- From: <noah_mendelsohn@us.ibm.com>
- Date: Tue, 4 Apr 2006 16:50:50 -0400
- To: "Roy T. Fielding" <fielding@gbiv.com>
- Cc: W3C TAG <www-tag@w3.org>
Roy Fielding writes: > The finding should be about non-identifying data included in the > URI, which is almost always stuff that is metadata about actions on > that resource. Why? Using an example from my note: http://example.org/book/chapter2 I think there is indeed implied metadata about the chapter. First of all, it looks like I've inferred it was indeed a chapter. More to the point, I inferred (however unreliably) that it might be one of a collection of chapters, all in the same book. I further guessed that it might have a sibling: http://example.org/book/chapter3 Seeing either URI on the side of a bus, one might (unreliably) infer not only that example.org can be contacted for authoritative representations of the resources separately (which the RFCs directly support), but also (unreliably) that there may be a containing resource which is the book as a whole, and that example.org may have some real world ownership or control of it (I.e. they might employ the author as well as serving the Web resources). All of that seems to be inferred metadata about chapter2. If the URI were something more opaque, such as a uuid:xxxxxx, there is no way I could have inferred such things. Why is it inappropriate to consider such information as metadata and to talk about it in the finding? Insofar as the inferred information proves true, it is definitely metadata about the chapter I think. I do think we want to preserve and emphasize lots of the existing warnings about the downsides of either inferring or relying on metadata from the URI when you don't need to. On the other hand, on the recent TAG call, there was considerable sentiment that people do the above chapter-ish munging all the time, and that it's a significant aspect of what makes the Web useful. I believe I've been tasked with drafting a finding that tells both sides of the story. We'll see if I can do it without the result being too long or confusing. -------------------------------------- Noah Mendelsohn IBM Corporation One Rogers Street Cambridge, MA 02142 1-617-693-4036 --------------------------------------
Received on Tuesday, 4 April 2006 20:51:02 UTC