Thanks for the response. Comments inline. I removed sections for which I have
no further comment.
Thanks!
Ben.
On May 16, 2013, at 10:19 PM, "Wang,Weiming"
<wmwang2001(_at_)hotmail(_dot_)com> wrote:
[...]
-- The draft mentions a couple of instances of tests that failed because of
an incorrect implementation or differing encapsulation formats. Does this
suggest that the specifications should be clarified? In particular, in the
case of encapsulation format mismatch, should the specs include stronger
requirements to be able to receive all encapsulation formats? Or should the
number of options be reduced?
[Re. by ΕΗ] The protocol provides a number of different approaches
[Re. by Weiming] The key issue is still from the deep understanding of the
protocol from implementations. I still have not seen need for any urgent
change for the protocol.
I don't have enough knowledge of the protocol to form a specific opinion, but
it's been my experience in other areas that when implementors interpret things
differently, there's often room for clarification, even if the text is formally
correct. I agree it doesn't imply an urgent need, but would it be worth one or
more errata?
[...]
-- section 4.4, last paragraph:
The text says that since the mentioned failures were likely the result of
bugs, it doesn't indicate an interoperability problem in the specs. I have to
point out that, it also doesn't prove interoperability in both directions for
the particular test. It would also be worth commenting on whether the
probably bugs were programming errors rather than misunderstandings of the
specification.
[Re. by Weiming] to change the whole paragraph to:
<t> The two test items failed. Note that Test #7 and #8 were identical to
the tests, only with CE and FE implementers were exchanged. Moreover, test
#12 and #13 showed that the redirect channel worked well. Therefore, it can
be reasonably inferred that the problem caused the failure was from the
implementations, rather than from the ForCES protocol itself or from
misunderstanding of implementations on the protocol specification. Although
the failure made the OSPF interoperability test incomplete, it did not show
interoperability problem. More test work is needed to verify the OSPF
interoperability.</t>
Works for me, thanks!
[...]
[Re. by Weiming] The section 3.2 para.3 has been changed to:
<t>... Because there came unfortunately a problem with the test system in
Greece to deploy IPSec over TML during the test process, this test only took
place between test systems in China and Japan.</t>
The sentence is still hard to parse. Do you mean the following?
"Because an unfortunate problem with the test system in Greece prevented the
deployment of IPSec over TML, this test only took place between the test
systems in China and Japan."
[...]