From: Larry Masinter (LM@att.com)
Date: Fri Mar 17 2000 - 10:08:34 PST
Me:
> > "Running code" is necessary but not sufficient.
> >
>
> False. Running code is necessary and sufficient. Proof statements:
> AOL, ICQ, Shoutcast, Napster. Running code, interoperating, from
> multiple implementors, is all of: desirable, necessary and sufficient.
Well, a couple of things. First, you mention a couple of things
beyond "running code", such as "interoperating" and "from multiple
implementors". And I'm not sure that the AOL situation is "desirable":
is it really reasonable that you only get to interoperate with AOL
if you sign a licensing agreement with them? So that, for example,
FaceTime can write an AOL client, but Microsoft and Yahoo and Mitre
can't? This is "desirable"? and "sufficient"?
Now, go into your AOL IM client, click on "setup", and then "Connection"
and then on "auto-configure". Now imagine running that (as I do) on
a laptop that is variously plugged into different networks with
different firewall and proxy configurations. Now, tell me that this
is "desirable" from an interoperability point of view.
If there were a standard for Instant Messaging and some analysis that
it wasn't an enormous security hole (as IRC seems to have been from
time to time), then the firewall administrators might actually turn
it on.
Convincing all of the firewall administrators and network managers
requires more than running code. Getting services actually deployed
widely requires convincing them.
Running code is necessary but not sufficient. It actually has to work
in a wider variety of circumstances than you can cook up in your
testing, and work stably, uniformly, and reliably. To get that requires
at least rough consensus that your application need is worthy, not
just the opinion of the area director.
> Convincing e.g. Keith Moore (sp? lazy. My former roommate? Dunno.)
> (pick your IETF AD) that your application need is "worthy" is neither
> necessary nor sufficient.
An odd strawman; you must be confused about the nature of the process.
Convincing the IETF AD that your application is "worthy" is neither
necessary nor sufficient, but you're playing the wrong game.
> $0.03 (inflation)
>
> jb
hmmmm
This archive was generated by hypermail 2b29 : Fri Mar 17 2000 - 10:09:48 PST