On Sun, Nov 15, 2009 at 10:18 PM, Alexander Johannesen
<alexander.johannesen_at_gmail.com> wrote:
> If you *really* wanted to be control freaks, you could use the
> User-Agent header [bad idea] (say that only browsers could slip
> through unattended, or registered agents) or agents with a (possibly?)
> registered FROM email address could have access [another bad idea].
> The convention is to simply track the FROM email addresses; you can do
> interesting correlations based on it. Both are easy enough to hack
> around, of course, but for LCSH, who would bother? :). All of this is
> a poor-mans version of control, though, as HTTP is quite transparent
> and easy to spoof.
>
> For serious control, jump through the authorization hoop, but since I
> suspect no one really wants that I guess we need to sort out what sort
> of data from users you're after. (I could at this point jump into my
> Identity Management spiel [for potential users of a resource], but
> I'll be good ...)
Perhaps I wasn't very clear. It's not control I'm after ... but usage
metrics. People are speculating about id.loc.gov not being used very
much, but really it's just some people on a discussion list talking
about what they *think* is going on. The Linked Data folks have their
high level view of what they think is going on too [1] ... but ideally
there would be a way for Linked Data publishers to look in their web
server logs and gather some sort of idea of how their stuff is being
referenced in other people's data.
/Ed
[1] http://richard.cyganiak.de/2007/10/lod/
Received on Sun Nov 15 2009 - 22:55:52 EST