Re: Test first as specification

Laurent Bossavit wrote:

>> Yes. I suggest requirements (REQ) = 1-8, 11.

> One reason I've called them "candidate requirements" is that I suspect
> they should not all make it as written into the final specifications. In
> particular, I'd want to see 8 broken down into smaller pieces. And it
> seems to me that 1) belongs with 5) through 7) - there will be many
> tests for which the board dimensions don't matter.

What you call them, how you divide/refine them is moot.

Do we have agreement on 1-8, 11 being the start point for both of us ??

If so, we now need to agree the information sources for the subject
domain of the game of Go.


>> 3. We need an independent source to hold the deliverables until we
>> have both finished. I do not want to be unfairly advantaged by seeing
>> your stuff thus potentially bettering what I might otherwise devise.

> I'm not sure what's "unfair" about that. The point, as I see it, is to
> explore different approaches, not to hold a contest.

The point is to give an objective demonstration of work process.
If I am beneficially influenced by seeing your work, then it is not a
fair reflection on my work process.


Regards,
Steven Perryman

.

Relevant Pages

  • Re: Test first as specification
    ... >> domain of the game of Go. ... >> The point is to give an objective demonstration of work process. ... specification" against one that believes that a precise specification ... And in an environment where all things are equal (information sources ...
    (comp.object)
  • Re: Test first as specification
    ... we now need to agree the information sources for the subject ... > domain of the game of Go. ... As long as it's online... ... > The point is to give an objective demonstration of work process. ...
    (comp.object)