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

Re: A question about requirements

At 10:21 -0700 9/22/03, Bob Hinden wrote:
Bernard,

They have whatever force the WG (and the IESG) wishes to give them.  Many
of the recent requirements documents might be better off being forgotten
entirely -- and some have been deemed so valuable as to not even be worth
finishing.

...

In practice there are elements of requirements documents which appear
foolish in hindsight -- and I would argue that they only need be minded
to the extent that they make sense once the WG understands the problem
better.


I agree that writing a requirement documents is a good way for the working
group to agree on what the problem is they are tasked to solve.  I think
there is starting to be a lot of experience that completing the
requirements document and publishing it as an Info RFC is not worth the
effort.  I am starting to think that most requirements documents should be
"working" documents and never be published as RFCs.

I have had the opposite experience in PKIX. in working on SCVP, we encountered debates about functionality that might have run on for some time, but the existence of a requirements RFC allowed us to quickly focus the debate.

I also think that developing requirements and a solution in parallel is
preferable to doing them sequentially.  It is too easy to write conflicting
or impossible requirements, where developing them in parallel allows for
feedback between the two and improves both.  It's also easier at times to
reach consensus on a solution (that generally meets the requirements) than
on the requirements document itself.  I think we tend to evaluate them
differetly.

I agree that parallel development can be attractive, but it does not help resolve the requirements debate sooner rather than later, a two-edged sword.