Richard Rinehart:

Hello Rhizomes,

I'm writing to follow up on Lauren's email about the Rhizome ArtBase and to kick off a conversation about the language we use to describe works in the ArtBase.

There are different types of metadata relevant to works in the ArtBase and some are fairly straightforward such as Creators, Dates, and Titles. But the type of metadata that is most problematic and at the same time most community-driven is descriptive metadata such as Type, Genre, and Keywords. The data-values used to fill out those metadata are terms taken from vocabularies (the lists of different types, genres, etc.) If you have ever submitted a work to the ArtBase, you know what these look like: Types include animation-art, audio-art, etc.; Genres include abstract-art, allegory-art, etc; and Keywords include access, animation, archive, etc. (a full list of Rhizome's data-values/vocabularies follows below).

Rhizome would like to update the vocabularies it uses for this descriptive metadata. Rhizome has cited three reasons for changing; the vocabularies are incomplete, the vocabularies are however key as they are how visitors search the vast Rhizome site, and lastly, but not least, there is no canon or authoritative source for terms related to digital art, so Rhizome can take this practical need and turn it into an opportunity to engage a community discussion about vocabularies and to set an example for others to follow. All metadata specific to one discipline, but especially vocabularies, need to arise from the community's practice and not be imposed from outside or the descriptions and the artifacts being described will never quite match up. It is also important to collaborate and coordinate with other groups working on digital art metadata and preservation, so that's another reason to have this conversation on RAW and why Rhizome will also be convening people from the Variable Media, Archiving the Avant Garde, and Canadian DOCAM projects to discuss this as well.

Some questions and considerations that might get the conversation started:

1) Do Rhizome's vocabularies need to be compatible with other metadata standards? If so, which, and how much?

Thoughts: Many other disciplines and communities use what they call "controlled" vocabularies or authoritative thesauri. For instance the art world has used the Getty's Art and Architecture Thesaurus for years ( ucting_research/vocabularies/aat/). Systems are then built using these vocab standards. If Rhizome were in some way compatible with these standards, then new search engines could search across distributed art resources online from Getty databases to Rhizome's ArtBase ensuring that digital art is not "ghettoized" because of incompatible languages. Interoperability is important in a semantic as well as technical sense, but luckily compatibility does not necessarily require that one adopt the "authoritative" vocabulary completely, or exclusively.

2) What can we propose here that Rhizome can practically accomplish given limited resources?

The larger cultural world is cursed with a plethora of metadata "standards" and vocabularies that are so complex that no one can afford to implement them and thus they go unused and interoperability remains a theoretical concept. We should be smarter than that. A simple system that works and can be realistically maintained is worth more than a complex solution that never happens.

3) Currently the metadata that uses vocabularies is divided into type, genre, and keywords -- are these categories sufficient? Should we add others?

Thoughts: Many other disciplines and communities use metadata categories similar to this. For instance, in various art-world/museum metadata standards they use Genre to indicate a broad category ("painting"), then Type to indicate a format within the Genre ("watercolor"), and then Subject (keywords) to indicate "intellectual access points" ("landscape") that people will search on to find the record.

4) Do we want to enhance/ elaborate/ add on to our existing descriptive terms or keep the current controlled vocab as is, and make folksonomy also an option?

Thoughts: Can one use folksonomies or other dynamic systems to keep a vocabulary fresh yet still retain some level of compatibility with other standards?

5) who is the artbase for? Who is its audience, and how does that affect our re-design of the metadata.

Thoughts: Related to this is the question of what the long-term use of the ArtBase should or will be and how can we support that with better vocabularies?