what are the requirements? [Re: XP Boundaries] 
Author Message
 what are the requirements? [Re: XP Boundaries]

Victor Goldberg wrote (much abridged):

Quote:
> So, at its core, XP assumes the existence of a client and a close
> collaboration with such a client.  XP development is client driven...
> When the client is an abstract entity, like The Market, the thoughts and
> motivation behind XP may still be valid, but should be thoroughly
> re-examined...

Thanks for an interesting article. I'd like to explore the concept of
'close collaboration with such a client' from a cynic's viewpoint. You
may take it as weekend humor if you like with a twinge of tragedy.

One would expect a human client's needs to be easier to discern than
those of the Abstract Market. In my experience on many projects large
and small, discovering the client's needs is often hindered by these
factors:

- the person assigned to be 'the client' has no interest in his duties
- if he is interested, he does not know what he wants
- when he does know, he finds it hard to articulate
- what he does articulate is imprecise and vague
- as it is fleshed out, he realizes it's not what he wanted after all
- he neglects to communicate critical details that later wreak havoc
- what he wants does not adequately represent his organization
- his organization was chosen by artificial criteria to state needs
- the organization does not adequately represent the enterprise
- the enterprise changes direction after the project is initiated
- his authority is later withdrawn and his requirements nullified
- in midstream a new 'client' replaces him and the cycle repeats

I can recall working on one major project that suffered none of these
defects, but almost all my projects have suffered at least some. My
experience assessing the needs of the Abstract Market is worse than
for the human client.

I think that needs assessment, or requirements analysis if you prefer,
is the weakest link in software methodology. Making it more robust
would bear much fruit. It would require major surgery on the business,
even at the highest levels. Does anyone know what happened in the end
to the classic XP project--Chrysler Payroll? Is it humming
away happily in production, or did it meet some organizational demise?



Fri, 24 Dec 2004 07:44:20 GMT  
 what are the requirements? [Re: XP Boundaries]
Hi Bruce:

<much good stuff snipped>

Quote:

> I think that needs assessment, or requirements analysis if you
> prefer, is the weakest link in software methodology. Making it more
> robust would bear much fruit. It would require major surgery on the
> business, even at the highest levels. Does anyone know what happened
> in the end to the classic XP project--Chrysler Payroll? Is it
> humming away happily in production, or did it meet some
> organizational demise?

Second question first -- AFAIK, C3 died.  I haven't checked, but I bet
the story is on the c2.com wiki.  What I *heard* was that it failed
when its users were stopped by the system from entering bogus info.  
The system it was to replace, allowed this and the business users
decided they wanted that capability.  I must assume that entering
erroneous data was *not* part of the stated requirements.

Interestingly, when Kent was here in Seattle last time, he spoke to a
combined Smalltalk, Java and XP users group meeting.  He talked
specifically about business organizational change.  He felt (and I am
paraphrasing here) that it was MUCH more difficult to get business to
change and that business organizational change was where he should
focus his energy.  Getting developers to be successful with XP was
easy by comparison.

--
Thanks!!
Joseph Bacanskas [|]
--- I use Smalltalk.  My amp goes to eleven.



Fri, 24 Dec 2004 12:09:49 GMT  
 what are the requirements? [Re: XP Boundaries]
Interesting.  One of our marketing folks just read Kent's XP book, and
his statement was:

"This is more than a programming methodology; it's a business
methodology"

On Mon, 08 Jul 2002 04:09:49 GMT, Joseph Bacanskas

Quote:

>Hi Bruce:


><much good stuff snipped>

>> I think that needs assessment, or requirements analysis if you
>> prefer, is the weakest link in software methodology. Making it more
>> robust would bear much fruit. It would require major surgery on the
>> business, even at the highest levels. Does anyone know what happened
>> in the end to the classic XP project--Chrysler Payroll? Is it
>> humming away happily in production, or did it meet some
>> organizational demise?

>Second question first -- AFAIK, C3 died.  I haven't checked, but I bet
>the story is on the c2.com wiki.  What I *heard* was that it failed
>when its users were stopped by the system from entering bogus info.  
>The system it was to replace, allowed this and the business users
>decided they wanted that capability.  I must assume that entering
>erroneous data was *not* part of the stated requirements.

