Observations and arguments.

Archive for DAM

Feeding the Beast

Statistics compiled from iTunes,
the work computer:

Total Songs: 5879 (20.31 GB)
Need Ratings: 2732 (9.29 GB)
Never Played: 180 (.82 GB)
Played Just Once: 1094 (3.8 GB)

The Played Just Once number is meaningful in that on any particular day I may get up from my desk several times without pausing iTunes. So that for a song to register (particularly one with which I am unfamiliar, e.g. those contributed by workmates), multiple listens are required.

Ratings are important within my listening habits; most of my playlists are Smart, and rating is almost always included among the determining criteria. (Unless it is a smart playlist of songs that need rating.)

Most music listeners, I have discovered, do not use ratings to help organize their music. They might wonder more specifically about the 3,000+ songs that I *have* made time to rate.

Of course, I am in the position of having to rate songs distributed across four separate computers. Moving songs from one machine to another is relatively simple, as Title, Artist, Album, Genre, Comment, and even Album Artwork is embedded within the MP3/AAC and travels from one iTunes database to another seamlessly.

Rating, however, is an attribute that resides in the master iTunes XML data file, so a song moved to another computer must be re-rated.

(Which would be less of a problem if my rating systems were the same on every computer on which I maintain an iTunes database–Although the distinctions are neither sharp enough nor the conversion process interesting enough to cover in depth.)

But the work computer is not merely a subset of the home iTunes database(s). Rather it functions as the portal for an increasing amount of new music, which in turn is fed back home.

For the first couple of years of working in an iTunes-centric environment (and certainly prior to April 2004), the primary influx of music came from those coworkers in my immediate vicinity. An ad hoc swapping system developed, as we offered each other the opportunity to rip CDs we’d each brought in from home. When iTunes 4 was announced in April of ’04, the greatest influx of music came not from the new $.99/track Music Store, but from the newfound ability to listen to the shared music on the local network.

And while the collective taste was weighted heavily away from traditional and towards electronica, one could often find a new or familiar album of interest amid the collection of a coworker six seats over, or even in the next room. Depending on the level of social interaction I had with that workmate, I could then ask her/him to drop me the album via a shared network volume, allowing me to add it to my own collection. In some cases the request was itself an excuse to initiate social interaction (e.g. “I was listening to your copy of the Jesus Christ Superstar original cast album. Doesn’t that rock?”).

But of course the technology that allowed for music streaming on a local area network can make it possible to download the MP3 or AAC files themselves. Enterprising programmers have seized upon this porthole ever since, creating tools with which one can acquire shared iTunes tracks without even needing to engage in the social interaction of asking for them.

(Each successive update to iTunes has attempted to disable these and other “helper” programs, but most of the attention has been focused on killing the Windows-based utilities. For whatever reason, Mac tools for downloading shared iTunes tracks have managed to survive.)

Now, imagine that one were to have such a tool in an environment such as my workplace, that features some dozens of computers running iTunes at any one time (this effect would of course be multiplied in a college dorm environment). The need to establish and maintain social relationships as part of the currency of filesharing is removed, and one is free to download shared tracks anonymously. Said downloading would be free of many of the vagaries of P2P networks, with a stable LAN connection and a pre-screening process by your workmates, who presumably wouldn’t keep poorly encoded music in their iTunes Libraries.

Of course over time using this tool one could identify trends in P2PLAN downloading, as a scan of multiple users on the network would reveal improbable clusters of duplicated tracks (in the libraries of downloaders who do not take advantage of iTunes’ ability to share merely a controlled subset of one’s complete library).

But most significantly this would allow a user to expand one’s music library at an extremely rapid pace (e.g., 2GB in a week) without cost or emotional investment.

Now should the RIAA’s representatives or allies raise a hackle at such behavior, I would remind anyone so hackled that the ongoing clusters of RIAA’s lawsuits have not targeted those who download but rather those hosting the files for downloading.

