Re: draft-ietf-secsh-dh-group-exchange-01.txt

[Markus Friedl <markus%monkey.org@localhost>]

On Wed, Aug 22, 2001 at 04:46:37PM -0700, Darren Moffat wrote:
> >provos%citi.umich.edu@localhost (Niels Provos) writes:
> >
> >> In message <200108092254.f79MsLB718759%jurassic.eng.sun.com@localhost>, Darren Moffat 
> wri
> >> tes:
> >> >Matching against the peer's version number to try and work out which
> >> >version of a particlar draft they implemented doesn't seem like a good 
> idea.
> >> Thats who incompatibilities for all the other drafts have been handled, 
> too.
> >> Look at the code, you will be surprised.
> >
> >Yes, but for example in our software it is mostly done against our own
> >versions. And that is ugly enough. There are many vendors already, and
> >there is no guarantee that the application version string stays in the
> >same format from version to version.
> >
> >I don't know of a better way however; changing request names is ugly
> >too, perhaps even more so.
> 
> This is just one of the ugly things you have to do to deal with a
> protocol that hasn't been standardized.  If the protocol had become and
> RFC already and then stuff had changed the easiest way would just be
> to rev the protocol version number.

yes, but i don't think it's a big deal, since not many people
did implement the old version of the draft.

-m