>Interestingly, when Kent was here in Seattle last time, he spoke to a
>combined Smalltalk, Java and XP users group meeting.  He talked
>specifically about business organizational change.  He felt (and I am
>paraphrasing here) that it was MUCH more difficult to get business to
>change and that business organizational change was where he should
>focus his energy.  Getting developers to be successful with XP was
>easy by comparison.



Fri, 24 Dec 2004 12:57:51 GMT  
 what are the requirements? [Re: XP Boundaries]
Victor Goldberg wrote these thoughts on which I wish to comment:

Quote:
> Communication is the best preventative tool available...

In the context of gathering software requirements, I'll relate two
stories, one of communication that failed and the other that succeeded.

In the failure case, I implemented a few pieces of one of the world's
most ambitious Smalltalk projects. The requirements for one piece were
communicated and negotiated on the web, in person, and by email. They
were discussed in excuciating detail. A high level committee
representing organizations and geographies endorsed them. But six
months into the implementation, some engineers were consulted who shot
the decisions down. Six months of my work was lost.

Moral: communicate early with the right people.

In the successful case, I implemented pieces of a medium scale fortran
and Assembly project. The requirements were again negotiated in great
detail and documented. They stuck, served as the basis of a superbly
modular design, and drove the best project to which I've contributed.
The communications hub was a manager who had deep business experience,
  amazing design skills, brilliant coding ability, and humility.

Moral: choose the chief communicator carefully.

Quote:
> Not only that, but 10% of he population is above the 90th percentile...

I draw inspiration from the very highest percentiles, most of whom
I've never met... Winston Churchill for his eloquent courage, Richard
Feynman (a physicist) for his piercing pursuit of truth and unbounded
enthusiasm when finding it, Alfred Brendel (a pianist) for magically
unifying art and intellect, and Mother Theresa (Missionaries of
Charity) for outshining them all by demonstrating the purest form of
love. I'm inspired too by the faithful laborer who lays roof shingles
under the summer sun while we enjoy our leisure. When I'm frustrated
at work, these people lift my spirit.

Quote:
> From that point on, I was hooked, and eventually went on to study
> Physics in college...

I got hooked on physics too, about 10 years younger than you.


Sun, 26 Dec 2004 09:22:41 GMT  
 what are the requirements? [Re: XP Boundaries]
Joseph Bacanskas wrote (abridged):

Quote:
> Second question first -- AFAIK, C3 died.  I haven't checked, but I bet
> the story is on the c2.com wiki.  What I *heard* was that it failed
> when its users were stopped by the system from entering bogus info.  
> The system it was to replace, allowed this and the business users
> decided they wanted that capability.  I must assume that entering
> erroneous data was *not* part of the stated requirements.

Perhaps someone from the C3 project can verify your account. If true,
it demonstrates how hard it is to get good requirements, even when
using XP. Did the project's official client(s) not adequately
represent the wishes of the business users? Boy have I seen that in
spades.

Quote:
> Interestingly, when Kent was here in Seattle last time, he spoke to a
> combined Smalltalk, Java and XP users group meeting.  He talked
> specifically about business organizational change.  He felt (and I am
> paraphrasing here) that it was MUCH more difficult to get business to
> change and that business organizational change was where he should
> focus his energy.  Getting developers to be successful with XP was
> easy by comparison.

I wholeheartedly agree that business improvements give more leverage
than software methodology improvements. If Kent is able to switch his
focus to the business side, he has more flexibility and vision than
most of us old crusty software engineers.


Sun, 26 Dec 2004 09:35:42 GMT  
 
 [ 5 post ] 

 Relevant Pages 

1. I am not deaf, but am I mute?

2. To XP or not to XP?

3. XP Home vs XP Pro - switch ???

4. XP mean Windows XP proffessional

5. Jtag/Boundary Scan

6. Jtag/Boundary Scan

7. LabView/Boundary Scan/JTAG

8. boundary mathematics and FPLs

9. fixed boundaries for a graph

10. jtag boundary scan for fucntional test

11. Quadratic Surfaces + Boundaries TO VRML ?

12. JTAG boundary scan test vectors

 

 
Powered by phpBB® Forum Software