- From: David Booth <david@dbooth.org>
- Date: Wed, 05 Aug 2009 23:06:36 -0400
- To: Larry Masinter <masinter@adobe.com>
- Cc: www-archive <www-archive@w3.org>
Hi Larry, On Wed, 2009-08-05 at 12:46 -0700, Larry Masinter wrote: > > Compare this with a corresponding http URI such as > > http://filbert.example/nadia/woosel/yes > > In this case, the chain of authority looks like: > > 1. AWWW delegates to RFC3986 (because the whole thing is a URI) > > 2. RFC3986 delegates authority for http:* URIs to the owner of the > > http scheme > > OK so far > > 3. Owner of http scheme delegates authority for > http://filbert.example/* URIs to the owner of filbert.example. > > No. The "Owner" of the http scheme is the IETF. The http > URI scheme is defined in a Draft Standard for HTTP, following > IETF process. The current definition doesn't delegate the > "authority" for http://filbert.example/* URIs to the > "owner" of filbert.example, it only allows for the HTTP > protocol to be used to connect the HTTP server being run > at the computer with the DNS entry for filbert.example. Yes, I realize that the HTTP spec doesn't say that. That could either be viewed as a gap that should be corrected or one could view it as unnecessary for the HTTP spec to make that explicit delegation, since the layering of extension semantics on top of http URIs is non-destructive (monotonic). It is not correct to say that the HTTP spec 'only allows for the HTTP protocol to be used'. The HTTP spec only *defines* the use of that protocol with the http scheme, but it does not prevent other protocols from being used as well. Section 3.1 of the AWWW addresses this explicitly: http://www.w3.org/TR/webarch/#dereference-uri [[ Although many URI schemes (§2.4) are named after protocols, this does not imply that use of such a URI will necessarily result in access to the resource via the named protocol. Even when an agent uses a URI to retrieve a representation, that access might be through gateways, proxies, caches, and name resolution services that are independent of the protocol associated with the scheme name. Many URI schemes define a default interaction protocol for attempting access to the identified resource. That interaction protocol is often the basis for allocating identifiers within that scheme, just as "http" URIs are defined in terms of TCP-based HTTP servers. However, this does not imply that all interaction with such resources is limited to the default interaction protocol. For example, information retrieval systems often make use of proxies to interact with a multitude of URI schemes, such as HTTP proxies being used to access "ftp" and "wais" resources. Proxies can also to provide enhanced services, such as annotation proxies that combine normal information retrieval with additional metadata retrieval to provide a seamless, multidimensional view of resources using the same protocols and user agents as the non-annotated Web. Likewise, future protocols may be defined that encompass our current systems, using entirely different interaction mechanisms, without changing the existing identifier schemes. See also, principle of orthogonal specifications (§5.1). ]] > > > 4. The owner of filbert.example delegates authority for > http://filbert.example/* URIs to the Filbert specification. > > I'm not sure who you think the "owner" of "filbert.example" > is. HTTP doesn't know anything about "owners". The AWWW defines the notion of URI owner: http://www.w3.org/TR/webarch/#uri-ownership > There are > protocols, domain names, operators, and an infrastructure > of DNS. The system doesn't allow delegation to a new > "specification". What the operator of the service at > filbert.example has as operational policy is, and should > be, opaque to the agent interpreting > "http://filbert.example/*" You seem to be ignoring the point of layered semantics. > > > > > There is one extra level of indirection in the http case, but the net > > effect is nearly identical. > > Only in some hypothetical world. > > > > For simplicity, but without loss of generality, I have made the syntax > > of these two URIs look very similar. > > You're building up an imaginary system which cannot be > put into practice. Huh? This isn't imaginary. It's the way semantic extensibility works *now*. > I could imagine a hypothetical world > in which you could make the associations you're providing, > but I don't think it's possible given the organizations > involved, nor do I think that the hypothetical world > you postulate is a better one than the one in which > http URIs are used for the HTTP protocol, and if you don't > want the HTTP protocol, you use a different URI scheme. You're missing the point. It's not about *not* wanting the HTTP protocol. An http URI still has its semantics as an http URI, which means that it is still legitimate for someone to try to dereference an http URI, and they might even get useful information back as a result. The additional Filbert semantics are in *addition* to the http semantics: they *further* license the use of that URI as a Filbert URI. That's what layered semantics is all about. This is no different in principle from the use of URIs for accessing weather forecasts. As owner of the weather.example domain I could decree that all URIs minted in the URI subspace of http://weather.example/zipcode/* are for accessing weather forecasts by zip code, such as http://weather.example/zipcode/02144 . I could publish this information to the world, and allow people to access weather forecasts for any zip code this way. Furthermore, if I chose to do so, I could define a new WEATHER protocol that is specifically used for accessing weather reports by zip code, and declare to the world that any URI in the URI subspace http://weather.example/zipcode/* may be accessed using the WEATHER protocol. TimBL has written about this layering in one of his "Design Issues" documents: http://www.w3.org/DesignIssues/Stack David Booth > > I could respond to the rest, but it's mainly repetitive. > > Larry > > > > -- David Booth, Ph.D. Cleveland Clinic (contractor) Opinions expressed herein are those of the author and do not necessarily reflect those of Cleveland Clinic.
Received on Thursday, 6 August 2009 03:07:14 UTC