In this case, those sharing their music through iTunes are engaging in what Apple has established with the RIAA as non-infringing activity. It is the hypothetical downloader in this scenario (read: don’t look at me) who is prying open iTunes’ sharing feature to get stuff for free.

So then is it the creators of the downloading utility that should be targeted as “bad”? This is of course tied into the current Grokster Supreme Court case, in which corporate media is attempting to shut down technologies that are used for copyright infringing activities, regardless of whether the technology has a legitimate application.

Which legitimate application I use at home when juggling music between three computers on a network. For a track I have bought and downloaded from the iTunes Music Store I have been granted the rights to host that file on up to five computers, but there is no straightforward way to transfer that file from one machine to another.

Using one of these utilities to transfer the song seamlessly across the network to another of my iTunes databases makes this non-infringing activity a breeze. In fact, it only highlights the absence of such a feature in iTunes itself.

One of the many ways iTunes makes it inconvenient for me to enjoy the music I purchase, even when I am not infringing anyone’s copyrights.

* * * Quick Blessings Count * * *


I am neither a Dave Matthews fan nor a Windows user. Which is good because apparently the current DMB disc cannot be ripped on Windows machines. Upon insertion, a program attempts to install that will allow you to copy rights-protected Windows Media 9 files of the CD’s content to your hard drive. But that means they won’t play in iTunes and won’t transfer to an iPod.

Mac-using Dave Matthews fans have no such problem.

So, I’m just saying if I were a Dave Matthews fan, that would be blessing number 2.

Before iTunes

Very few of the millions of iTunes users worldwide have much need for more than a small subset of its features.

I would imagine that for most users iTunes functions primarily in service of the activity of listening to music, whether on the computer or an iPod.

For a smaller, but significant number, iTunes helps feed (or even inspires) a second, distinct leisure activity: acquiring music. Whether ripped from CDs or downloaded from P2P networks or purchased from the iTunes Music Store, hours of time are consumed each month simply adding music to the Library. My coworker KT, a compulsive acquisitioner, bemoaned that the number of songs in his iTunes Library that he had never listened to threatened to surpass the number to which he had.

I’ll discurse upon this hobby/addiction in a future post, but would like here to focus on a third leisure activity toward which iTunes applies its computing power, one enjoyed by a passionate minority, the pastime that elevates this “jukebox” beyond web browsing and email to make it my favorite reason to own a computer: database management.

It is a hobby but implied by Nick Hornby’s pre-digital High Fidelity, in its leading male characters’ compulsion to generate ranked lists of songs, albums, etc. according to a narrowly defined set of criteria. For the most part the characters in Fidelity create their lists in their heads or in conversation, their existence recorded only through Hornby’s printed prose.

But place that book forward into 2005 and I expect that iTunes and the playlist structure on each character’s iPod would be integral to the story. Moreover the characters would be posting comments to 43 Folders arguing whether their various top ten lists should be entered into the computer using DEVONthink or Hog Bay Notebook.

Because the compulsion Hornby captured so well in his novel of a certain subset of mostly male humans to make lists, to categorize, to sort, to select, to recategorize and recontextualize can turn the at face value humdrum computing task of database management into a thrilling (and nearly limitless) endeavor, particularly when it comes to music, and especially with a database like iTunes.

Before iTunes I had developed my own music database, using FileMaker. Between 1996 and 2001 I had entered 8424 individual records, each representing a song I owned on vinyl, CD, or cassette.

In truth I had created multiple related databases, because FileMaker was at that time a flat (i.e. non-relational) database, and to view the data from different angles multiple databases were required:

Music Tracks: Each record represented a single music track.
Artists: Each record represented a single artist or band, and its associated tracks in the Music Tracks database.
Collection Tapes: Each record represented a single side of a mixtape (or, after 1999, a single mix CD), linked to their individual tracks in the Music Tracks database.
Music Line Items: a hidden database of field relationships that allowed the various databases to speak to one another

(The other thing that allowed the various databases to communicate and to understand, for example, the difference between fifteen recordings of “John Henry”, was that I could not create a record without assigning a unique ID#, typically a three letter abbreviation of the artist’s name–useful for purposes of recall–followed by three numbers. As the database topped three or four thousand it often took several attempts to come up with an ID# that hadn’t already been taken.)

