To:
Brad Knowles <brad.knowles@skynet.be>
cc:
Ed Sawicki <ed@alcpress.com>, Randy Bush <randy@psg.com>, dnsop@cafax.se, namedroppers@ops.ietf.org, ngtrans@sunroof.eng.sun.com, ipng@sunroof.eng.sun.com
From:
Robert Elz <kre@munnari.OZ.AU>
Date:
Wed, 17 Jul 2002 09:13:29 +0900
In-Reply-To:
<a05111b0eb95a3d13cbc1@[10.9.8.228]>
Sender:
owner-dnsop@cafax.se
Subject:
Re: (ngtrans) Re: RFC 1886 Interop Tests & Results
Date: Tue, 16 Jul 2002 23:17:57 +0200
From: Brad Knowles <brad.knowles@skynet.be>
Message-ID: <a05111b0eb95a3d13cbc1@[10.9.8.228]>
| At 12:23 PM -0700 2002/07/16, Ed Sawicki wrote:
|
| > Yes, but isn't there value in knowing who the implementors are so
Sometimes a list of who participated is published, without
identifying which implementation belongs to which implementor.
| Moreover, if you are someone looking to actually use products
| such as the ones under test,
This is not the purpose of the tests - that's a worthy purpose, but it
isn't the IETF's role. What is being tested, as Randy Bush keeps on
saying, by the IETF is the standard upon which the implementations are
based.
We want to know if the standard describes something that can be implemented
in an interoperable way. When a failure occurs, then we need to know
whether of not that is because of a faulty standard, or just an implementation
bug ("the standard was clear, I just screwed up..."). Apart from that,
whether bugs exist or not is irrelevant - and the tests are not designed
to attempt to discover implementation bugs - so that bugs weren't found in
some implementation or other doesn't mean that there weren't any.
kre