marypcbuk: (Default)
I've written a lot about developments in identity systems this year; this time I've been writing not about new features but about old problems and whether the new approaches will make a difference. It turns out that some of the old systems provide good principles. If someone changes the address on your credit card but not the address you've set with an online identity provider, the credit card company can cross-check with your preferred address - or they can just choose to trust you. The less information a company keeps, the fewer liability issues. Small pieces, widely distributed; stealing all of my identity would be like a treasure hunt. Plus, why Dale Olds from Novell thinks identity might be the wrong word to use for all of this: read on at Developer Register...
marypcbuk: (Default)
Two pieces published yesterday, one in the Developer Register, the other in the FT Digital Business section, one about Higgins and the attempt to simplify the way developers work with identity, the other about the horrible state of the average inbox and what Marc Smith, Microsoft's research sociologist, thinks software should do about it.

Higgins is one of the interesting individual developments in identity that will go to make up an identity metasystem; enough small pieces and I won't have to call it Kim Cameron's idea for an identity metasystem, or designate it in any way because it will be widespread enough to really be a metasystem. Breaking identity up into little pieces tightly managed is one of those ideas it's easy to dismiss because it's a big thing; everyone has to play if it's going to work because it has to work with everything. It's like my childhood reaction to learning about communism; 'what a nice idea, it's a shame people aren't actually like that' (a hardened cynic by the age of 11). TCP and printer drivers were big ideas; one of them won because it was obviously a better solution, one because it made things easier for users and developers. (Guess which I think is which!). There are enough people and pieces and players and financial penalties coming together that we might get Identity 2.0. I'll be writing more about this for DevReg, covering Intel Research's project and what PGP is up to these days.

SNARF is one of those nifty tools that can dig you out of a hole (I'll point it at the email I skimmed whilst travelling in case I missed anything crucial) but it's only a prototype done to find out what people need. The nice thing about that is that if baby steps are useful, bigger lessons might be another big shift. The principle I took from my AI degree was that we don’t know enough about why we work the way we do to emulate or simulate it usefully, but we do know enough to start making interfaces that make it easier to work the way we do.

Marc Smith is hugely fun to talk to and a joy to interview, because he comes out with lines like No one is giving me more heartbeats per day or more minutes; there is no Moore’s Law for humans. I am not becoming twice as intelligent and half as cheap; if anything the cost is going up and I’m slowing down."
marypcbuk: (Default)
My first piece for the Developer Register is online now, covering Implementing InfoCard. There have been plenty of pieces on the philosophy and the politics of InfoCard and the identity metasystem, but I wanted to concentrate on the technical and implementation details - all four will have to work for anything to succeed. MIX 06 was excellent timing because I was able to get the latest details from the InfoCard team and talk with Kim Cameron and Pault Trevithick together, explaining why InfoCard and Higgins actually complement rather than compete.

Profile

marypcbuk: (Default)
marypcbuk

March 2022

S M T W T F S
  12 34 5
6 7 89101112
13141516171819
20212223242526
2728293031  

Syndicate

RSS Atom

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Jul. 12th, 2025 08:43 am
Powered by Dreamwidth Studios