[trustable-software] The P90X of Software Development
Andrew Banks
andrew at andrewbanks.com
Wed May 23 12:00:41 BST 2018
Heresy, heresy... burn him at the stake!!!!
-----Original Message-----
From: trustable-software [mailto:trustable-software-bounces at lists.trustable.io] On Behalf Of trustable at panic.fluff.org
Sent: 23 May 2018 09:13
To: Trustable software engineering discussion
Subject: [trustable-software] The P90X of Software Development
A recent Register Article states "Agile Development exposed as
superstition"
https://www.theregister.co.uk/2018/05/16/agile_development_exposed_as_techie_superstition/
Software Development is enacted like a religion. That is to
say it is full of faith that particularly behaviours will bring around
salvation but there is remarkably little science and research around what
delivers reliable software.
The great problem with the various 'Development Religions', such
as Agile, Lean, XP, Waterfall, DevOPS, is that they say 'any way you can
make it work is good enough'. The problem is they don't define the
measures or the techniques in any detail except hand wavy approaches. [For
example, there is no agreed way to use Version Control, record
requirements, estimate effort or costs.
There are a growing number of religious cults around offering
salvation, such as the DevOPS institute and ITIL Certification but all of
these have a get out.. 'You didn't do it the right way' but
don't publish the 'right way'..
So... lets look at what works in other arenas. One such arena is
personal training. Tony Horton's P90X program
https://en.wikipedia.org/wiki/P90X
is prescriptive and declares its measures ahead of time. He even
encourages people to publish their success.
It's my believe that what is missing in the development market places
is a fast prescriptive way to lead people to be productive. In effect
program that says, like P90X, spead 90 minutes every day for 90 days and
you'll have mastered the processes critical thinking, measurement and
development which will give you a faster and more reliable development
process.
At the heart of my thinking is importance of encouraging measurement
and experimentation. Doug Hubbard's book "How to measure Anything"
https://www.amazon.co.uk/dp/1118539273
is a good starting place. As is Fredrick P. Brooks "Mythical Man Month"
https://www.amazon.co.uk/dp/0201835959
and an understanding forecasting. Philip Tetlock and Dan Gardner
"Superforecasting" is a great introduction to training people
https://www.amazon.co.uk/dp/B00Y78X7HY
Most recently even Gene Kim and colleages are begining to call out
the DevOPS/Lean approach in "Accelerate: The Science of Lean Software and
DevOps: Building and Scaling High Performing Technology Organizations"
https://www.amazon.co.uk/dp/B07B9F83WM
So why raise these issues... because increasingly I'm working with
organisations who need prescription and evidence. Just like those who take
to P90X they want something which isn't perfect but at least 'trustable'
and has evidence behind it's delivery and reliability.
I wonder if within this forum we could agree on even the branching and
merging strategy used within source code control ?
If we are unable at least to get to that point perhaps there is
nothing we can agree on about software development and we should agree
that all we have is 'WOO' and no science
Edmund
--
========================================================================
Edmund J. Sutcliffe Thoughtful Solutions; Creatively
<edmunds at panic.fluff.org> Implemented and Communicated
<http://panic.fluff.org> +44 (0) 7976 938841
_______________________________________________
trustable-software mailing list
trustable-software at lists.trustable.io
https://lists.trustable.io/cgi-bin/mailman/listinfo/trustable-software
More information about the trustable-software
mailing list