Re: xp and simple design
From: Robert C. Martin (unclebob_at_objectmentor.com)
Date: 04/11/04
- Next message: Robert C. Martin: "Re: Class/Object Diagram in UML"
- Previous message: EventHelix.com: "Re: Class/Object Diagram in UML"
- In reply to: Daniel Parker: "Re: xp and simple design"
- Next in thread: Daniel Parker: "Re: xp and simple design"
- Reply: Daniel Parker: "Re: xp and simple design"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Sun, 11 Apr 2004 07:11:41 -0500
On Sat, 10 Apr 2004 23:31:13 -0400, "Daniel Parker"
<danielaparker@spam?nothanks.windupbird.com> wrote:
>"Robert C. Martin" <unclebob@objectmentor.com> wrote in message
>news:7qdh70pp90h1obdjdlhus19cg6ksgo29nk@4ax.com...
>>
>> The range of systems that XP-like methods have been used on is quite
>> broad. Insurance, banking, internet security, word processing,
>> project management, call processing, embedded real time, medical
>> systems, etc, etc.
>>
>The issue here is much narrower than XP proper, it is strictly concerned
>with the requirements side. The model of "customer in room" + 3x5 + tests
>is just too simple a model; it just doesn't fit with many types of
>organizations, organizations that would benefit from iterative development
>in other ways. I think the iterative community would widen its appeal more
>if they thought about the business and requirements side more. I would very
>much doubt, for example, that the insurance and banking applications
>referred to above were developed strictly within this model.
Some XP teams use narratives as part of their requirements process.
i.e. when a story is scheduled, they write up a technical description
of it along with the tests. Indeed, one of the central ideas behind
FitNesse is to allow the tests and narratives to be a single
collaborative and online document.
I know of one company that writes up highly formal detailed and
complex Marketing Requirements Documents for each release, however
they write it *after* the development is complete. (I'm trying to
talk them out of writing it, they don't really need something that
formal and detailed, they're just scared that Training, Support, and
Pre-Sales will freak if they don't get it. (They won't).
To reiterate your argument, you feel that the XP model of requirements
is too simple for most complex business domains. Please consider that
*all* requirements, no matter what process is used to create them,
begin as gleams in someone's eyes. From that point they grow and
evolve throughout the lifecycle of a project. They start simple,
informal, and undocumented and over time grow into formal, complex,
documented requirements.
XP does not change this fundamental fact. It simply changes when the
developers get involved, and the form of some of the documentation.
The developers get involved earlier, and become partners in the task
of formalizing and elaborating the requirements. They help the
business analysts and customers capture the details of the
requirements in terms of highly formal and executable specifications.
The code that meets these specifications evolves concurrently with the
specifications themselves.
XP does not eliminate the detail. Indeed you *can't* eliminate the
details. XP simply changes when developers get involved and the form
in which the details are captured. The end result is even more formal
and detailed than a non-executable requirements document.
-----
Robert C. Martin (Uncle Bob)
Object Mentor Inc.
unclebob @ objectmentor . com
800-338-6716
"Distinguishing between the author
and the writing is the essence of civilized debate."
-- Daniel Parker
- Next message: Robert C. Martin: "Re: Class/Object Diagram in UML"
- Previous message: EventHelix.com: "Re: Class/Object Diagram in UML"
- In reply to: Daniel Parker: "Re: xp and simple design"
- Next in thread: Daniel Parker: "Re: xp and simple design"
- Reply: Daniel Parker: "Re: xp and simple design"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|