RE: non-http uris

Mike Schinkel writes:

> Since you believe so strongly in having a geoloc: scheme, why 
> not create an implementation that requires it and attempt to 
> see it become widely deployed/utilized?

I have very mixed feelings about this.  In general, putting trial 
technology out there and letting the market decide is a good thing, but 
I'm not sure there are good ways to experimentally "squat" on parts of URI 
space.  If Erik starts using geocode: then (a) I would think he should 
register it and (b) that would mean that for all time nobody else could 
use it either for an entirely different purpose or, more likely, for 
geographic codes implemented in a different way.  My intuition would be to 
not deploy a new URI scheme unless you, and preferably a significant 
fraction of the affected community, agrees that it's a good long term 
investment.  Just my opinion, and it may well be that there are 
conventions for experimenting with schemes and that I am unaware of them. 

Also:  the fact that something like a geocode scheme starts to get 
widespread adoption doesn't entirely settle the question of whether it was 
the right architecture for the greater good.  It's quite possible that 
it's conveniently meeting the short term needs of early users, while 
sacrificing potentially important long term benefits such as integration 
with the browsable Web, etc. 

Noah

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------

Received on Thursday, 3 January 2008 21:30:11 UTC