brian at karoshi.ucsd.edu
Sun Aug 30 23:33:34 PDT 1998
>...please note that as a formal matter
>interoperable implementations aren't required for a move to proposed standard.
No, of course not. I consider that a SERIOUS flaw in the current
process. One does, after all, want some proof that it's going to
work before we go standardising it - this isn't the ISO, and the
"make it work before you write the standard" was what made the
Internet such a big success in the first place. It's a great way
to find the flaws in the protocol design before they're fixed in
too much code. My experience has been that theoretical designs often
have to be adjusted when they encounter the real world.
However, that philosophical argument aside, I was hoping for
something to test clients against without having to write the
servers myself too. And it would be an invaluable chance to
see how practical people find each of the enhancements.
I'd strongly encourage the developers to PLEASE advise this list
when functional code is available to test against. I (and undoubtedly
others) anxiously await it.
More information about the ietf-nntp