what are the real problems

Thomas Narten narten at us.ibm.com
Fri May 23 09:52:33 CEST 2003


Aaron Falk <falk at isi.edu> writes:

> You've been around the IETF much longer than I.  Do you sense that the
> level of conservatism in the IESG has increased over time?  My
> perception is that, as more non-computer scientists get on the
> Internet, it has been viewed as more of a utility, i.e., "it just
> works."  It wouldn't surprise me if there is a reduced tolerance of
> risk by the IESG to protect these users.  In other words, the
> interpretation of stewardship of the Internet has changed from "let's
> grow something neat" (or "let 1000 flowers bloom") to "let's keep this
> cool thing from breaking" (or "let's not confuse the marketplace").

It probably has. Speaking personally, the longer I've participated in
the IETF, the more I can look at past decisions and conclude "we
shouldn't have done that", or "if only we had pushed back a bit more".

Ten years ago, the Internet wasn't being viewed as something that had
to be 7x24 and part of the basic infrastructure of our economy and
society.  But times have changed. Protocols that might easily be
considered "90% finished" (as in "Good Enough" for PS) ten years ago
don't get the same deference today. In many cases, one just can't
assume that we can fix the serious bugs after we get experience. In
today's world, where a PS can be implemented/shipped via a Windows OS
release or cell phones to tens-hundreds of millions of devices, stuff
that gets deployed has to be dealt with as part of the infrastructure
for years afterwards. Security is the obvious example where if we
screw up in what gets deployed, the potential for damage is huge.

This goes back to a thread earlier (Brian?) about how the industry has
matured and expectations have changed as well.

Thomas


More information about the Problem-statement mailing list