6 Jul
2010
6 Jul
'10
1:08 p.m.
Why would it be a disaster? Could you elaborate more?
You're turning an opaque identifier into a semantic key that eventually will get overloaded to handle whatever kinds of variability people come up with. This happens in databases all the time, and it leads to maintenance problems. Your solution also doesn't work with existing standards for the representation of attributes, which include the use of opaque URNs to identify attributes that have unambigious names. The problem here is the protocol, so that's where the solution needs to be. Your protocol needs to be fixed to handle the extensibility you need in negotiating information release. -- Scott