Antone Roundy wrote:
> do you have figures on how much push reduced bandwidth
> use vs. RFC3229+feed? My guess would be that the difference wouldn't
> be as dramatic as between feed and RFC3229+feed, but I recognize that
> I may be wrong.
Sorry, we don't have precise numbers. I've done the calculations in
the past (it's a fairly expensive process) but then haven't kept the
detail... I did the calculations for me, not for review or publication. But,
you are absolutely correct in your guess that the savings of push over
RFC3229+feed is less dramatic then the savings over the normal file polling.
This was one of the motivations for our proposing RFC3229+feed in the first
place. We wanted to make polling as efficient as possible since we get
polled a lot and since we poll a lot... RFC3229+feed diminishes the
bandwidth savings you get from push but doesn't eliminate those savings.
That weakens the argument for push, but results in improving the overall
system.
> But I also don't think that we should require all Atom Documents
> (other than Entry Documents) to do so. In other words, I don't think
> this is an either-or question--I think both should be possible.
I agree that this should not be "either-or". I would be happy if the
spec said something like "multiple entries with the same atom:id MAY appear
in a single feed document." I don't think it would be reasonable to insist
that people insert all changes in their feeds. My goal is simply to permit
this, not force it. Whether the feed contains all versions of entries should
be up to the feed writer.
bob wyman
