- From: Brian E Carpenter <brian@hursley.ibm.com>
- Date: Wed, 10 Feb 1999 15:59:06 +0000
- To: spreitze@parc.xerox.com
- CC: Graham Klyne <GK@dial.pipex.com>, discuss@apps.ietf.org
My point is only that the word "subset" is open to the wrong
interpretation; if we just speak of layering there is no
possible misinterpretation.
Brian
spreitze@parc.xerox.com wrote:
>
> > ... that approach (true subset of HTTPng)
> > could imply that HTTPng is the basis for all future applications
> > protocols.
>
> It seems to me that the implication goes the other way (assuming we're willing to assemble a given application's protocol stack from various protocols that solve subsets of the application's problems): the smaller the set of protocol problems you address, the larger the set of applications whose problem set includes the set you address. Put another way: if HTTP-NG is layered over APPLCORE, then APPLCORE is used for everything HTTP-NG is and potentially also things that HTTP-NG is not used for.
Received on Wednesday, 10 February 1999 11:02:05 UTC