A number of people are working towards the ability to search
selected sites, though I haven't heard of anyone trying to
put the result in a newletter format.
Harvest allows the user to custom build his own database,
which is then locally accessed at search time.
(http://harvest.cs.colorado.edu/)
MetaCrawler, Silk, IBMinfoMarket, and no doubt many others
query multiple pre-configured search databases at search
time.
(http://metacrawler.cs.washington.edu:8080/home.html
http://services.bunyip.com:8000/products/silk/silk.html
http://www.infomkt.ibm.com/about.htm)
I'm looking forward to the day when two of these
"meta" search services point to each other and create
an infinite search loop....
PF
ps.
If you're going to the WWW conference in Boston, I'll
be chairing a BOF on distributed searching. Please see
http://rodem.slab.ntt.jp:8080/paulStuff/