- From: Booth, David (HP Software - Boston) <dbooth@hp.com>
- Date: Mon, 14 Apr 2008 10:50:23 +0000
- To: Pat Hayes <phayes@ihmc.us>
- CC: "wangxiao@musc.edu" <wangxiao@musc.edu>, Tim Berners-Lee <timbl@w3.org>, Michaeljohn Clement <mj@mjclement.com>, "www-tag@w3.org WG" <www-tag@w3.org>
> From: Pat Hayes [mailto:phayes@ihmc.us] > > At 12:52 PM +0000 4/13/08, Booth, David (HP Software - Boston) wrote: > [ . . . ] > > Imagine that you have some information, I, that you > > wish to provide through your Web server. > > [ . . . ] What do you mean by 'information' here? [ . . . ] > Any input to a computational process has to be encoded > in bits somehow, because bits are what computations work on. Right. For the purpose of this discussion, it doesn't matter how I is expressed, provided that it can be viewed as a set of distinct pieces of information, so that we can talk about RI being a subset of I, and of course provided that it is the right type for then encode/decode functions. If you wish you could think of set I as being a set of RDF statements, for example. > [ . . . ] > > However, content negotion is *only* > > for sending information that is in I. Content negotiation is > > *not* for sending arbitrary information (or metadata) that is > > not already in I, such as the fact that the photograph was > > taken by "David Booth" and the cat depicted in the photograph > > is named "Cheshire". > > > BUt if this metadata were part of I, then this would be OK, > right? Correct. > So, here one is on the receiving end, and I get handed > a lossy JPEG image and some RDF metadata about it. What > should one do? Protest to the Gods of the Internet about this > mis-use of conneg, or simply assume that I - to which one has > no other access - did after all contain the metadata. The > fact that one has not seen it before can be chalked up to the > other kinds of lossyness inherent in the other encodings one > had previously chosen. Correct. Unless you have more information about what I is supposed to contain, you won't know. On the other hand, the owner *might* have told you, in some other document, what it was supposed to contain. > > The point Im getting at is that this strict doctrine which we > keep hearing about, regarding what conneg MUST be used for > and MUST NOT be used for, seems to be nothing but doctrine. > There is no way to determine that it is being obeyed. If we > all just go on assuming that it is, then some resources will > seem to have more information in them than they previously > seemed to have. As long as this extra information is useful, > why should anyone complain about this outcome? To hell with > doctrine when it ceases to be useful and becomes simply an > encumbrance. Hold it. I'm not quite following your point here. I thought the issue was that Xiaoshu was suggesting that it should be okay to use a URI to denote a non-awww:InformationResource (such as a sumo:Human) while using content negotiation to return a 200 response with either HTML or RDF. It's perfectly fine for content negotation to return either HTML or RDF. The only part that is *not* okay is when he tries to use that same URI to denote both a person and an awww:InformationResource. That causes URI collision: http://www.w3.org/TR/webarch/#URI-collision For example, suppose a GET on URI http://example/mycat sometimes returns a 200 status with the following HTML: <html> <body> This is a page about my cat. </body> </html> and, because of content negotiation, sometimes returns a 200 status with the following RDF (shown in N3): <http://example/mycat> a :MysteryType . If :MysteryType is awww:InformationResource, then there is no problem. But if :MysteryType is sumo:Human, and sumo:Human is disjoint with awww:InformationResource, then there is a URI collision, because in essense, the 200 response implicitly declared URI http://example/mycat as denoting an awww:InformationResource, whereas the RDF content declared URI http://example/mycat as denoting a sumo:Human. In such case the URI owner has done something wrong, but whether you consider the error to be a misuse of content negotiation or something else is a matter of interpretation: however you choose to atribute the cause, the parts don't fit together. David Booth, Ph.D. HP Software +1 617 629 8881 office | dbooth@hp.com http://www.hp.com/go/software Opinions expressed herein are those of the author and do not represent the official views of HP unless explicitly stated otherwise.
Received on Monday, 14 April 2008 10:51:42 UTC