Cart
[Home] [By Thread] [By Date] [Recent Entries]
If the name was the thing, I might be persuaded, but it seems the name follows the application. I don't think a protocol spec followed by a path is much of an identifier until it is married to a domain octet somewhere. Yes, I read the specs. I consider them a sort of adaptive justification. It was a handy thing to do and works in the niche. A system ID invaded the information domain. If it turns blue and is clickable, it's a control, not an identifier. Of course, hidden in a structure where I can't click on it, it can be whatever that structure allows. Len http://www.mp3.com/LenBullard Ekam sat.h, Vipraah bahudhaa vadanti. Daamyata. Datta. Dayadhvam.h -----Original Message----- From: David Brownell [mailto:david-b@p...] > URIs aren't designed to > be namespace identifiers per se. They are co-opted into > that as a side effect of their uniqueness. If that "I" in "URI" didn't signify "Identifier", I might be tempted to agree, but as it is ... I can only hope that you mean to say that using identifiers to denote one kind of thing ("namespace") is as arbitrary a choice as using them to denote another ("person", "location", and so on). Identifiers are void of meaning. Meaning comes from the context in which they're interpreted. Much confusion came from assuming that the only context in which URIs would be used is a "fetch contents" operation, despite specifications clearly stating otherwise.
|
