Subject: rtprio(1) plays mp3s, hangs record companies, and passes death sentence on the banks
From: Ian Grigg <iang at systemics dot com>
Date: Sat, 23 May 1998 16:04:36 +0200

<someone> said:

  > Just discovered ""idprio"", the proper way to run processes when
  > the system is idle.

Aha! What's more, ""rtprio"" is the proper way to play MP3 files. No more glitching when netscape says ""me mozilla, me crunch.""

Speaking of MP3s. Last night, at the way-too-much 5th birthday bash of xs4all, we saw the one and only mpman outside Korea. (Peter, who wasn't selling them, also had more he could sell you tomorrow :-)

It's about the size of a skinny fag packet, with a bunch of tiny buttons and a little display (I think, my brain was hazed by the alcohol at that stage). Sound was great, although standing in the middle of the paradiso dance floor at the time was not a good test. 900 nlg (about $450?) for the 64M version. Peter was also using it just to carry data around, like a small disk drive. Why ever not?

Which takes me back to dinner at Zooko's last year where Sebastian explained his design for them. I passed the death sentence on the record companies then, last night the hangman arrived. I'd even go so far as to say that this is the clearest example of how the ""new tech"" can change things. Kudos to Sebastian for being ""first"" on that one.

Back to the mpman and the end of the world as we know it. The concept needs about 64M to make it work, and this sort of memory is now available for those plugin memory banks for Psions, and will soon be plausible in the new generation of Psion-killers, the Palm Pilot from USRobotics.

Now, one of the things that palmtops always lacked was a killer app.

If you look at the devices, they are mostly good at organising, and perform their job as well as a Filofax, more or less. So they have enjoyed steady growth, with obvious advantages (you can add a spreadsheet, etc) and obvious disadvantages (battery failure kills the memory, etc) over the paper equivalent.

MP3 is the killer app for the palmtop. Suddenly, we have this wonderful new app that makes these a consumer tool. MP3 playing on a little device that also handles names&addresses, notes, spreadsheets and household accounts, is such an essential feature of ones life that anybody with more money than an Indonesian depositor must, must, must have one, and now!

Which means that we are about to witness an explosion of these devices. The numbers that USRobotics have sold are going to appear like small change when the non-computer literate get the idea that all this power is good for something. Small, personal organisers with displays, heaps of storage, and a secure connection to a network.

Which, finally, irrevocable, brings in the last piece of the puzzle that has bedevilled our efforts over the last 4 years or so in the production of payment systems: the secure device.

Every architect worth his salt in the spice-fields of Internet payment systems has worked out that you need a secure device on which to store the keys. Without getting into the logic, the ideal was a small computer like a Psion. Or a smart card, but they were an unfortunate compromise that lacked the ability to tell the user what was being signed for. Either way, it isn't a PC, and it definately wasn't a PC for Rabobank last week, who found the wiley hacker had convinced a few customer cents to go in non-customer-directed directions.

Smart cards had other subtle advantages. They were big (in terms of infrastructure) and expensive (in terms of implementation), something that I claimed in my 1994 Critique made them attractive to banks, in a way that the Internet was not. As laying out a smart network was only for well-heeled institution, and as it solved, mostly, the nightmare scenario of Internet money being uncontrolled on the one hand and indefensible on the other, this was obvious strategy for the banks.

And so it came to pass. Banks around the world united in competition for the smart card system to give everyone the 21st century technology that they told us we deserved. Central Banks cheered from behind, and rushed out a series of well researched papers that said this was the right thing to do. Internet connectivity was added as an appendix, because that was right, too.

With an MP3-driven explosion of pocket computing power, the strategy is crumbling like castles in the face of seige cannon. The user infrastructure cost barrier was the one that the banks were desparate to pay for, as the two other barriers to payment systems, reputation and customer base, are proving insufficient; witness the shenanigans in UK supermarkets, the onslaught of US NOW accounts and those sneaky Finnish telcos.

The death sentence is hereby passed on the banks. Judged by the impartiality of Hettingarian technology, the banks will go to the grave to a dirge of mp3-playing pocket HINDEs. Some will lament that taking the payment systems subsidy away from the banks was not meant to kill them. Oh dear no, we didn't intend to do that, did we?

Exposed in all their ignominious imbalance, banks that concentrate on just banking (the taking of deposits and the making of loans) will collapse in the face of their two worst enemies in glorious alliance - the investor and the freely accessible markets. It has been shown a long time ago in pre-Internet times that making loans to the banks was based on convenience, and the free investor, Markowitz style, was as capable of securing returns for their money as any bank.

What to do with the money collected was really the banks' speciality, and this was based on the two-pronged attack of the payment system (needed for any working business) and the special knowledge that a bank had in assessing a loan - access to cashflow records. This, in effect, meaning that they had an advantage over the other lenders.

Have you ever wondered why a bank likes your business to have just one banking relationship? Bankers will be, soon. Indeed, their last expiring thought will be ... I wonder what it is like when my borrowers have dozens of payment systems at their disposal?


who's playing sweet financial funeral music at real-time-priority (rtprio(1) on FreeBSD) today.