I had a nice long phone chat with John Hiler of Xanga and Microcontent News. He's just recently read my thesis draft, and is intrigued by the idea of using them in larger-scale blog services. The main application will probably be to suppress spam and abuse, and perhaps to ease support.

The prospect of having the trust metric ideas implemented for a mass audience site is exciting. I'm sure it will happen sooner or later, and this might just be the sooner.

Teletruth

Bruce Kushnick has written an "Unauthorized Bio of the Baby Bells", available for free download. I don't really have time to read through it right now, but I'm intrigued by his thesis that the telcos in this country are basically a scam. He comes across as a little bit of a crank, but my gut feeling is that he's substantially right.

While my experiments so far with VoIP have shown it to be not quite ready for prime time, I strongly suspect that it will become so in a couple of years or so, at which time the telco business will find itself quite disrupted. For example, imagine a cell phone that has an onboard 802.11 radio, and logs on to the Internet (instead of GSM or whatever) whenever it finds a base station within range. That would be cooool.

I'd be interested to hear what any Advogatan (or critical thinker from any corner) has to say about this book.

(via JOHO the blog via boingboing)

DNS, phone numbers, and ENUM

One of the many problems that needs to be solved before Internet phones become usable is directory services. What kind of name space do we want for IP phones? One obvious choice is phone numbers, which have the advantage of smoothing migration from POTS. Other choices include DNS-based email-like human readable addresses, or entropy-rich tokens such as hashes of public keys. As a stopgap, people might use IP addresses (these seem to be standard for NetMeeting and friends), but those have serious usability problems.

One proposal to map phone numbers to IP addresses is ENUM, which hacks the phone number namespace on top of DNS. The basic idea seems sound, but in practice it's almost certain to inherit the problems of DNS, not least of which is corrupt governance.

In particular, it sure looks like the telephone companies are going to be the primary owners of namespace chunks. This is the worst choice possible, because they have a powerful business interest to fuck up deployment of IP phones as long as possible, so as to preserve their price-gouging in the POTS business.

I want to be able to map my phone number to my Internet address. I'm skeptical that the ENUM implementation will actually be able to provide this service for any reasonable price and terms, but of course I'd love to be proved wrong. In the meantime, people interested in making IP phones happen should be thinking about other forms of namespace. Or, perhaps the right answer is an independent service that also uses the phone number namespace, but is actually trustworthy, as opposed to the telecom industry. Worth thinking about, in any case.