IETF-SSH archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Other comments on draft-ietf-secsh-publickey-subsystem





On Wednesday, August 30, 2006 08:22:18 AM +0200 Jon Bright <jon%siliconcircus.com@localhost> wrote:

Sam Hartman wrote:
"Jon" == Jon Bright <jon%siliconcircus.com@localhost> writes:
    >> - I'd rather the "mandatory" attribute of attributes be named
    >> "critical"...

    Jon> This would change a sentence like "If the server does not
    Jon> implement a mandatory attribute, it MUST fail the add.." to
    Jon> "If the server does not implement a critical attribute, it
    Jon> MUST fail the add..".  The first seems preferable to me.

My personal opinion is that critical is far preferable to mandatory in
a security protocol.  The usage you seem to be objecting to is quite
common in PKIX documents and is becoming more common in Kerberos
documents and other things throughout the security area.

I didn't make the common as an AD because I thought it a bit late, but
I support this change as an individual.

OK, since everyone seems to want this, I'll change the wording to
"critical".  I'm still confused about why this is an improvement, though.

In normal English usage, "critical" has several meanings.  One of these
is "indispensible, essential" (and having looked at several dictionaries,
that's usually the meaning right before the definitions involving nuclear
physics begin).  "Mandatory" seems to have only the meaning wanted in
this document.

The distinction here is subtle but important.
The word "mandatory" usually describes things that MUST be implemented.
The word "critical", in a security context, is used to describe things that need not be implemented, but which have the effect of producing incorrect behavior if blindly ignored. Such features/extensions/whatever must be rejected if they are seen by an implementation which doesn't understand them.

-- Jeff



Home | Main Index | Thread Index | Old Index