Is It Push? (11:15:04 PM) 
Over in James' blog, he says "It's push technology that actually succeeded because it wasn't being forced on users by any single technology vendor", and then (to Tim Appnel) "You're right, it ain't push, it's pull." Don't some of you really wish it was push, or at least better pull?
I am the first to admit that I am in no way an RSS guru, but the thought of all the wasted bandwidth pulling over an entire feed's worth of content on every scan when 95% of it is likely to be unchanged from the last time you scanned has been gently but insistently tweaking the back of my mind since I started reading blogs. I know there have been conversations about this plenty of times, but has anyone yet suggested some kind of standard HTTP header or query string parameter which would let you specify the last time you succesfully grabbed the feed? Then dynamic engines which understood it could only hand you the new bits... Yes, I know that the standard HTTP last-modified trick will get you a good bit of the way there, but for big feeds it still seems like it would be nice to be able to get just the new part. Static feeds could simply ignore such a thing, and dynamic ones could DTRT.
So I guess maybe what I'm actually asking is - has anyone implemented a weblog which parses If-Modifed-Since and only sends back the new RSS?
Conversations like this and this happened a while back - anything new on this front?
Edit : OK, I may have answered my own question by fully reading some of the conversation I pointed to. The problem with doing things as I'm thinking is that the backend needs to potentially regen the RSS feed on each request, which may be more expensive in terms of speed/processing than getting the whole feed would be in terms of bandwidth. Hm. (incidentally that was a really good conversation on Phil's blog there)
|
