Re: Problems with draft

[nisse%lysator.liu.se@localhost (Niels Möller)]

der Mouse <mouse%Rodents.Montreal.QC.CA@localhost> writes:

> >> Also, section 7 writes
> >>    The hash H is computed as the HASH hash of [...]
> >>      string    I_C, the payload of the client's SSH_MSG_KEXINIT
> >>      string    I_S, the payload of the server's SSH_MSG_KEXINIT
> >> It is not clear whether these are compressed or uncompressed;
> > It refers to the payload after the transport layer transformations,
> > i.e. after decryption and decompression, with padding removed.
> > Again, I don't think a change in the core drafts is necessary.
> 
> Perhaps.  I think I disagree, though; I don't see how someone
> implementing from scratch is supposed to know, from the drafts, that
> I_C and I_S are the decompressed forms.

Except for section 4, which describes the transport-level mangling of
compression, padding and encryption, all mention of "payload" in the
specs refer to the uncompressed, unencrypted, unpadded payloads. It
may be a little unfortunate that both section 4 and section 6 of the
transport spec use the word "payload" for different things, but I
don't think it's a big problem.

> I still think less ambiguity on each point would be good.  This is,
> after all, supposed to be a standard, so clarity is a Good Thing.

And conciseness is also important; if every detail that someone might
ever misunderstand should be explained full, a spec can easily get too
large to be readable.

> Is there some high overhead associated with changing the drafts or
> something?

We've been trying to get this stuff to be published as a proper
standard for ages, and every little change somebody asks for delays it
even further. 

That said, it's of course a good thing to get feedback on the draft.
If you really think this needs additional clarification, please write
the text you want and the wg should consider it. I'd recommend
rewording section 4 so that it doesn't use the word "payload" to refer
to a string that sometimes contain the payload in compressed form.

Regards,
/Niels