> - it uses DNS to gaurentee uniqueness The DNS does not guarantee uniqueness outside of a TTL; you need a timestamp to accomplish this. > - next is a domain name in *small caps* Do you mean lower-cased? > - a standard httpish path per URI specification. There are a lot of "does this match that" rat holes here. as an example, are: pkg://clarkevans.com./ and pkg://clarkevans.com/ the same? How about pkg://clarkevans.com/data-type and pkg://clarkevans.com/data-typ%BC or pkg://clarkevans.com/data-type%bc Saying something other than "httpish" may be useful here. > > Examples: > > pkg://clarkevans.com > pkg://clarkevans.com/data-type > pkg://clarkevans.com/2002/my-data > > The end goal is (a) to have a DNS based URN and (b) not have > this URN imply any sort of access mechanism (http,ftp,etc.) > > For XML-DEV people... What do you think? Pretend for a moment that > this is 1997 and XML is just about to emerge. Would something like this > have helped? Please limit comments to what the URI change will impact, > not about general problems with XML namespaces and such soap boxes. > Ok. Now, given today, if it took a while to catch on would it improve > the situation (mod other namespace problems)? If not, why? > > Anyway, if you are interested, I'm going to attempt to carry > out this conversation over on the ietf's discussion list at > http://www1.ietf.org/mail-archive/ietf/Current/index.html > in a manner independent of XML. I'll be observant of replys > to the xml-dev list, but I'm not at all interested in the > xml-namespace soapbox. ;) The URI mailing list may be a better home. Ted Hardie