ietf
[Top] [All Lists]

RE: RFC conformance:

2002-04-23 08:00:45
Liao Wei:

My recommendation for a way to show interoperability with your product is
first contact the area directors or the chair of the particular working
group in which you have RFC interest.  I have found everytime that they are
very helpful in establishing contacts with members whose companies may have
participated in the interoperability, and for me, they have had great
suggestions.  Then contact the individuals they have identified, and arrange
for either test time in their lab, or buy their equipement for your own test
lab, and start testing.

For an appropriate market application, I never had a problem checking in 3rd
party hardware or software into my test lab for interoperability testing and
verification.   My opinion is that we all want the standards to be
effective, applied, and work well.

My experience is that even with "certified" products, there will still be
interoperability problems.  Therefore, as you find issues like Valdis
mentions below, simply generate a FAQ page with the work-arounds.  Then guys
like him and me who go try to make this stuff work have a reference that we
can sift through to help us set those configuration flags for
interoperability.

Thanks, and good hunting.

Jim

Jim Busse
Chief Architect
NEC Solutions (America), Inc
email: jim(_dot_)busse(_at_)necsam(_dot_)com
Phone:  408-844-1234
Fax:  408-844-1250


-----Original Message-----
From: Valdis(_dot_)Kletnieks(_at_)vt(_dot_)edu 
[mailto:Valdis(_dot_)Kletnieks(_at_)vt(_dot_)edu]
Sent: Monday, April 22, 2002 7:40 AM
To: Jiwoong Lee
Cc: IETF
Subject: Re: RFC conformance:


On Mon, 22 Apr 2002 09:24:23 +0900, Jiwoong Lee said:
Someone may have to state one's implementation as

      - RFC 3180 compatible with errata version Mon, 22 Oct 2001

instead simply

      - RFC 3180 compatible.

Hm.. I think I see RFC numbering does not scale well. : )
Could you please share your thought on this errata series of RFCs ?

After having actually *checked* the page, and found mostly typographical
errata, and not many true show-stopper protocol errata, I don't think
that there is any real requirement to have to state "errata version"
when claiming RFC compatibility.

A *MUCH MUCH* bigger issue is when a product says "Supports RFCnnnn",
but only implements the MUST/MUST NOT level of requirements, and does
totally gratuitously stupid things with the "SHOULD/MAY" level items.
If you're looking for something for your engineers and tech writers to
do, rather than worry about the errata, have them make a list of all
those places where they do something other than the SHOULD(NOT), and
justify why they do so, and then if they are still bored, have them
do the same for the MAY items.

Of course, I'm just biased - I've had more operational headaches from
software that doesn't do a SHOULD right than I've ever had from
anything in those errata....

-- 
                                Valdis Kletnieks
                                Computer Systems Senior Engineer
                                Virginia Tech



<Prev in Thread] Current Thread [Next in Thread>