> From: "Jason Wharton" <jwharton@...>
> Date: Fri, 16 Jun 2000 15:19:29 -0700

> > But unless and until we agree on the requirements for disaster
> > recovery, we can't evaluate possible alternatives.
>
> This would probably spin us into deadlock and nothing would ultimately get
> accomplished. You don't seem to be the agreeable type (nor am I). <g>

This has nothing to with whether Jason or Jim is an agreeing kind of a
guy or an arguing kind of guy: the starting point in software
engineering, just as much as in civil or mechanical engineering, needs
to be a statement of the requirements that the chosen design must
meet. To solve a problem, we first must know what the problem IS,
what requirements *any* chosen solution will have to meet in order to
actually *be* a solution to the problem.

I started earlier today to write a reply describing difficulties which
arise from Jason's proposal (like a transaction which is active for a
week, and prevents GCing of back versions), but I abandoned it when
Jim's message reminded me that I, just as much as Jason, was putting
the cart before the horse. Worrying about difficulties which arise
from a proposed solution (or the proposed solution itself) is
pointless until we have a clear statement of the problem we're trying
to solve.

First the problem, then the solution. First the question, then the
answer. Only in a Johnny Carson comedy routine does the answer come
before the question.

--
Chris Jewell developer/sysadmin voice: 831-431-6531
cjewell@... InterBase Software fax: 831-431-6510