Conferences
FOM: Alternative Compensation Panel
The Alternative Compensation Panel at the Future of Music Summit was the one panel I was most looking forward to. Having read and panned most of the proposals on the table, I was hoping to get more dirt on these proposals and see if I missed anything. There was bound to be something that would make these proposals make more sense.
Unfortunately that was not the case. Given the scope of the panel and the audience in question, the panel couldn't do much more than introduce the concept and debate the higher level points of these proposals. After the session I asked about my biggest concern: the issue of fraud. I know that PayPal must process upwards of 1M transactions a day before (I believe 1M was the number -- if you know anything about this, please post a comment) before they can make a profit, due to the vast amount of money they need to spend to prevent and address fraud issues.
The answer that the panelists gave were not satisfactory enough -- at least Chris Amenita acknowledged that fraud is a large concern:
Chris Amenita:
ASCAP uses an outside agency to select what's being sampled, so that people inside our organization can't stuff the ballot box. There are technologies that one can use, but it's never going to be 100%, and people might be able to rig the system - it's a possibility. The fact that the large amount will go to fraud-prevention is important.
Neil Netanel:
I don't think there's much of a fraud problem when you're not talking about hundreds of thousands of individuals, but you're talking about manufacturers in business. If you engage in some less precise methods that do some statistical sampling but more than are in effect today, groups like ASCAP and BigChampagne can help.
Neil Netanel must not have understood my question. If a new compensation system uses statistical sampling to determine which artists to reward, then the small artists will be left out of the picture. If a census method is used, then this census data must come from P2P clients. How are the P2P clients going to differentiate between legitimate P2P downloads and fraudlent downloads initiated by bands (or just plain crooks) trying to inflate their download numbers?
I think I am now starting to understand where the confusion between the panel and myself comes from. I am a computer geek and I hack on code for a living (well, kinda) -- these folks are economists. I don't think they are aware of the technical limitations that need to be overcome in building a system that accurately collects information on who needs to be paid. Just as Napster 1.0 could not effectively filter out infringing works, these systems will not be able to do a good job gathering valid census data. This is a really hard, if not impossible problem. Period.
Posted by Mayhem at May 5, 2004 03:16 PM
It was a great question. You were exactly right about the audience thing -- I would have loved (LOVED) to get into the nitty gritty technical details, diving into the various ways to game the system and technical and legal measures we could put into place to prevent and mitigate fraud on the system. But The audience was made up mostly of musicians and music-industry people, and the most important thing was to get them to understand (and hopefully support) the basic idea and leave the legal and technical issues to other discussion fora. Had we gotten into the stuff that really interests me -- for instance, legal stuff like how we make a system that doesn't violate international treaties, or technical stuff like how we do this without outlawing open-source P2P software -- we would have lost most of the audience.
That said, I completely agree that most of the people who have made proposals on this aren't tech people -- they're law professors. For a particularly egregious example, see Lon Sobel's piece on ACS, in which he reveals a fundamental misunderstanding of the way both the Internet and personal computers work. (http://www.quicktopic.com/19/D/RkCdhaVqwpBck.html)
What I tried to do in my piece (http://www.joegratz.net/files/Joseph%20Gratz%20-%20Reform%20in%20the%20Brave%20Kingdom%20Dec%2019.pdf) is come up with a way to do this relatively accurately without breaking existing P2P systems or basic Internet architecture. I propose doing the measurement based on what people are sharing, 'cause that information is public by the nature of the P2P network.
(Incidentally, though I don't mention it by name, I'm basically assuming that we build MusicBrainz into all P2P apps -- at least the TRM bit -- for fingerprinting.)
I'd be interested to see what you think of the technical feasability of my proposal.
THEY BETTER THINK OF SOMETHING FAST,AS THERE IS ALOT OF PIRATING,STILL THE MUSIC GETS OUT THERE.THE WEB SITES THEMSELVES ARE RESPONSIBLE FOR KEEPING TRACK.I GUESS ITS A QUESTION OF PRIVACY OR NOT , I SAY PRIVACY,THE MUSICIANS WANT THIER WORK HEARD NO MATTER WHAT, MONEY IS SECONDARY, BUT LETS TRY TO GET IT TO THEM ,SHALL WE?
