[OpenID] Anti-XRI FUD

[Jens Alfke]

On 3 Jan '07, at 11:36 AM, Peter Davis wrote:

> FWIW, there is nothing decentralized about openID.  It's rooted in  
> either
> the '.' DNS namespace (centralized), or via IANA-controlled address
> assignments. Please explain how DNS centralization is any better or  
> worse
> than iNames centralization.

Simple: DNS is better because it already exists, is ubiquitously  
deployed/implemented/documented/understood, and is a mission-critical  
piece of the entire Internet as we know it. It may not be perfect, but  
it's pretty good, and we already depend on it for everything else we  
do online.

So perhaps I should have said that OpenID is incrementally  
decentralized. It introduces no new centralized services or single  
points of failure. The same is not true of iNames, and I'm quite  
reluctant to give my support to any new protocol that relies on  
centralized services.


Kaliya wrote:

> There is more then 'one goal' in the user-centric identity  
> community...we can all work together to get a big vision happening  
> or we can squabble amongst ourselves...the leaders in this community  
> chose the former thank goodness.

I agree; cooperation is good! I work on RSS/Atom stuff in my day job  
and would hate to see that kind of nastiness spread into other  
communities.

I'm in this thread because I want to figure out whether iNames/XRI is  
something that I personally want to put the effort into learning and  
supporting. Some of my bad attitude last night stemmed from my initial  
excitement being thwarted by an inability to find real documentation,  
and the letdown of discovering creeping centralization.


> I-names does not have the problems that sxip had. For starters the  
> IP is in a public roylty free situation [...]

I understand those legal and organizational differences, but the  
situation is still the same from the technical perspective that there  
is a single point of failure, so if one domain's servers go down, or  
their top level DNS record is hijacked, the lights go out everywhere.


> Well read more and if you get a chance to talk Drummond or Andy or  
> Gabe  in person [...]

To be honest, the world is so full of a number of protocols and  
architectures that I'm sure I could stay busy 24/7/365 reading them  
all. I try to read a little about a lot of things, enough to get the  
gist, but I'm not going to focus my energies on something unless I  
think it's either really good, or likely to be very successful, or  
both. At this point I don't see a reason to dig into XRI/iNames. If  
new and better-organized docs appear, that could change.

Currently, along with OpenID, the identity-related technologies that  
really excite me are
*  Unmanaged Internet Architecture (http://www.brynosaurus.com/pub/net/uia-osdi-abs.html 
)
* Petnames (http://www.skyhunter.com/marcs/petnames/IntroPetNames.html)
* YURLs (http://www.waterken.com/dev/YURL/Analogy/)
All of these are 100% decentralized, using DNS only as an optimization.

--Jens
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-general/attachments/20070103/67cd4a0e/attachment-0002.htm>