[SystemSafety] Collected stopgap measures
Steve Tockey
Steve.Tockey at construx.com
Fri Nov 16 17:34:52 CET 2018
Derek,
I kindly request that you don¹t put words into my mouth.
You wrote: ³I'm assuming that you were paid to do the second
implementation because the first was such a mess.²
While that is a correct assumption in a few of the cases, it is certainly
NOT a correct assumption in all cases.
Please re-read and pay attention to what I wrote: ³Consider an
organization that has built several systems of Type X before, whatever
Type X happens to be. So when they are given a new system of Type X they
can estimate it based on their past experience.²
These organizations HAVE successfully built several systems of some type.
Not one, several. They have enough experience with successfully delivering
systems of Type X that they can estimate a new product of Type X based on
their past experience.
I am comparing the actual results of using an engineering approach to
(successfully) building a system to a somewhat reasonable estimate of what
it would take to (successfully) build that same system using the
mainstream approach. Their estimate is based on that organization¹s actual
experience of having (successfully) delivered using the mainstream
approach.
So, given that the organization should likely have been successful in
delivering one more system of Type X (because they have already shown to
have been able to do it in the past) then it should be entirely reasonable
to compare our actual results to their estimates for this new system.
Again, as I said, I am ignoring the fact that they would most likely have
over-run their own estimates had they built that new system in the
mainstream way. I am comparing our actual performance to their estimated
performance, not to a more reasonable expectation of actual performance
based on industry data. It would actually be safe to say that the
performance data of the engineering approach is even better than what I
stated.
Now, there have been a few cases where we were called in because, as you
say, the first was such a mess. In one specific case, the organization had
already tried TWICE and failed miserably both times to build a highly
complex, mission critical system. Using the engineering approach, this
system was delivered on-time against OUR estimates. So I take the fact
that we actually delivered when TWO previous attempts had failed miserably
as more evidence that an engineering approach can and does make a
difference.
As for your statement, ³Then, second time around you have the benefit of
what has been learned from the first project.² Well, if that were true,
then why was their second attempt also such a miserable failure?
‹ steve
-----Original Message-----
From: Derek M Jones <derek at knosof.co.uk>
Organization: Knowledge Software, Ltd
Date: Friday, November 16, 2018 at 7:37 AM
To: Steve Tockey <Steve.Tockey at construx.com>
Cc: "systemsafety at lists.techfak.uni-bielefeld.de"
<systemsafety at lists.techfak.uni-bielefeld.de>
Subject: Re: [SystemSafety] Collected stopgap measures
Steve,
> Of course I wish I could have built several of exactly the same system
>in parallel. Unfortunately I have only been able to do that once. You
>realize it is too expensive to do that in the general case.
If governments are really interested in learning about software
production, they need to fund such experiments.
> When we implement that same new Type X system using an engineering
>approach we, as I said, finish in about half the time, with about half
>the cost, and with no more than one tenth of the delivered defects when
>compared with the mainstream estimate.
I'm assuming that you were paid to do the second implementation because
the first was such a mess.
So your data points are based on a sample that is so bad somebody is
willing to pay for another implementation.
Then, second time around you have the benefit of what has been learned
from the first project.
How can the second implementation not be faster and cheaper?
Almost any approach you use would be faster and cheaper, given
you are comparing against such a poorly performing project who
did not have the benefit of prior implementation experience.
--
Derek M. Jones Software analysis
tel: +44 (0)1252 520667 blog:shape-of-code.coding-guidelines.com
More information about the systemsafety
mailing list