FW: HTTP-NG and application core

From: Larry Masinter (masinter@parc.xerox.com)
Date: Fri, Jan 29 1999


From: "Larry Masinter" <masinter@parc.xerox.com>
To: <ietf-http-ng@w3.org>
Date: Fri, 29 Jan 1999 15:25:30 PST
Message-ID: <001101be4bde$a96f0200$15d0000d@copper.parc.xerox.com>
Subject: FW: HTTP-NG and application core

-----Original Message-----
From: Chris Newman [mailto:Chris.Newman@innosoft.com] 
Sent: Friday, January 29, 1999 3:09 PM
To: Larry Masinter
Subject: Re: HTTP-NG and application core


On Fri, 29 Jan 1999, Larry Masinter wrote:
> What do you think of the possibility that HTTP-NG might narrow
> its scope enough to be a candidate for the application core?

I would welcome a proposal from the HTTP-NG team for a "core protocol" 
which meets the restrictions of this proposed charter.  I was also
thinking of asking the people doing DIAMETER in the AAA BOF to do a
proposal.  And Rob Earhart has already done an IMAP-based draft as well. 
If we take the best ideas from these and illuminate them with the
historical perspective of past successes and failures, I think we'll end
up with something great. 

> Certainly 'web browsing' is a very important application, and
> anything that is a general-purpose application protocol should
> support one of the most common Internet applications.

"Web browsing" is out of scope of APPLCORE since we only have one IETF
protocol for it (unless you count FTP as a web browsing protocol). Suppose
we layer the "mailcap" protocol on top of APPLCORE.  Should mailcap also
do web browsing?  No.  I'd conclude "web browsing" isn't part of the core.
Furthermore, most of the negative comments on this charter think it's
still too broad, so adding "web browsing" would be a step in the wrong
direction.

However, I'd have no problem with the HTTP-NG team independently layering
a web browsing application on APPLCORE in addition to being active
participants of the APPLCORE work.

> The HTTP-NG working group proposed charter wasn't accepted as
> proposed, and the intent seems to be to create two working groups:
> one in transport, to focus on the transport related issues, and
> one in APPS, to focus on the problem of creating an 'application
> core'. I thought you did a good job of writing down some important
> requirements for such an application core, but I believe that
> many of the HTTP-NG working group members believe they're actually
> working on the problem that you've articulated, despite the name
> of the group as "HTTP-NG". In fact, part of the pushback at the
> BOF in Orlando was "don't call it HTTP".

I've had enough positive response from this proposed charter that I'm
going to ask for a BOF in Minneapolis.  I'd welcome constructive input
from HTTP-NG people.  If someone from HTTP-NG was willing to help with
document editing of the historical/problem-statement draft, that would be
wonderful.

		- Chris