Re: Robot Discovery

["Bill Kearney" <wkearney99@hotmail.com>]

> One thing to consider is large/deep sites, such as multi-user sites 
> using URLs of the form www.site.com/user. If there are thousands of 
> users, a global discovery file could be huge. Perhaps what's needed 
> is one per user, such as www.site.com/user/discovery.xml.

This is why I look at it from the perspective of the currently viewed 
page.  If you're looking at something that might have a feed, meta-
data would help find it.  A site full of user sites (or other 
hierarchies) wouldn't be as viable.  However, if you're look at a 
site directory of feeds, well that might be interesting.

-Bill Kearney