- From: John Erickson <olyerickson@gmail.com>
- Date: Thu, 12 Aug 2010 07:35:28 -0400
- To: public-lod <public-lod@w3.org>
I realize this is a dangerous question, but...what is the cause of the change? * Has the "content" actually changed (revision), or simply moved (URI change)? * Must the content at the "old" URI go away, or can it be represented as a previous version? All of this is useful knowledge --- a form of provenance --- and should not be lost. The question is, how to handle it systematically and efficiently... This brings to mind a number of conversations I've had in the DOI/Handle System alternate reality, in which we've discussed a vocabulary of types (think: predicates) that could serve as hints to a HS proxy on what "physical" URI to return to a resolution request, including and especially as a "smart" way to handle conneg. Not only could these type assertions help the proxy distinguish between data vs. document, but it could nicely handle version ala Memento <http://mementoweb.org/> My point is not to harp on the Handle System, instead to point out that the HS (architecturally) builds-in the kind of encapsulation that can facilitate this --- although the type-based conneg I'm referring to has only been seriously considered in the past few months (AFAIK). John On Thu, Aug 12, 2010 at 5:12 AM, Kjetil Kjernsmo <kjetil@kjernsmo.net> wrote: > Hi all! > > Cool URIs don't change, but cool content does, so the problem surfaces that I > need to permanently redirect now and then. I discussed this problem in a > meetup yesterday, and it turns out that people have found dbpedia problematic > to use because it is too much of a moving target, when a URI changes because > the underlying concepts change, there's a need for more 301s. > > The problem is then that I need to record the relation between the old and the > new URI somehow. As of now, it seems that the easiest way to do this would be > to do something like: > > <http://example.org/old> ex:permanently_moved_to <http://example.org/new> > > and if the former is dereferenced, the server will 301 redirect to the latter. > Has anyone done something like that, or have other useful experiences relevant > to this problem? > > Cheers, > > Kjetil > > -- John S. Erickson, Ph.D. http://bitwacker.wordpress.com olyerickson@gmail.com Twitter: @olyerickson Skype: @olyerickson
Received on Thursday, 12 August 2010 11:36:01 UTC