Re: Robot Discovery

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

> Or, you can put
> references in the 'root' file to all of the user files, but then you
> still need to maintain all of those links, which kind of gets us 
back
> to square one; you might as well put the metadata in the Web pages
> themselves.

How about using a programmatic interface like SOAP for this?  Pass it 
an e-mail address and see if it returns a feed URI.

Or better yet, use something like DNS resource records to find the 
feed server for that domain.  You can use an MX record for the proper 
mail server, why not one for feeds too?  If you look into MS 
Hailstorm you'll see some stuff along these lines.

Start thinking about live queries and returning XML data.  It's a lot 
more flexible.

-Bill Kearney