- From: Richard Light <richard@light.demon.co.uk>
- Date: Wed, 13 Apr 2011 16:26:34 +0100
- To: Ross Singer <ross.singer@talis.com>
- Cc: Karen Coyle <kcoyle@kcoyle.net>, Dan Brickley <danbri@danbri.org>, Ed Summers <ehs@pobox.com>, public-lld@w3.org
In message <BANLkTinY6HZ2Fy1niUauNb2Ws6F4KLBykQ@mail.gmail.com>, Ross Singer <ross.singer@talis.com> writes >On Wed, Apr 13, 2011 at 10:49 AM, Karen Coyle <kcoyle@kcoyle.net> >wrote: > I can see this becoming unwieldy, however. > > person_123_who_was_walking_along_Main_Street_on_Saturday_July_7_2011 > > It seems that there's a noun-verb struggle here. The person is the same > person, the activity is different. I know that some of these distinctions > are bound into our subject headings, but rather than change the identity > of the person I would prefer that we use the name in a context where > possible. > > George Bush --> served as POTUS --> dates > George Bush --> author of --> autobiog > > It's the same person, but the activity of the person has changed, not > his identity. > >But as a subject, these make a bit of a difference (and, realistically, is >where narrower terms come in) and you'd (potentially) want to be able to >distinguish between them. > >A book written about the executive branch of the United States in the first >decade of the 21st century probably should generate a significantly >different graph than book about the ownership of the Texas Rangers >(although there is, of course, one notable overlap). > >If the subject is significant (and distinct) enough, we should be able to >specify it. You're absolutely right, though, there is a limit to what's >practical. Isn't this just the classic pre- vs. post-coordinate indexing debate in another guise? In a Linked Data context, I suspect that you get much more utility (in the sense that you can answer a much wider range of queries from a single set of assertions) from analysing out in the way Karen suggests. Richard -- Richard Light
Received on Wednesday, 13 April 2011 15:27:04 UTC