- From: Kendall Clark <kendall@monkeyfist.com>
- Date: Thu, 26 Feb 2004 13:23:42 -0600
- To: public-webarch-comments@w3.org
Through my work at UMD's Mindlab and XML.com (as well as trying to write a chapter on web architecture for the REST book I'm doing for O'Reilly), I've been spending a lot of time since December with the AWWW document. I've found a few issues that I want to present formally as Last Call comments. I've tried to summarize the issues below, and in many cases I've included links to more substantive discussions (in which, in some cases, there is some prose which is more suited to a generalist audience than to the TAG, for which my apologies). If some of these sound, well, crabby, I apologize; my intent has simply been to take the document as seriously as it was offered. I've tried to raise specific issues tied to concrete textual problems, rather than more general complaints about tone or intent. I'll be at the TAG public meeting on Tuesday in Cannes, and I'd be happy to elaborate on any of these if that would be helpful. ------------------ Last Call Comments ------------------ 1a. Fragment Identifier Semantics The new way of explaining fragment identifier semantics -- that is, talk of primary and secondary resources -- is, in my view, very weird. I amplify this claim at <http://monkeyfist.com/kendall/awww-issues/frag1.txt>. While I suspect that the older language for describing these semantics had its own problems, I would be happier either with (1) its return or (2) some further amplification or clarification of the existing language. 1b. Conflicting Secondary Resources I find this discussion in AWWW totally underdetermined. I discuss this at length at <http://monkeyfist.com/kendall/awww-issues/frag2.txt>. 2. What Kinds of URI Ambiguity Are There? AWWW abjures URI ambiguity; but in trying to think carefully about this, I've realized that it's important to distinguish two kinds of URI ambiguity: diachronic and synchronic. The AWWW only addresses the former kind, and I think it should address the latter kind, too. Diachronic URI ambiguity is the case where at time T, URI U identifies resource R; but at time T2, U identifies R2. Synchronic URI ambiguity can arise from "URI overloading" via content negotiation. Consider a URI U that identifies a magazine article resource, which is available in three different data formats, via con-neg: HTML 4, the XML variant of DocBook, and plain text. Consider, also, that via con-neg one can get an RDF representation of the article's metadata (say, the standard DC terms which apply to it). In this (common!) case, I suggest that U is synchronically ambiguous: which resource does it really identify, the article or the article's metadata? They aren't the same resource, though they are related; this synchronic URI ambiguity is as problematic, if not more so, than the diachronic type. I'd like to see some language in the AWWW about avoiding synchronic ambiguity by avoiding the "URI overloading" mistake with content negotiation. 3. Willy-Nilly Resource Change The AWWW says that one may conclude that agents or representations are each referring to the same resource if they are using identical URIs. But that's problematic; it suggests that the relation between resources and URIs is in some sense timeless and static. Once a URI has been coined to identify a given resource, it can only ever identify precisely that resource; else, we have to embrace the willy-nilly change problem. The willy-nilly change problem really bothers me; this is probably more pointed in the SemWeb 'social meaning' context; but even if so, there should be some indication in the AWWW that other people's URIs can change *radically*, ad-hoc'dly, and with neither forewarning nor warning. Yes, the AWWW implies this, but I'd be happier if it stated it outright. (I think what I'm calling "willy-nilly resource change" is an implication of the issues about ownership and authority raised by Peter Patel-Schneider. In some sense, then, my objection is parasitic upon his. If in responding to his objection the TAG substantively changes the AWWW -- which I doubt will happen, for what it's worth -- some such change may moot my objection here. If it does not substantively change the language about authority & ownership, then my objection is that it should warn people about the resulting willy-nilly resource change problem.) 4. Hypertext Issues 4a. Hypertext Good Practice Redundancies There seems to be a redundancy between two of the good practices in the discussion of hypertext. (Or, alternately, I simply fail to see the difference(s) which justify the distinction.) The first good practice says, in my paraphrase, that (1) good representation types allow users to make links to other resources and to parts of representation-states of resources. The second good practice says, again in my paraphrase, that (2) good representation types allow users to make "Web-wide" links rather than merely "internal document" links. Aren't these redundant? A link to "other resources" just is a link to something else on the Web or to a resource in some other information system which the Web encompasses. In other words, the reasonable reading of the first good practice is that data formats that allow links to other resources and to parts of representation-states are better than ones that don't, while the reasonable reading of the second good practice is that data formats should allow links to resources on the Web and not just to parts of representation-states. I fail to see the distinction which makes a difference here. If not precisely redundant, it seems that (2) is obviously and trivially entailed by (1), so we really only need (1). Of course, perhaps there's an aspect I'm missing completely; if so, could it be make more obvious? 