RE: [Asrg] 6. Proposals - Pull System (revisited)

["Hallam-Baker, Phillip" <pbaker@verisign.com>]

> Yes, but that HTTP trick has a lot of built-in failure modes, which 
> is really why RSS came about.

Absolutely, the HTTP infrastructure is designed to support caching. It is
not a change notification system and was never intended to be. At this point
in the Web I am not sure that anyone really uses caching in the network. We
are not working at 28.8K anymore...

The HTTP protocol is intended to support use as a gateway protocol to a data
source. There is no way that the HTTP server can require change notification
from that data source.

> >For the respectable end of the market (the confirmed opt-in, or
> >whatever the current phrase is), I agree completely---there's a space
> >for some use of RSS or something much more appropriate than email.
> >Actually, I suspect HTTP with perhaps some automated browser
> >configuration (to switch on monitoring of the relevant web 
> page) would
> >be easier to get going.
> 
> That's a technique with 8 years of hard failure modes behind it. It 
> would be nice if there was a reliable lightweight standard way for 
> web browsers to detect significant changes of HTTP-accessible 
> documents, but there isn't. There's about an 80% solution, and the 
> ways that it does not work are non-trivial to solve.

Hence RSS, it is a fairly widely deployed solution at this point. I do not
see the reasoning behind creating a new protocol that has the same
structure. I do not see the value of research that is intended to duplicate
what is deployed.

It would be quite easy to configure a client so that it allowed automatic
subscription to RSS feeds. Probably the thing that is missing is a URL to
identify the protocol so that the browser knows what to do with the link
when you click on it. 

		Phill

_______________________________________________
Asrg mailing list
Asrg@ietf.org
https://www1.ietf.org/mailman/listinfo/asrg