Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
I, Cringely - The Survival of the Nerdiest with Robert X. Cringely
Search I,Cringely:

The Pulpit
The Pulpit

<< [ Lock and Load ]   |  Tough Love  |   [ Nothing Changes ] >>

Weekly Column

Tough Love: More on Media Storage Lockers and Bob's Idea for Getting Rid of Web Window Shoppers

Status: [CLOSED]
By Robert X. Cringely

Last week's column was all about the idea of media lockers -- servers on which we can load (or accept delivery of) media files that can then be accessed by us through any compatible system anywhere on the Internet. Two examples cited were MP3tunes' Oboe music locker, and Apple's rumored video locker service supposedly to be announced at Macworld next month. The audience of this column being what it is, of course, you launched into nearly a thousand thoughts on what was right or wrong about the media locker concept in general, and these proposals (or rumored proposals) in particular. And there is a lot to be learned from this shared wisdom.

Starting with Apple, readers were skeptical of the Thinksecret rumor, not because they thought it was wrong as much as they thought it was incomplete. The most glaring omission, it seemed to many, was the iPod, which Thinksecret ignored completely. How can Apple enter any new video or audio businesses that don't somehow involve making us buy an iPod? Certainly, at this point Apple is as much an iPod company as a computer company. Of course the readers are right. Whatever Apple DOES announce in January, it is very likely to have an iPod component, which itself brings into question the idea that Apple would be streaming its movies and TV shows, rather than downloading them. How can you stream to an iPod?

Readers thought, too, that Apple couldn't possibly intend to actually stream to us all those thousands of movies and TV shows, either (iPod or no). For exactly as I described, streaming is technically demanding and even market restricting, while plain old downloading works a charm. Maybe, dozens of readers thought, Apple would download the encrypted video but stream the key to unlock it. That may be so. As Steve Jobs said, the point of all this Digital Rights Management stuff is to "keep the honest people honest," so maybe some hybrid distribution system will be good enough.

Certainly, Apple is up to something and several readers pointed to Front Row's ability to play movie trailers almost instantly as proof of that. These would be the readers who weren't complaining about poor Front Row performance (I had some of those, too).

Those who commented on my thought that Apple would have to do a deal with if the company were to actually roll out something like the Thinksecret rumor -- well, they were more cynical than skeptical. Some thought Microsoft's investment in Apple and the patent cross-licensing agreement the two companies signed would have made it a non-issue; Apple would just ride on Microsoft's license from Burst.

Nope. Microsoft sold its Apple stock long ago for a big profit and the patent cross-licensing agreement was for only five years, starting in 1997, so it has long expired. Besides, Microsoft's license from Burst did not include rights to sublicense. Burst may be small, but it is not stupid. Microsoft has in the past been very successful at licensing intellectual property then using its sublicense rights to compete with and eventually destroy the business of the inventor. Burst was not about to allow something like that.

Reader reaction to MP3tunes' Oboe unlimited-capacity Internet music locker centered mainly on the idea that the system wouldn't really have to hold 5000 different copies of "Louie-Louie," would it?

"What happens when two users upload the same exact (down to the bit) copy of, oh, let's say 'Muskrat Love' by Captain and Tennille?" asked one reader who clearly survived the 1970's. "If the hard drive surface actually retains the song despite its cloying sweetness, let's assume the system uses a compression system where the minimum not-compressible unit is not a byte or a word, but a song. In some ordinary compression systems, the first instance of something is stored in its actual full value. Subsequent copies of that same value are replaced by what are effectively pointers or references to the original value. The first instance of 'Muskrat Love' is stored in full on a hard drive, and customer number two also uploads a copy. This second copy would be an identifier rather than an actual copy of the song. Since the copies that were uploaded were identical, each customer still has 'their' copy in the system."

Since most of these copies came from CDs and are therefore genetically identical, this scheme would probably work. It would also reduce Oboe's storage requirements from petabytes down to terabytes and turn the company's prospects from technically-challenged to almost a no-brainer. Well, of course, this is the one point where the otherwise-loquacious Michael Robertson clammed-up, so I'm guessing that means the readers are correct. But I'd also guess that for legal reasons it would be better to compare those checksums up on the server AFTER the upload where the extras can be discarded in a process another reader called "upload compression." Fortunately, this is only a small hardship for Oboe, which, like many Internet services, has a lot more excess upload capability than download, so deleting after upload these duplicate copies wouldn't be painful at all.

But can Oboe use the same trick for music that wasn't taken from CDs? Probably not. Music that arrives at the user's house on vinyl or tape will show enough differences after going through analog-to-digital conversion that every copy WILL be unique. Who cares? Vinyl enthusiasts probably wouldn't use such a service anyway, so this class of recording would probably constitute only one percent of the songs, though 10-plus percent of the total storage.

You can bet the RIAA will be uploading a stack of old Sinatra records just to test the system. And if they don't play back with every pop and crackle intact, the RIAA will sue.

Thinking very much outside their boxes, two readers came up with whole new uses for a service like Oboe. Noting that the upload process is essentially a bit-level examination of your entire music collection, one reader wondered if it could be used to scan for pirated copies. His idea was that many pirated copies will have the same noise signature and this signature could be used not only to determine if the song was pirated, but even to trace the original source. While this is technically possible, it is hardly in the interests of MP3tunes to narc on their subscribers.

The second idea, from The Famous Brett Watson, is much kinder and gentler. He'd turn Oboe from just a music storage system into a music DISTRIBUTION system.

"Suddenly Oboe looks like a great distribution medium for independent artists," wrote Famous Brett. "The artist uploads one copy of the song, and then Oboe sells it to other Oboe users on behalf of the artist, giving the artist their cut. The actual effort involved on Oboe's part is almost purely a matter of accounting; the technical side doesn't even involve making another copy of the data -- they just make a link to the data appear in the user's locker. Oboe looks like a promisingly disruptive business to me."

I LIKE this idea, but it could be easily taken to a higher level. How would Oboe promote music to its users? The easiest was would be by creating an Internet radio station specializing in just these independent artists. The radio station would be free for anyone to listen to, not just Oboe subscribers, so it would effectively promote not just the independent music but also the locker service. Artists would be paid only for locker sales, not radio play. But to encourage people to buy rather than simply mooch on the free radio service - here comes my clever twist - the Internet radio station could be programmed to limit each song on the system to no more than three plays to each IP address.

It would be like one of those disaster radios you have to crank to charge-up: If people don't buy any music, their selection withers and eventually fades away. And what's wrong with that? THEY WEREN'T GOING TO BUY ANYWAY, so losing these customers wouldn't be losing customers at all. I really like this technique because it eventually bans non-buyers from the store completely. Imagine how useful this concept would be in almost any retail industry where smaller crowds of buyers are always preferred to larger crowds of lookers.

I wish I'd thought to patent that.

Comments from the Tribe

Status: [CLOSED] read all comments (0)