[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message] RE: URI concerns continue
All forms of human communication require consent. So, shaky ground is the best we get. Your post is a fair condensation of the original discussions on URL/URN/URIs as namespace identifiers. Essentially, if you conflate identity, name, and location, you face some problems and you have others if you don't: 1. Record of authority: who owns it, who controls it, how is it known and controlled; can this be cited and tested/proved 2. Resolution: is a name a location identifier and can it be resolved. Identity may have a mathematical definition but it is for the most part, philosophical in implementation. That is why I regard those who think they can always discriminate rhetoric from logic and assert the superiority of logic with suspicion. Again, what is the age of the dead man? Answer: it is not a fact. It is derived and without stating the semantic of the derivation, arguable. This is the central problem of contract/consent based systems: semantics. It is the central dilemma of the so-called semantic web. 3. Is there a means to determine a superstitious assertion? In behavioral science, this is a behavior acquired as a result of an observation leading to a false association of cause and effect, thtat is, one that is only reliable some of the time and then only coincidentally. See B.F. Skinner. This is a central issue of the reliability numbers in logistics analysis. Essentially, URI/URN/URL are simply system specific aggreements about these problems and are based on a working philosophy (say, implementable) about how things will be named, found, and discriminated on a system named the World Wide Web. The only guarantees are behavioral: repeatable results at some satisfactory level of sustained repetition (eg, nothing can guarantee the network is always up or available). The question then is should such a system specific naming philosophy become part of the information which that system handles? The SGML answer was no. Public identifiers (contractual and yes similar to ISBNs) are there for formal naming. System identifiers are there for resolving a location to a thing which the FPI may or may not name (not required in the instance). There are few guarantees in the FPI; it is a convenience for a separate process of contract based communication. System identifiers are guaranteed; must resolve. A third entity, the catalog, was created to enable a system implementor to agree to a means to tie these together for authoritative resolutions. The WWW answer has been we should be able to resolve or not resolve the location by a name in which the system implementation is the guarantor insofar as it can be guaranteed to have a unique identity. This simply avoids the catalog but it makes it difficult in all cases to determine the correct behavior. There is room for both approaches in the identification of the schema or DTD. In the namespace of an aggregate, we don't seem to be in any more or less trouble unless we attempt to resolve the value of xmlns. Heretofore, that was considered outside the scope of the namespace spec, or so we were told until implementors began to resolve them simply because with a protocol identifier, they could in some cases. On the other hand, that isn't guaranteed. Len Bullard Intergraph Public Safety clbullar@i... http://fly.hiwaay.net/~cbullard/lensongs.ram Ekam sat.h, Vipraah bahudhaa vadanti. Daamyata. Datta. Dayadhvam.h -----Original Message----- From: Lee Anne Phillips [mailto:leeanne@l...] Sent: Tuesday, July 11, 2000 9:44 AM To: XML-Dev Mailing list Subject: Re: URI concerns continue I agree the URI's are built on quicksand for the most part, once you leave the shaky ground provided by URLs, which have their own problems. The whole concept of "authorities" presumes an exactitude in the world that simply doesn't exist in any real sense. Even URL's can be (and have been) subverted. The authority of URL's depends entirely on the goodwill and skill of all (or almost all) participants in the DNS scheme of things.
|
PURCHASE STYLUS STUDIO ONLINE TODAY!Purchasing Stylus Studio from our online shop is Easy, Secure and Value Priced! Download The World's Best XML IDE!Accelerate XML development with our award-winning XML IDE - Download a free trial today! Subscribe in XML format
|