCG, but it's not clear to me whether or not their assigned numbers are stable. There also appears to an international number registry which assigns numbers of the form Z99999.99 where the alpha prefix is a geographic zone, but I don't know what agency maintains the registry.

Having said all that, Robert is the only domain admin here and I'm not sure how active he is these days. We may need Jeff or Kirrily to help with assigning a backup domain admin.

It does help with notifications! I don't know if Robert's still interested in maintaining this base or not. As Tom says, changing the expected type is doable, though not completely trivial. The data appears to have been scraped from WP infoboxes, which do not include the "list of lights" volume prefix. I'd say that we pretty much need the publication prefix if we want to use these as keys (which would be better than machine readable string, as long as we're changing the data type anyway), but we actually have no way to convert the current prefix-less numbers to prefixed ones, so that's kind of a problem.

The NGA's data set includes a correspondence to what are probably the British Admiralty numbers (I think this is Tom's agency with the Z99999.99 format), which are also in the WP infobox. The infoboxes also have an id for the Amateur Radio Lighthouse Society's list.

Thanks for your comments. You are probably right that I should do this on the sandbox, but since I barely know how to use freebase anyway, and someone made this base for me (which I thought I could just play with till I was happy) I haven't tried sandbox. Its confusing enough working here. Also I understand that all my work is wiped each Monday on sandbox, which would make it harder to do things until I've worked out how to script things and then I can rebuild what I had already.

I can't work on this full-time so this has to be done fairly slowly. I'd love to work somewhere where I could happily develop something and then see what others thought of it. Sandbox sounded too shifting sands for that 8-).

I'm also still getting my head around terminology and the way CVT's work, so bear with me.

In answer to your question - no, citations do not cite a previous ruling, they are the ways in which this ruling is cited. I'm aiming to add a property indicated any references to previous rulings. The citations property that is there at the moment is for the ways in which this case is cited.

However, I think you are right that a type is needed for citations. Citations aren't unique and there's not always a canonical citation for any decision. So each court decision would have a number of citations. Is that something one would use a CVT for? One doesn't want a topic for each citation, surely? I may be getting muddled with terminology again.

Citations have a number of different structures - as you have pointed out - so one (type?) should exist for each form of citation and have the structure you suggest. Does that sound right? Most "English" citations will take the form (with a few variations) ditto, US cases have a nice system, as do those in other jurisdictions and I can put together a collection of such things if that sounds the right thing to do.

I appreciate any help or feedback, I really do. Its going to take a few weeks for me to get my head round the freebase system.

Hey, feel free to model here , where it won't get wiped away. When i get a brainstorm of an idea that is kind of complicated, I often model it out on Sandbox if it's a type/base/commons that is critical/well-used by external sources. But I only do it if i intend to reproduce on Freebase proper almost immediately or I'll make a copious notes, screenshots, learn MQL enough to export the schema off site. etc.

We really appreciate your interest in a somewhat neglected area of Freebase!

Initially I read the description of Governing Body of Protected Sites which seemed fine to fit with my initial question. But the definition of Protected Site seems a bit stricter? At least it says:

"A "protected site" is any location that is protected under law" "usually under the control (at least in part) by some form of government agency."

Could an NGO also be added, as well as govt agency?

I'm a little confused as to what you are trying to model. Is it publicly owned places or is it "protected" places (which is an ambiguous word) or is it some combination or something else? For example, US National Forests are actively managed and US National Parks are not but both are "protected". There are government programs (e.g., US Farm Bill) that "protect" areas via subsidies for limited amounts of time, where protection is defined in terms of a list of acceptable activities (e.g., planting native grasses, not harvesting timber) that may change every couple years. Sometimes governments or NGOs will purchase timber rights but not mineral rights and vice versa.

My interpretation of the current schema descriptions are that they focus on public lands. It would be nice if they were broader than that. Y'all might want to take a look at how other people are defining these data. For example, the California Protected Areas Database(not for commercial use) includes NGO holdings. TheUS Protected Area Database design looks to model easements separately. Both of these projects focus on "fee-owned lands".

