1. no thanks. i'll take my files free and unrestricted, thank you very much.

Posted at 4:23AM on Dec 19th 2005 by catflap

2. There is nothing wrong with cental databases, and I like the notion of a distributed database even more. But the problem with DRUMS is that it introduces unnecessary burdens of control. Why do we need a central authority that determines whether or not a piece of work is DRUMS-compliant?

There are many problems with such central points of control - one of the most obvious is that they obstuct the creative process. Take something like the Wired / Creative Commons CD for example. Musicians are free to sample any of the songs of the CD and incorporate them into their own works. After they have done so they might upload their works on the web and release them under the same Creative Commons license themselves.

Now imagine a P2P system that uses DRUMS to filter content - I believe that's what Scott Matthews has in mind as one of the premier applications for DRUMS. The original Wired CD cold be shared easily.

But every offspring song would have to go through a clearinghouse before it was avaliable on the system - probably producing costs for the musicians, since DNS registration isn't free either. Right now the Creative Commons website features 193 remixes of the original CD. I'd imagine that there are hundrets more out there on the web somewhere. DRUMS would very likely only capture a small percentile of these works.

So why bother with a centralized clearinghouse? I'd suggest the notion of a public peer review of the submitted information instead. Think if it as Wikipedia vs. Encyclopedia Britannica - and ask yourself which one is more up to date.

Posted at 4:23AM on Dec 19th 2005 by Janko Roettgers

3. Would you similarly fault the central and authoritative DNS database that drives the Web as stifling creativity?

Of course you wouldn't.

Perhaps you favor the notion of some ACS ("alternative compensation system")? If so, wouldn't that similarly need a central and authoritative database?

-Scott

Posted at 4:23AM on Dec 19th 2005 by Scott Matthews

4. One other point: following your own comparison of Wikipedia to Britannica, let's remember that both are dependent on central databases -- is just that the rules of maintaining "trustworthiness" are a bit different.

-Scott

Posted at 4:23AM on Dec 19th 2005 by Scott Matthews

5. Sorry, one more point...

I don't see filtering P2P as the appeal of a system like DRUMS.

The appeal is its potential to create a new infrastructure for the discovery of art.

Following your example, most people were exposed to the Wired / Creative Commons CD because it was in Wired magazine and not because it was licensed for sharing.

The point is, with P2P you already have to know what you're looking for -- it's not a tool for discovery.

On the other hand, a system like DRUMS could provide a neutral infrastructure to drive the discovery of art.

-Scott

Posted at 4:23AM on Dec 19th 2005 by Scott Matthews

6. P2P used to be a great tool for content discovery, and in some cases, still is - e.g. where you can still browse people's folders, chat and search for similiarities. The key to discovering new content is community, not authority. And you certainly don't need a central gatekeeper to build P2P-based communities.

Plus I can't really see the similarities between a DRUMS-database and DNS. You built your proposal on the assumption that they are similar, and one could certainly build a system that distributes database information in the same fashion that DNS root server entries get distributed.

But does the task require such an architecture? I don't think so. DNS is a - some people would argue clumsy - way to administer limited ressources. How is this comparable to an art metadata space?

And what can be achieved with a central, controlled, authoritative registry that a dezentralized, semantic web approach could not deliver?

Posted at 4:23AM on Dec 19th 2005 by Janko Roettgers

7. Scott, regarding your Wikipedia comment: I said I don't oppose databases, but the gatekeeper authority that your model is based on.

Posted at 4:23AM on Dec 19th 2005 by Janko Roettgers

8. Janko, do you favor the notion of some new system that would compensate rightsholders for what is now illegal filesharing?

If you do, wouldn't a trustwothy registry be a prerequisite? Should such a registry be closed and proprietary, or open for all to use?

-Scott

Posted at 4:23AM on Dec 19th 2005 by Scott Matthews

9. How exactly would DRUMS help to compensate rights holders for the use of their works in P2P networks if DRUMS would "compete with P2P", as you state in your proposal?

And how can it be open if it is based on a central reigistry "with an authority (or a handful of authorities)"?

Plus if you want to build an open system, don't use DNS as a blueprint, because it not open and participatory at all.

Posted at 4:23AM on Dec 19th 2005 by Janko Roettgers

10. By "open" I mean that anybody can search it, and build applications on top, like DNS. But only trusted authorities would change the contents, like DNS. Imagine how much of a mess the Web would be if DNS followed the Wiki model.

But now it seems you're avoiding the question.

Do you favor the notion of a new system that would compensate rightsholders for what is now illegal filesharing?

If you do, wouldn't a trustwothy registry be a prerequisite? Shouldn't such a registry be open for all to use?

-Scott

Posted at 4:23AM on Dec 19th 2005 by Scott Matthews

11. Scott -

How is this different from Bitzi or MusicBrainz?

While I agree that such a central meta data source *could* be a key component of a "new infrastructure for the discovery of art," it isn't really necessary to produce a P2P based recommendations system today given the data that is out there...by that I mean even crude data sources such as the mp3 filename the user has assigned, its SHA1 and the data available from MusicBrainz can get you pretty far down the discovery engine road.

