|
It would clearly be useful to track who has read what.
Users would benefit by being able to find out just
the things they have not seen. The system could also
be smart about finding similar things of interest to
the user. Advertisers would like it, etc.
But tracking each and every response read by every reader
would be too expensive for the server to do. There
are about a 1000 members of this HyperNews server, and
I don't consider it really widely used yet (though the
server is starting to melt down recently). Furthermore,
there are probably 10 times that number of non-member
readers - many
of whom might want to take advantage of a tracking service.
No way this could be a global service.
Tracking ought to be done closer to the client.
It will
be reasonable to track readers when replicas of
HyperNews exist and a
single server only serves a moderately small community.
If you wanted to limit your HyperNews server to a small
community, then it makes sense to track readers. And
a HyperNews replica will be exactly a HyperNews server
limited to a small community. So it will happen.
| 
Tracking would be expensive 
Does HyperNews track read/unread articles?
(Steve Herman)