However, check out this backup-holder technique in Massachusetts. Looks like pretty permanent protection.

Just wondering what the consensus is for the quick bit of work I did yesterday on these Designations ? http://www.freebase.com/view/m/0cj5w33

I wasn't sure if the 23rd Annual listing of "America's 11 Most Endangered Historic Places" is a good name. Perhaps the naming is too catchy? And should be renamed simply "Endangered Historic Place" ?? But I went with their own designation. Thoughts?

Also, Listing Agency, Categories, and Sites properties do not contain a description. i would like to see some, as it would have helped me.

Now back to the topic at hand, those Governing Bodies that protect sites... Guam Preservation Trust is one, I think. They acquire title to protected sites, http://guampreservationtrust.com/archives.php which are also listed with National Register of Historic Places, such as http://www.freebase.com/view/en/san_dionisio_church_ruins.

How does one go about linking "Guam Preservation Trust" that acquires title to some of those sites within NRIS that they manage ? Are they simply the "owner" of the designated site ? hmm.. perhaps so, since they have the title ? (See the About Us page for Guam Preservation Trust's website)

I think an example may help with this one: Fingal's Cave (a geographical feature) is an example of a sea cave (a geographical feature category). Sea cave is a subcategory of cave, also a geographical feature category. (Cave was until about a minute ago mistyped as a geographical feature).

More generally: a geographical feature is a specific place you can go and visit. A geographical feature category is a type of geographical feature.

Does that help?

What properties would the type have?

At the moment, this can be modelled by typing the topic as a geographical feature and setting the category to "beach" - in general, we try and avoid types with no properties as they don't add any real value.

True. But the way the schema is now it is then impossible to tell which plane type was operated by which company, and how many passengers and crew each plane carried.

Would it be better to have a property/CVT for each aircraft involved in an accident? Or is that too big a change for this type?

Iain: Good point. A CVT makes sense to me.

Paul: searching for "SwissAir 111" and "Tenerife crash" already produces the appropriate topics as the first hit. Even "Tenerife" gives the crash as the 4th match; looks to me like things are working well at the moment. We can always add aliases if we think they would be helpful.

It looks like this type needs some refactoring. I'd recommend something like this:

1. Move these properties to a CVT that will represent each aircraft involved:

/aviation/airliner_accident/aircraft_type

/aviation/airliner_accident/tail_number

/aviation/airliner_accident/operator

/aviation/airliner_accident/passengers

/aviation/airliner_accident/crew

/aviation/airliner_accident/survivors

/aviation/airliner_accident/flight_origin

/aviation/airliner_accident/flight_destination

1. These properties should actually be pulled from the Disaster type:

/aviation/airliner_accident/site

/aviation/airliner_accident/fatalities

/aviation/airliner_accident/injuries

1. These properties should actually be pulled from the Event type:

/aviation/airliner_accident/date

1. These properties should expect Airport rather than Location:

/aviation/airliner_accident/flight_origin

/aviation/airliner_accident/flight_destination

The definition might also better be expanded to "aircraft accident" rather than strictly airliner accident to allow the inclusion of topics like 1948 Gatow air disaster.

Wow. That's hideous. Sorry about that. Let me try again:

1). Move these properties to a CVT that will represent each aircraft involved:

• /aviation/airliner_accident/aircraft_type
• /aviation/airliner_accident/tail_number
• /aviation/airliner_accident/operator
• /aviation/airliner_accident/passengers
• /aviation/airliner_accident/crew
• /aviation/airliner_accident/survivors
• /aviation/airliner_accident/flight_origin
• /aviation/airliner_accident/flight_destination

2). These properties should actually be pulled from the Disaster type:

• /aviation/airliner_accident/site
• /aviation/airliner_accident/fatalities
• /aviation/airliner_accident/injuries

3). These properties should actually be pulled from the Event type:

• /aviation/airliner_accident/date

4). These properties should expect Airport rather than Location:

• /aviation/airliner_accident/flight_origin
• /aviation/airliner_accident/flight_destination

The definition might also better be expanded to "aircraft accident" rather than strictly airliner accident to allow the inclusion of topics like 1948 Gatow air disaster.

