Extreme Programming - Research shows that timeboxing ...

Author

Research shows that timeboxing ...

Isaac Gouy

2005-02-01, 4:00 am

"Research shows that timeboxing itself brings benefits in terms of
increased productivity." page 54
Craig Larman, "Agile and Iterative Development: A Manager's Guide"

"Timeboxing by itself has been shown to have a productivity effect.
DuPont, one of the earliest timebox pioneers, found developer
productivity around 80 function points per month with timeboxed
iterations, but only 15 to 25 function points for other methods.
[Martin91]" page 77
Craig Larman, "Agile and Iterative Development: A Manager's Guide"

Unfortunately I don't have a copy of James Martin's old Rapid
Application Development book, maybe someone else can check what it
says?

The only information I've found online *doesn't* say that "timeboxing
by itself" was found to be more productive; it does say that
'timeboxing' used with a CASE tool, was more productive than
'traditional' used with COBOL Pl/1 4GLs.

'Small teams of developers at DuPont Fibers, using "timebox"
development and CASE tools, achieve an average life-cycle productivity
of about two hours per function point, or 80 function points per
person-month. This is more than four times the productivity DuPont
achieved when using a traditional life cycle and fourth-generation
languages in conjunction with COBOL and PL/l.'

http://sysdev.ucdavis.edu/WEBADM/do...age-studies.htm

This couple of pages also sounds interesting: 'The term 'timebox'
was first used by Scott Shultz of DuPont, as a key component of Rapid
Iterative Production Prototyping (RIPP), a predecessor of RAD, which
Shultz developed in the early 80's. James Kerr and Richard Hunter
interview him in Inside RAD: How to build fully functional computer
systems in 90 days or less. (McGraw-Hill, 1994, pp. 14-16.)'

John Roth

2005-02-01, 4:00 am

You might also look at "Lean Software Development",
by Mary and Tom Poppendieck.

See p27ff Iterations. They say:

"If a manufacturer wants to start applying lean production
principles, there is one starting point that always works:
use just-in-time inventory flow.
....
There is an equivalent universal starting point for all
agile software development approaches: iterations.
An iteration is a useful increment of software that is
designed, programmed, tested, integrated and
delivered during a short, fixed timeframe.
[end quote]

Notice the qualification "fixed" in the last
sentence.

John Roth

"Isaac Gouy" <igouy@yahoo.com> wrote in message
news:1107220640.650150.271750@c13g2000cwb.googlegroups.com...
> "Research shows that timeboxing itself brings benefits in terms of
> increased productivity." page 54
> Craig Larman, "Agile and Iterative Development: A Manager's Guide"
>
> "Timeboxing by itself has been shown to have a productivity effect.
> DuPont, one of the earliest timebox pioneers, found developer
> productivity around 80 function points per month with timeboxed
> iterations, but only 15 to 25 function points for other methods.
> [Martin91]" page 77
> Craig Larman, "Agile and Iterative Development: A Manager's Guide"
>
>
> Unfortunately I don't have a copy of James Martin's old Rapid
> Application Development book, maybe someone else can check what it
> says?
>
> The only information I've found online *doesn't* say that "timeboxing
> by itself" was found to be more productive; it does say that
> 'timeboxing' used with a CASE tool, was more productive than
> 'traditional' used with COBOL Pl/1 4GLs.
>
>
> 'Small teams of developers at DuPont Fibers, using "timebox"
> development and CASE tools, achieve an average life-cycle productivity
> of about two hours per function point, or 80 function points per
> person-month. This is more than four times the productivity DuPont
> achieved when using a traditional life cycle and fourth-generation
> languages in conjunction with COBOL and PL/l.'
>
> http://sysdev.ucdavis.edu/WEBADM/do...age-studies.htm
>
>
> This couple of pages also sounds interesting: 'The term 'timebox'
> was first used by Scott Shultz of DuPont, as a key component of Rapid
> Iterative Production Prototyping (RIPP), a predecessor of RAD, which
> Shultz developed in the early 80's. James Kerr and Richard Hunter
> interview him in Inside RAD: How to build fully functional computer
> systems in 90 days or less. (McGraw-Hill, 1994, pp. 14-16.)'
>

Isaac Gouy

2005-02-01, 4:00 am

> You might also look at "Lean Software Development",
> by Mary and Tom Poppendieck.

No thanks ;-)

Thanks for your previous mention of Crystal Clear, it prompted me to
take a look at the articles
http://alistair.cockburn.us/crystal...irsarticles.htm

(And the dissertation gave quite a nice reprise of his thinking, once
it got past the research method justification.)

