Discussions on Artwork

Good suggestions. Will look into adding these.

sorry for not specifying; most artworks have their dimensions given as inches (in the US) or cm (in Europe). is it possible to do conversion or allow for different units to be entered?

For now, a single system of measurement is used for artworks for easier comparison, so that the relative dimensions of multiple artworks are obvious.

We settled on using the metric system instead of the imperial system because metric conversion is easy: 1m = 100cm, so an art work of 7.32m is 732cm -- math that people can do in their head. The imperial system of foot to inch conversion, however, is not so straightforward. Most people can't tell how many inches are in 7.32 feet (it happens to be 87.84 inches).

The use of meter over centimeter was debated. Meter was chosen eventually because although the measurement of small artworks in meters can look rather odd, the measurement of large artworks in centimeters look even odder. A small painting like the Mona Lisa, for example, is 0.77 x 0.53 meters, but Brancusi's Endless Column is 2933 centimeters tall, and Spiral Jetty is 47250 centimeters long.

I agree that having an automatic conversion tool would be useful. For now, you can easily convert artwork dimensions you know in cm to m by dividing 100 (or moving the decimal point to the left two digits). As for inch to meter conversion, I would suggest using an online tool since the calculation of 1m=39.37" and 1"=0.0254m can be (understandably) error-prone.

Sounds good. Let me try it on sandbox and I'll send out a link.

I've added the property on sandbox here. One concern I have is that there may be quite a bit of overlap between the subject and the title of an artwork. The subject of Monet's "Waterlilies" series is waterlilies, for example. And the plethora of Madonna and Child and Crucifixion paintings so popular during the 14th to the 16th century are aptly titled. I have to admit that even so, the link from artwork to subject as structured data has its advantages.

looks good.
the subject of that painting is actually Claude Monet's garden in Giverny which we dont have a topic for yet.

Hi, I gave this more thought and decided to make Visual Art Subject a type of its own. The benefit of this approach is that not only will artworks have a property for subject, but perhaps more importantly when you view a topic that is an Art Subject you can see a list of artworks on that subject. I've modeled this on sandbox and filled out some data. See the list of artworks in Freebase depicting the Annunciation (now typed as Visual Art Subject), for example.

I agree, this is much better.

Ship it! Looks fine.

The places lived type used in the Person type may be the best option. Though the name seems a little weird.

Hi Alexander, you're right that location of artworks can change. So can ownership, and both properties are overdue for a redesign. Change of ownership usually causes change in location, though not necessarily vice versa. Special exhibits should probably be its own type, with Event as an included type.

How specific would you like to model artwork location data? The level of detail required can help drive the schema design. At the museum-level, or perhaps including the museum gallery or wing info? Should it be noted when an item goes from storage to display and back at the same museum? When an art piece is en-route from one museum to another?

In terms of capturing reproductions or non-unique works of art, there's been a lot of debate but not a lot of break-through idea. How do we model the various Rodin sculptures like "The Thinker", all pieces of which share the same name but can differ in size, material, ownership and location? We'd like a schema to capture all relevant information that can be applied to special cases like this, but not unnecessarily complicate data entry on more typical "unique" pieces of artwork that don't suffer the same level of identity crisis. Ideas?

You've presented quite a few of the issues that I've become more aware of as I explored trying to find a database that records sales/movements of art that was public domain (so far, no luck). I think that it might be a good idea to have a seperate compound for tracking sales of art. The most recent buyer becomes the holder. Supporting pieces with multiple reproductions still isn't entirely clear though. Location(s) could have a similar compound value that specifies if the item is on-loan, exhibit, or owned by the location, and when they received it for exhibit. I would _love_ to see detailed location information since what I really want to do with this data is build virtual art exhibits, where you can browse a collection at a point in time.

Hi Alexander, I think you'll be happy to know that Jeff and I are going to refactor Artwork's Location and Owner properties to be time-based. Pricing info will be added at the same time. The project is currently scheduled for March. We will also revist modeling non-unique pieces of art (Rodin's "The Thinker"), although the complexity may require more time.

Speaking of detailed transitory exhibit/on-loan info, we would love to load it if you have access/rights to such data that you'd like to share with the Freebase community. Until we have enough data or users willing to put in such extensive data though, modeling it might be premature. It's my experience that trying to design schema without data to fill it out can quickly become an academic exercise, not something very useful to users I'm afraid. Let's revisit it in a bit.

Oh, if I haven't mentioned before, all users are encouraged to play with schema in their domain. Well-used user-types have been successfully promoted and integrated into the Freebase data community here.

Great news to hear! I've been looking for a provenance dataset since I started this thread actually. The Getty has a fairly good database but doesn't provide access to the data in a way that I can really make a dataset out of it (search only) and also has copyright on the data so the licensing issues come up. I'm still looking, and I keep meaning to chat with some Art History folks I know to see if they have a source for this kind of information.

 I've actually been hesitant to make my own type for it because I don't have a dataset yet, so I totally understand that issue. If you just want a sense of what the data might look like, then the Getty's provenance database(s) are quite useful. (http://www.getty.edu). 

I've put up a model on sandbox to try to deal with most of these issues. You can see it here (at least until next Monday evening, PDT).