For each track I could record the song’s Artist, Composer, Album, Medium (CD vinyl, etc.), and Length (two fields, Minutes and Seconds). Genre was categorized using checkboxes, a system that allowed for easy genre-blending (e.g. Holiday AND Bluegrass). Fields for producer and record label were of minor interest, but I can’t think of a time when I actually did a search against these.

In the Music Tracks database the “Artist” field was repeating, so that for example on a bluegrass or a jazz recording I could list all of the session performers, and the track would become associated with each of them in the Artists database.

With that kind of cross referencing, for example, bluegrass/old-time bassist Mark Schatz could have more tracks to his name (80) than The Police (78).

With cross-referencing I was also able to identify which mixtape(s) I had placed any particular song on. I could also determine how many songs I had placed on a mixtape (1205, 14.3% of the total songs catalogued), but not–at least not in a manner I could figure out–which songs appeared on the most mixtapes.

The Collection Tapes database was a relatively comprehensive record of my mixtaping history–53 in all, dating back to 1988 (tapes made from ’88 to ’96 I had to re-enter from scratch, having no way to port the data directly from the previous music database I had created in, you guessed it: HyperCard).

But of course this Filemaker solution was a database wholly separate from the music itself, which was recorded on a variety of media scattered through several rooms within the house (or in some cases, boxed up in the garage). But in 1996 I couldn’t have imagined a solution in which the music was contained within the database.

Enter iTunes.

Title, Artist, Album are entered automatically (drawing on data from the error-riddled Gracenote CDDB). Genre is provided, but is typically wrong. Thankfully corrections can be batch applied. Time is a direct measurement, although it’s thrown off by extra silence or other intra-song noise. Point being less time is spent entering data, so more time is available to manage data.

But since the database now = the music, the creation of special custom mixes takes only as long as it takes to apply, rearrange, or filter the data.

Example: Create a mix of music embodying “the Bakersfield Sound”, the electrified 1960s country popularized by Buck Owens and Merle Haggard.

Using the old technology I would do a search in my database for tracks by Owens, Haggard, the Derailers, and Dwight Yoakam. I would attempt to remember which of the individual tracks sounded the most “Bakersfield” to me, and narrow the list to under 90 minutes to fit on a cassette tape. I’d then have to spend the time (typically three hours) to make the tape: picking an order, setting levels, etc.

Now I can batch apply “Bakersfield” to the Comments field to all the songs by these artists, and create a Smart Playlist that captures all songs with “Bakersfield” in the Comments field. In under twenty minutes I can probably listen to a few seconds of all the songs searching for ballads, etc., that don’t belong, and remove “Bakersfield” from their Comment field, thus removing them from the playlist. Done. It’s ready for listening on random play, making it a new playlist every time I listen.

And the beauty of it is that I can add future songs by adding the tag, or remove songs that I don’t feel work, at any time in the future, without having to create an entirely new tape.

I can also make further enhancements to the playlist, e.g. narrowing it to those songs rated at 4 or 5 stars (Best of Bakersfield), or to those songs I haven’t heard in the past three months.

So with all these productivity enhancements I should be spending a lot less time managing my data and a lot more time listening to my music, right?

Of course what I’m winding up doing is spending a lot more time doing both. And add to that even more time engaged in acquiring music, because the easier you can understand what’s in your music collection, the easier you see what’s missing from your collection. It’s a pernicious cycle, and one that has fed the 450 million songs purchased from the iTunes Music Store (of which I can claim hundreds).

More on this in an upcoming post.


the commute
printing to the plotter
survivor/the amazing race
digital asset management
soon enough: home repair/maintenance