At present, I'm trying to reconcile the "Timeboxing by itself" comment
with 'using "timebox" development and CASE tools'.
Maybe the productivity was due to the CASE tool.

Isaac Gouy

2005-02-01, 8:57 pm

Isaac Gouy wrote:
> "Research shows that timeboxing itself brings benefits in terms of
> increased productivity." page 54
> Craig Larman, "Agile and Iterative Development: A Manager's Guide"
>
> "Timeboxing by itself has been shown to have a productivity effect.
> DuPont, one of the earliest timebox pioneers, found developer
> productivity around 80 function points per month with timeboxed
> iterations, but only 15 to 25 function points for other methods.
> [Martin91]" page 77
> Craig Larman, "Agile and Iterative Development: A Manager's Guide"
>
>
> Unfortunately I don't have a copy of James Martin's old Rapid
> Application Development book, maybe someone else can check what it
> says?
>
> The only information I've found online *doesn't* say that "timeboxing
> by itself" was found to be more productive; it does say that
> 'timeboxing' used with a CASE tool, was more productive than
> 'traditional' used with COBOL Pl/1 4GLs.
>
>
> 'Small teams of developers at DuPont Fibers, using "timebox"
> development and CASE tools, achieve an average life-cycle
productivity
> of about two hours per function point, or 80 function points per
> person-month. This is more than four times the productivity DuPont
> achieved when using a traditional life cycle and fourth-generation
> languages in conjunction with COBOL and PL/l.'
>
> http://sysdev.ucdavis.edu/WEBADM/do...age-studies.htm
>
>
> This couple of pages also sounds interesting: 'The term 'timebox'
> was first used by Scott Shultz of DuPont, as a key component of Rapid
> Iterative Production Prototyping (RIPP), a predecessor of RAD, which
> Shultz developed in the early 80's. James Kerr and Richard Hunter
> interview him in Inside RAD: How to build fully functional computer
> systems in 90 days or less. (McGraw-Hill, 1994, pp. 14-16.)'

The good news: the library had a copy of James Martin's "Rapid
Application Development" book.

The bad news: the chart of IT productivity for DuPont Fibers shown on
page 225 of the book does match the chart and information shown here:
http://sysdev.ucdavis.edu/WEBADM/do...age-studies.htm

The chart tries to show average annual productivity in 2 categories:
1) traditional lifecycle using COBOL, PL/1, FOCUS, NATURAL
2) timebox methodology using the Cortex CASE tool.

Clearly, this is not "Timeboxing by itself"; it's "timeboxing" with a
CASE tool.

"DuPont uses the Cortex toolset and originally built its methodology
around the Cortex Application Factory operating on a DEC VAX...
DuPoint is achieving an average lifecycle productivity of about two
hours per function point (80 function points per person month)" page
224

Earlier in the "Rapid Application Development" book, there's a report
of an empirical survey of 30 applications from 20 different sites that
used the Cortex toolset. The report gave 2 to 45 (average 13x) more
productivity with Cortex than with COBOL. (page 132)

In other words, the improved productivity at DuPont is less than the
average improvement seen by other users of the CASE tool.

Isaac Gouy

2005-02-25, 3:57 am

> This couple of pages also sounds interesting: 'The term 'timebox'
> was first used by Scott Shultz of DuPont, as a key component of Rapid
> Iterative Production Prototyping (RIPP), a predecessor of RAD, which
> Shultz developed in the early 80's. James Kerr and Richard Hunter
> interview him in Inside RAD: How to build fully functional computer
> systems in 90 days or less. (McGraw-Hill, 1994, pp. 14-16.)'

Question: 'How important is the tool set? Is it possible to apply the
new skills and methods without investing in CASE tools and all the
training you need to make them work.'

Scott Shultz: 'The tools are critical. My original name for the process
you call RAD was Rapid Iterative Production Prototyping, RIPP. It's a
kind of evolutionary development, where you build successive iterations
of full-scale prototypes very rapidly, evolving into the finished
product. You can't do rapid iterations of a prototype, or anything
else, in a hand-coded environment. ...'

....
Question: 'How do you measure RAD projects?'

Scott Shultz: '... I prefer to measure and manage something I call
"Balanced Delivery," which is essentially the point at which the users'
ability to absorb and turn around deliverables is equal to the team's
ability to produce the deliverables. At Balanced Delivery, a given
project is moving at the optimum speed for that particular project and
the circumstances that surround it. ..."

Sponsored Links