- From: Sandro Hawke <sandro@w3.org>
- Date: Thu, 29 Jul 2004 11:20:45 -0400
- To: Yosi Scharf <syosi@MIT.EDU>
- Cc: public-cwm-bugs@w3.org, connolly@w3.org
> Yes it should. But figuring out what the referer IS is a decidedly
> nontrivial task. In particular, the command
>
> cwm file1.n3 file2n3 --think
>
> can combine use a combination of rules from the two files to generate a
> URI and read it. Figuring out what, if any referer should be given in
> cases like this is not an easy task at all.
I disagree. Any URI U1 from which you obtained text which contained a
URI U2 is a referrer you could correctly pass along when dereferencing
U2. I suggest the first such U1 is probably the best and most-useful
one, with the idea that it's simplest to think of spidering being
depth-first.
It seems to me this could be easily implemented with a dictionary
mapping U2 to U1. Entries added by or near the parser, and looked at
by the dereferencer. It could be fun to dump and examine the
dictionary, too.
-- sandro
Received on Thursday, 29 July 2004 11:17:27 UTC