Wednesday, October 6, 2010

Metadata in the Real World.

Library school is in full swing for me. It's hard, but not impossible. Lots of reading, which is annoying only because when people ask me what classes are like, I have to reply, "It is a lot of reading." This furthers the stereotype that all the field consists of is reading. Wrong. Librarianship isn't about books, it's about information. One main concept we've started to talk about now is "metadata," essentially, data about data. Informally, metadata is "a cloud of collateral information around a data object" - Clifford Lynch. Metadata are the descriptions of items or records in a catalog that allow you to search by keyword or description rather than the search engine having to search loads of full text and taking much longer. Example time.

So for a particular album or book or DVD or journal article, metadata and data are different. Let's look at a random album to make this a bit more real, say, the album 19 by Adele. Metadata would include things like, album title, artist, record label, executive producer, and the data for the specific album would be 19, Adele, Columbia/XL, and Jim Abbiss. So that's metadata.

Metadata allows better cataloging and more efficient searching. But it's also super complex and confusing, far more than I ever really thought it would be, mainly because no specific format for metadata has achieved real standardization in the information world. So you've got millions of different individuals, corporations, libraries, and groups using all sorts of different search methods using different metadata. Confusing.

It's an interesting concept to me though, and it will be especially interesting once I learn how to apply it to real life or actually get to work with metadata in a real way. Since that won't happen for awhile, I've decided to undertake a giant personal project to get my head around the idea. I'm metadata-ing my iTunes library. Hawt.

I'm going to try to get one album a week catalogued, which means I'll have this whole project done by the year 2022. Not bad. To get more specific, I'm going to basically be adding info about each album I have into the comments section of a song or album's info. So I'm starting with some cataloging already done, a few standard fields all filled in:

1. Album Title
2. Artist Name/Group Name
3. Year released/recorded
4. Genre
5. Album Artwork

Five fields, all already filled for every song and album in my library. But I'm going a bit further. I'm going to add as many additional authors as I can for these songs. For example, I just catalogued the album of earliest release date in my library, the Columbia Symphony Orchestra's recording of George Gershwin's masterpiece, Rhapsody in Blue, and the New York Philharmonic's recording of another of his beautiful works, An American in Paris. Here's the metadata I added:

1. Additional author: Gershwin, George (1898 - 1937)
2. Add. author: Bernstein, Leonard (1918 - 1990)
3. Add. author: New York Philharmonic and Columbia Symphony Orchestra [each one labeled to the specific track they recorded]
4. Original recording date: June 23rd, 1959 (for Rhapsody) and December 21st, 1958 (for AAIP)

This will be unique for every single album, and very possibly for every single song. For these two Gershwin songs, they have different information because they were released and recorded before the standard "album" as we know it was commonplace and they were two totally different works done by the same artist; it's only for continuity's sake that I've put them together under one album label. Once I get into newer albums, there will be a lot of additional metadata fields for record label, additional performers (featured artists), executive producers, etc. There is tons of extra information I can add. If I wanted to go totally nuts, I'd start cataloging sampled works in the music I have. Holy cow that would be insane. I could label samples of samples and also do covers. The data is seriously unending.

So far, the biggest obstacle I'm running into is that iTunes isn't set up for this kind of data entry. The comments section of song info is really the only place I can put this much superfluous information, and I can't organize it as well as I'd like. I can't even put info into separate lines. Blah. The other main problem is that while iTunes is referred to as a "library," there is no good catalog setting through which to search. Anyone who pops onto my iTunes and is curious to see what items I have that have Leonard Bernstein somehow included in their metadata has to create a smart playlist that matches the following rule: "Comments" "contains" "Add. author: Bernstein, Leonard". Not a lot of room for error here. But it's a work in progress. And (I hope) it's getting me ready for some sort of real world work in the field of library/information science, whereas in all of my classes we're basically just gabbing at each other all day long.

Oh hey, if anyone wants me to completely revamp their iTunes collection, making sure no duplicates are there, everything has correct data entered in all fields, then add lots of metadata for better cataloging purposes, and you're interested in paying me hundreds of dollars to do it, just let me know.

-Jon

1 comment:

Mrs. Robinson said...

Of course, you realize that your target completion date is only valid if you don't add any new albums between now and then...unlikely, I think...maybe go for two albums a week...:-))