[
Lists Home |
Date Index |
Thread Index
]
Michael Champion wrote:
> "aggregation" of information that is unique to me is
> a bit of a misnomer, so I don't see the value of
> leveraging the RSS+Aggregator infrastructure to do
> that. Better to build a real pub-sub network?
I can see nothing about the RSS+Aggregator infrastructure that
results in any particular impedance mis-match with a "real pub-sub
network." The "RSS+Aggregator" stuff is, or can be, only part of a
fuller network of systems which includes real publish/subscribe
systems either upstream or downstream from one or more RSS files. The
complexity of the overall system needs to be adjusted based on the
specific need and the time-sensitivity of the data which flows through
the network.
For instance, as many know, at PubSub.com, we have a "real
pub-sub" system, but we can deliver results to an RSS file. This works
wonderfully for delivering notifications or content that aren't
time-sensitive. For most folk, blog and newsgroup posts aren't things
that must be seen within seconds of their posting. Thus, the whole
RSS+Aggregator+Polling thing works fine as a delivery mechanism. On
the other hand, we also support notification of SEC Edgar Filings and
there are whole communities of folk who insist on getting these
filings instantly. So, we provide email and other "fast" notification
methods (not all public) for SEC Edgar Filing notifications -- while
also providing RSS delivery for folk who aren't quite so insistent
about getting the notifications instantly.
The same is going to hold for "information that is unique" to
you. Some stuff you want to see immediately, other things you really
don't care about as much. What is needed here is a variety of delivery
mechanisms or perhaps a single mechanism with a variety of Quality of
Service (QOS) levels.
bob wyman
|
