I have been testing a new version of the Swish-E spider that can handle 
HTML documents with frames.

If the spider program detects that the document is a framed HTML, it 
recursively builds content by reading through the <frame src> pointers, and 
builds up a list of all the <a href> links that it finds. It then passes 
*this* back to the C program, which indexes the document as if it were one 
big HTML. The spidering will work as if all the <a href> links found in in 
the frameset HTML files were at level 1.

Would this be of use to anyone?

Chris Humphries