Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!ut-emx!tivoli!alan
From: alan@tivoli.UUCP (Alan R. Weiss)
Newsgroups: comp.software-eng
Subject: Re: design to cost
Message-ID: <392@tivoli.UUCP>
Date: 13 Feb 91 23:36:54 GMT
References: <15038@uswat.UUCP> <1991Feb06.102233.19301@eecs.wsu.edu>
Reply-To: alan@tivoli.UUCP (Alan R. Weiss)
Organization: Tivoli Systems Inc., Austin, TX
Lines: 59

In article <1991Feb06.102233.19301@eecs.wsu.edu> pcooper@yoda.UUCP (Phil Cooper - CS495) writes:
>In article <15038@uswat.UUCP> gbeary@uswat.uswest.com (Greg Beary) writes:
>>
>>We're having some difficulties in that we are trying to apply "normal
>>software business" lifecycles to in-house development. What happens is
>>that a internal client has the concept for a product. They "pitch" the
>>idea to a funding board. The funding board allocates $x to use to build
>>the product. We now begin to gather detailed requirements, estimate the
>>effort- $y, and architect the product. In no case has $x >= $y. What 
>>your are left with is a client that never has the dollars for the
>>functinality they want and a development organization that always "comes
>>up short" in the eyes of the client. 
>>
>
>    Forgive my ignorance, but how in the world can your funding board
>allocate a sum of money for a particular software development effort
>before any requirements are known or estimation done? I don't see how
>they could possibly hope to be accurate in their forecast of the cost
>of the project.  It seems to me that if the in-house client wants a
>particular software application developed badly enough, they can foot the
>bill for at least a preliminary estimation of the scope and cost before
>requesting funding from the board.  Just $.02 worth from a future
>(hopefully) software engineer. 
>
>Phil Cooper

Right.  In large corporations such as Unisys and IBM with formal
process models (Phase Review and Plan of Record, respectively),
little experiements are often funded during Phase 0 and Phase 1
without a lot of paperwork.  At some predefined point, a "stake in
the ground" is planted and the experimenter has to bring their project
under the process.  In doing so, they have to justify the project and
request additional funds.  This can occur at the exit of Phase 1
(Requirements) or Phase 2 (Specifications).  Phase 3 is Development,
Phase 4 is QA and Manufacturing, Phase 5 is Support, Phase 6 is Death
(this is from Unisys' process, which was lifted, er, gleaned from
IBM's process).

I am supportative of internal customer-supplier types of business
models, or even internal market structures such as capital funding
and transfer pricing schemes.  However, I have yet to have heard of
any software development organization who has employed internal markets
successfully.  Anyone else in net.land?


_______________________________________________________________________
Alan R. Weiss                           TIVOLI Systems, Inc.
E-mail: alan@tivoli.com                 6034 West Courtyard Drive,
E-mail: alan@whitney.tivoli.com         Suite 210
Voice : (512) 794-9070                  Austin, Texas USA  78730
Fax   : (512) 794-0623
_______________________________________________________________________



"Quality is never an accident.  It is always the result of high intention,
sincere effort, intelligent direction, and skillful execution.  It represents
the wise choice of many alternatives."