At upto11.net we've found that you can produce great recommendations, and help people discover something that's *new* to them (meaning that they didn't know its name prior to finding it), using existing open source "answer keys" to refer to. Would things have been easier and better had there been a more definitive meta data source like DRUMS? Sure! But my point is, you don't need DRUMS to create the P2P/Community-based discovery service like the kind that Janko describes.

p.s. our site goes live in the next day or so.

Posted at 4:23AM on Dec 19th 2005 by David Sabel

12. Hi David,

I'd say that a very big chunk of the work done by Bitzi, MusicBrainz, SnoCap, ASCAP, and so on goes into reproducing largely similar databases. What if, instead, that database was provided to all, as a common and shared piece of infrastructure?

My suggestion is that perhaps the notion of a trustworthy registry could appeal to both "conventional rightsholders" and "new visioneers" -- and if so, that both "sides" should come together to nail the details.

I don't suggest that the only way to discovery is via some registry -- but simply that, to follow your example, recommendation engines (and lots more) would benefit tremendously by having access to such a registry, and likewise to have "WorkID's" that could be shared between sites.

-Scott

Posted at 4:23AM on Dec 19th 2005 by Scott Matthews

13. No doubt it would be more efficient. But who pays for this common and shared piece of infrastructure and why?

"Conventional rightsholders" folks would likely focus on control/filtering features while "new visioneers" would likely want "help me get discovered" features.

While it might be ok to have a DB with lots of columns that meet the needs of only a subset of the constituents, if each "side" focused on populating the columns they got the most value out of and blew off the others, it would render the DB far less useful.

And on top of all this, you'd need to get the parties to agree on who has access to edit/ update the database and where the "atomic" items list is coming from (the sha1/filename entities) that all this meta data is being applied against - i.e. who verifies and how that an sha1 *IS* the song that someone purports it to be.

Posted at 4:23AM on Dec 19th 2005 by David Sabel

14. Hi David,

Conventional rightsholders are certainly interested in getting people discovered. And the new visioneers who favor an ACS (alternative compensation system) would likely need a trustworthy registry anyway.

Who would pay? Well, first we'd need to find out if something like DRUMS feels like a step forward to the different "sides." If it is, my suggestion would perhaps be to turn to the tech industry for funding.

(Regarding the number of "columns" in the database, I'd suggest that we start with the bare minimum. For example, is "WorkID" even appropriate? How do we define it?)

-Scott

Posted at 4:23AM on Dec 19th 2005 by Scott Matthews

15. Hi Scott -

Agreed. Both "sides" share in the motivations of

"let's get tracked" and
"let's get found"

but my point is the first is of less importance to a CC artist in the absence of an ACS and the second is of much less importance than the first for control oriented "conventional rightsholders."

As far as who pays - why the tech industry? And who specifically? Do you mean the download stores?, the P2P app companies? the portable device mfgs? The streaming svc sites? Oracle?

As far as the "workid" concept, I think you have to go as atomic as possible - the sha1/filename pair. From there, you move up a level where the meta data would be applied (the workid perhaps). All sha1/filenames related to a workid would share in that workid's meta data. The big challenge is in defining the rest of the columns so as to determine the level of uniqueness you need to represent and then gathering up and relating all the *right* sha1/filename combinations that map to that entity.

Posted at 4:23AM on Dec 19th 2005 by David Sabel

16. The tech industry is the primary beneficiary of p2p (especially broadband ISPs, storage manufacturers, MP3 Walkman devices, etc.) -- and they'd stand to gain a lot more if they could aggressively participate.

btw, I don't think SHA1/filename would be the most appropriate WorkID. Apart from the questions about the long-term prospects for SHA1:

http://it.slashdot.org/article.pl?sid=05/02/16/0146218

It seems the WorkID should represent the "idea" of the work, rather than a file. For example, a given song might be compressed to MP3, OGG, WMA, RM, AAC, and so on -- but they're all representations of the same song.

So, even just that -- establishing a common ID system -- would be tremendously useful.

-Scott

Posted at 4:23AM on Dec 19th 2005 by Scott Matthews

17. Scott -

Sorry I wasn't clear - I was suggesting the same thing as far as the workid

many different sub entities mapping to one "container" entity which keeps all the meta data

that way, different encoders, bit rates, filenames would all be tied to one "work" that has a set of meta data

Posted at 4:23AM on Dec 19th 2005 by David Sabel

18. we are Live On The Internet streaming Our radio station from salina,kansas usa, our music format:R&B, world-news,weather,sports & talkshows and more please Call(785)-820-9512 for your song request or comment on our programming or email us ksjfnetradio@cox.net

Posted at 4:23AM on Dec 19th 2005 by ksjfnetradio

19. i happen to think that internet ardio streaming is the way of future radio, its clean and it is easy to setup, everything is right at your pCs fingertips, i would just like to commend anyone with the tech-mind of creativity, In all aspects of the knowlege, there is a new way of communication in a small city or area that leaves out the Other kind of radio, But;; KSJF-NETRADIO Is Here to stay and you URL http://www31.brinkster.com/ksjfnetradio

Posted at 4:23AM on Dec 19th 2005 by ksjfnetradio

20. Can I just point something out? If the idea is for this system to compensate rightsholders in some way for files that are shared on p2p networks won't there need to be compensation for the administrators that compensate them? QED: not surprised Scott and David are supporting it the potential financial rewards are huge no matter how altruistic you might say your motives are!!

Posted at 4:23AM on Dec 19th 2005 by Beetlebum