Message-Id: <199602011601.KAA07560@library.wustl.edu> Date: Thu, 1 Feb 1996 09:45:20 -0600 From: Heath Rezabek <mailto:hrezabek@FIAT.GSLIS.UTEXAS.EDU> Subject: Re: Re[2]: yet another charter revision To: Multiple recipients of list WEBCAT-L <mailto:WEBCAT-L@WUVMD.WUSTL.EDU>
On Thu, 1 Feb 1996, Martin Hamilton wrote:> With URCs we ideally want something which is compact, unambiguous,
> and easy to parse.
-- And, I presume, consistently present from record to record / resource to resource. So a truly fundamental metadata set would have to be easily generated from any given resource. If the only resources in question were web pages, that would mean, perhaps, Title [from <TITLE>], its host server, and the URLs which that resource, in turn contains. I can't think of many other elements which are inherent; maybe a bit count and a last-modified date, which could have their uses. Is this impression accurate [my background is somewhat broad..]?
> The critical thing is that two programs which want to exchange URC
> information can do so via a common protocol.
Mm hm. But it seems like there *are* inherent elements; and establishing base-level records recording only those elements seems pheasable. [Labelling of URLs contained within a page & its peers is done in the DubCore via the designations "child" [contained in] "sibling" [same directory as [?] and theorhetically "parent" [accessed from?]. At this point I'm thinking out loud at best, so.
Heath M Rezabek mailto:hrezabek@fiat.gslis.utexas.edu http://www.gslis.utexas.edu/~hrezabek/index.html GSLIS Masters Candidate University of Texas at Austin Technical Staff Assistant Perry-Castaneda Library EIC