Looks good to me.

A slight tangent as we're in the area: Pan Am Flight 103 wasn't an "accident" in any sense that I'd use the word. Do we want to rename this type?

I've started work on this (Aviation incident/aircraft relationship -- the link from Aviation Incident is currently hidden while the modeling is in progess), and have naturally come up with a few further questions.

1. Should "survivors" be on the CVT (as I originally suggested), or should it just be a "total survivors" property on Aviation Incident itself?
2. What should the expected type of "Flight Origin" and "Flight Destination" be? I wanted to change it to Airport from Location, but what if the flight originated at a helipad on a hospital or the like, or from an aircraft carrier? Neither Location nor Airport really seems to answer.

Also, just to mention, I've made the expected type of "operator" be Aircraft Owner on the CVT, so that military and private owners can be correctly entered. I also made "tail number" expect a machine-readable string, rather than text, since the value should be language-independent.

Work has started on this task (DA-1058). We're running into a problem, though, which is that a lot of the flight destinations and flight origins are for cities, not airports. I've set up two To Do Lists for these incidents, if anyone would like to help clean them up:

• Aviation Incidents where Destination is City/Town
• Aviation incidents with Origins that are city/towns

Thanks skud.

As much as I'd like to go crazy with my new powers and move everything to the commons, I'll hold back as I'd like some critical review of the types I mashed together in the association football base earlier today.

Well, since you asked... :)   Bear in mind that I know exactly zilch about association football, other than which three English teams contain swear words in their names (which is what you learn when Billy Bragg is your main entree into the sport).

I notice that there is no direct relationship between players and teams (only via squad or match), and that in the case of Manchester United, you've co-typed it as both a squad and a team, which seems confusing based on the type descriptions.

Should soccer field include sports arena?

For goals, I think the disambiguator "team" is ambiguous -- at a glance it's not clear whether that's the player's team or the team the goal is credited to.

A soccer league does not seem to contain any teams.

Why is a soccer league also a soccer competition?

Soccer competition's property "seasons" expects "Event" rather than Soccer Soccer competion season.

Speaking of soccer competition season, the "matches" property should probably be reciprocated on the Match type.

But overall this is a very impressive schema!

Maybe add attendance to the soccer match type?

If possible, it'd be great if these types could inherit properly from the existing sports events/sports league championship events .. 

i.e. a soccer match (is a) sports event (is an) event .. a soccer championship event (is a) sports league championship event (is a sports event) (is an) event ..and so on ..

I'd mainly like this to reduce data duplication across types, and to enable querying across sub-types i.e. give me all the sports events held at Wembley in 2008 (although I suppose this is possible at the moment, but there is some duplication of data required i.e. location, date/time etc)

Just to add: 

The "sports league championship" type naming is probably misleading but the schema does fairly accurately describe all football cup events (Essentially the FA cup is a league championship for the FA associated teams, and the 1999 FA Cup Final was for the 2008-2009 FA Cup Season)  

The biggest difference for football I can see is in the multi-format competitions (although these are similar to american sports which the sports event types seem to have been created for) where you start with a league format and then a set of play off matches to arrive at a single winner i.e. world cup, euros, champions league

>I notice that there is no direct relationship between players and teams (only via squad or match) 

I've added a time-mediated property soccer player tenure.

 >and that in the case of Manchester United, you've co-typed it as both a squad and a team, which seems confusing based on the type descriptions. 

Now fixed, I've split the squad from the team.

 >Should soccer field include sports arena? 

now typed with sports facility.

>For goals, I think the disambiguator "team" is ambiguous

 now called "point awarded to"

>A soccer league does not seem to contain any teams.

I've created soccer league participation which is time-mediated using seasons instead of dates. 

>Why is a soccer league also a soccer competition?

