| Isaac Gouy 2006-06-24, 7:02 pm |
| Vladimir Levin wrote:
> Isaac Gouy wrote:
-snip-
>
>
> Well, I think we would not have the degree of interaction that we
> currently have. Nor would we have the flexibility to make adjustments.
> For example, we recently were working on two allocation methods
> separately, but someone noticed that it should be possible to merge
> many aspects of those two methods into one. We have begun merging the
> code and the business people have adjusted the way they prepare
> requirements accordingly. It should speed development up a lot. To
> answer your question, I think we'd still have the subject matter
> experts, but would they be they working right next to the developers
> and discussing the stories being worked on every day? I assume not. I
> assume that they would get the job of preparing requirements and would
> leave (most of them anyway) when the developers arrived to begin the
> implementation phase. That seems like a very different dynamic.
>
> If you disagree, please describe the kind of interaction you would
> expect to see if things were decided way up front. Maybe we're simply
> using different terminology to talk about the same things.
You've brought in a new example which talks about coding, and that's
fine, but what I was really asking about was your original scenario
because it seemed to be about noticing something in the story (the
requirements).
We might assume that the subject matter experts would not be working
with the programmers. (I've always liked the term software developer
but I like it because it covers so many different roles.) Would it be
unreasonable to imagine yourself interacting and exchanging information
with the subject matter experts and an interaction designer, in just
the way you originally described, and then describe that role as a
requirements specialist?
>
> Fair enough. It's somewhat off topic - more of a general discussion -
> but I'd say that it's easier to make those opportunity cost decisions
> as you go along with software rather than up front. Again, to
> illustrate: Instead of X and Y, I ask for x1 and x2 to be devloped. I
> have a look at the result. Now it seems more clear that y1 should be
> next. Earlier on I may have believed that x3 would be the way to go.
That seems possible - it doesn't seem to necessarily follow, it depends
on the assumptions we choose, it depends on what other artifacts
(sketched, screen painted, prototyped) we imagine as part of the
process.
|