4b. "Expected UI Paradigm"? The fourth good practice says good representation types allow users to make hypertext links when "hypertext is the expected user interface paradigm". Surely the AWWW also wants to say that for those kinds of web application or scenario -- Service Oriented Architecture and Semantic Web being the two obvious examples -- where hypertext is not the "expected user interface paradigm", by virtue of the fact that there really isn't a UI per se, one still wants to prefer representation types which allow users to make hypertext links between resources. REST and SOAP and RDF and WSDL and a lot of other fun stuff works precisely because -- even in the absence of any human-facing UI -- what's happening is that messages are being passed around between machines, some of which contain assertions about resources, and they are messages which contain hypertext links to other resources. The real problem here is that there is no real formalization of "hypertext link" in the AWWW. If it means A-HREF links simpliciter, then my point about SOA and Semantic Web exceptions to this practice is unmotivated and null. But if, as seems likely from Section 4.5.2. Links in XML, "hypertext links" encompasses any link mechanism (that is, XLink and friends) whereby HTTP URIs identify resources with which agents may interact with the resources-states thereof, then something like my point is needed. That is, in the weakest form, good representation types allow users to make hypertext links in many types of applications, but especially when hypertext is the expected UI paradigm. I would be happier with a stronger form of the claim: good representation types allow users to make hypertext links. Period. 5. Silent Error Recovery Always Harmful? I don't agree with the exceptionless form of this principle. I think one can imagine silent error recoveries which aren't harmful. I suggested an amended version: silent error recovery is harmful if, and only if, it does some harm beyond mere failure to notify; or, put better: mere failure to notify isn't always a harm. (I'd be just as happy with the smallest possible weakening of the principle, something like: "Silent recovery from error is usually [or "typically" or "often"] harmful." 6. Separating Presentation From Content This is often harder than the AWWW lets on, and sometimes it's simply not possible at all. I think the language should be modulated to reflect that reality. Further discussion, with concrete suggestions for change, is at <http://monkeyfist.com/kendall/awww-issues/prescon.txt>. 7. More Ambiguity "the ambiguous use of terms" is ambiguous; and, contrary to the AWWW's (fairly casual, of course) claim, ambiguity does *not* always impose a cost in human communications -- a research result demonstrated by UK cognitive scientists, among others. (If you want the full cite to this paper on CiteSeer, I can drum it up.) 8. Section 3.4.'s Unmotivated Paragraph There is a paragraph about URI ownership in Section 3.4, and I can't understand what it's doing there. I would strike or amend it. Full discussion of this issue is at <http://monkeyfist.com/kendall/awww-issues/para.txt>. 9. "Safe" and "Unsafe" Interactions I think I "know what you mean", but I really really hate the way this discussion is framed in the AWWW. I consider other ways of talking about this at <http://monkeyfist.com/kendall/awww-issues/safety.txt>. 10. Out-of-phase frag-id wordings in "Good Practice: Link Mechanisms" There are a few places in the AWWW where the language is out of phase with the new primary-secondary resource talk. This falls under wordsmithing and copy editing, but these out-of-phase bits really need to be fixed, since the issue is complicated and there have been at least two different ways of describing the semantics. 11. The "great power" of URIs and their "vastness of choice" I find this sentence, from Section 2. Identification, to be garbled at best. I discuss in some detail some of the issues it raises and implies in <http://monkeyfist.com/kendall/power.txt>. 12. Needless Propagation of URIs? I think this, as stated, is too strict. Further discussion at <http://monkeyfist.com/kendall/awww-issues/prop.txt>. Thanks to my colleagues at UMD's MIND Lab, especially Bijan Parsia and Jim Hendler, and O'Reilly's Edd Dumbill and Simon St.Laurent for discussions of, and reasons to discuss, these issues. Thanks, Kendall Clark
Received on Thursday, 26 February 2004 14:25:02 UTC