This was where I was having issues.  I'm trying to get my head around the Sports commons and also get some abstract inheritance in. I was trying to make league competitions different from cup competitions, and possibly different from multi-format competitions - yet all have a common inheritance from soccer competition.

>Soccer competition's property "seasons" expects "Event" rather than Soccer Soccer competion season.

fixed this. 

>Speaking of soccer competition season, the "matches" property should probably be reciprocated on the Match type.

now reciprocated. 

 >If possible, it'd be great if these types could inherit properly from the existing sports events/sports league championship events .. 

absolutely, my problem is that I'm not too familiar with the sports commons. But if you could point out which types to include, I'll update the schema.  I've done soccer match now.

I've also added a schema for player transfers and loans. 

All my types are still in association football, and I'd appreciate any further comments on the types.  If we feel some of the types/properties are OK I will start replicating them in the soccer commons.  (with the plan to eventually move everything over and delete association football).

I'm planning to move the schema from association football over the next few days.

If anyone has further comments or improvements to the schema, please let me know.

Job done!  The schema is now up and running in the soccer commons.

Hopefully I got everything from football association across (I purposefully left out squads and competition season types as I think they're a bit flaky), but let me know if there's something wrong. 

Ideas for the next feature are welcome :)

I use this thread as a general one about the domain.

A simple question: Football Competition type should be used for event or recurring event ?

I have a good list of exhibitors at various annual industry conference events. Those events are usually linked to particular subjects, such as Transportation Engineering, but broader use at an industry level "Transportation" at minimum would be acceptable.

Somehow linking the Industry with a Conference Event. However that can happen, I'm all for it.

Example: NYSATE Technical Conference http://www.freebase.com/type/schema/user/thadguidry/default_domain/exhibitor

What about those truly ambidextrous players, for example David Warner? This isn't the Kevin Pietersen style switch-hit where he changes position after the ball has been bowled, but he actually takes guard from both stances.

+1 on the stats.

(this time with formatting) i'd like to:

delete /online proceedings

delete /proceedings

delete /sessions

promote /base/public speaking event schema (all of it)

promote /geographical_area_for_recurring_events

sorry

delete /conferences/conference/online_proceedings (map this now with /common/resource)

delete /conferences/conference/proceedings (replace it with 'presented work' type from /base/public_speaking)

delete /conferences/conference_session (an unreciprocated old type)

promote /base/public speaking event schema (all of it)

promote /conferences/conference_series/geographical_area_for_recurring_events, (which is a type on the /events base)

Quick thoughts:

1. Conference Proceedings is not the same as Presented Work -- a proceedings is a publication of (a selection of) the papers presented at a conference; a presented work is one such paper.
2. Online proceedings does seem unnecessary now, so deleting it would make sense. Same goes for /conferences/conference/url.
3. I'm not sure about Conference Session. I guess the intent is to replace it with Public speaking event?
4. /conferences/conference/papers should not expect Conference Proceedings (see above). Assuming we go with #3, the "papers" property would be superfluous because Public Speaking Event already takes you there (albeit through a CVT).
5. /conferences/conference/venue needs to expect a Commons type.
6. Why not use a conference-specific Scope type? (E.g. "Conference Geographic Scope"?)

If Conference Session is deleted, what links a generic Public Speaking Event to its associated conference?

Note also that conference have presentation types which don't really have much to do with public speaking. I guess a panel session could be considered a form of public speaking (although not a speech), but poster presentations (ie put a few words/diagrams on a posterboard, stand it on an easel, and stand next to it answering questions) aren't. I think papers can also get "presented" just by dint of being in the proceedings, even if they don't have an associated session/talk.

yes. agree here.

there are many types of barcode systems thpough, so maybe it belongs in a seperate type?, which this type could eventually cotype.

i've got my hands on a big database of upc codes....

as long as this schema gets some +1's i'll upload it to this type this week.

Can GTIN be made a disambiguator?

I'd love to just cut and paste in http://www.earthyself.com/blog/dvd-barcodes or do CSV exports of a whole lot of point of sale systems to quickly beef up this data set.