: [Fwd: Re: Corrections to RFC-to-be 4004]

Hello folks,

Recently, we received notice that RFC 4004 was about
to be published, and entered the "48 hours" comment period.
I had not paid much attention to the document until that
point, but was asked to finally read through it carefully.
After reading it, I made the following suggestions (among
others which were considered even more editorial), and
then it was suggested that I send them to the mailing list
to see if there were any objections.

If not, then I think the corrections could be submitted
and the document published very soon.

Regards,
Charlie P.

PS. I can supply the instances of one or two sentences
       if the desired change is approved.

=================================================


            ...          I would like to see the
document changed in the following ways, but not at the
risk of delaying its progress.        .....

In section 4.1.1, it is stated that certain actions MUST be
taken if AAAH is stateful, and that certain other actions
occur otherwise.  There is no way to enforce the MUST.
Thus, it is pointless.  Non-conformance cannot be detected.
It does not affect interoperability.  Overall, I think the
mandate ought to be reduced in force to a SHOULD,
or conceivably a MAY.  In fact, it could potentially be
eliminated entirely.

At the top of page 23, something is said about destroying
"all" session keys.  It should be explicitly specified just WHEN
this action of deallocating all the MN's "resources" actually
occurs.  Clearly, it's not when the session keys expire!
(or... ???)

I do not understand what "*" means for certain AVPs in the
lists of required AVPs for the new AAA commands.
Surely it's part of the Diameter vernacular, and thus familiar
to one steeped in the lore.  But nevertheless I think for good
document cleanliness, the meaning of the "*" should be
repeated somewhere in this document.  It would only take
one sentence.

The name "Foreign-Home-Agent-Available" is clunky.
I would prefer "Visited-Home-Agent-Available".

The difference between two of the flags is not immediately
obvious, and in fact I had to skip around in the document
to figure it out (I think I get the point now).
           8   Foreign-Home-Agent-Available
           128 Home-Agent-In-Foreign-Network
It would be nice to have a sentence or two explaining
this, perhaps around the first time that the feature vector
itself is introduced.

In section 7.7, there is mention of a "RADIUS-style"
computation.  This needs a citation to the relevant
document specifying the computation.

         .........

Regards,
Charlie P.