People who confuse RUP with Agile
- I had the following response to an analysis of RUP, XP,and Agile in the OTUG
I appreciate your effort to cast light on the discussion of agile and
heavyweight, or - as it is now being referrred to - empirical and defined. I
wish your future efforts would allow cutting and pasting, however, because
the effort of recollecting your thoughts manually is tedious.
You state that "a methodology specifies who does what activities on what
work products, including how, why and when such activities should be done
for various types or projects; and a process ... (more like the previous)
In this statement you have dissected the difference between agile and
heavyweight, not elsewhere in your paper. As first discussed by Jim
Highsmith's book on adaptive development and further clarified in a number
of other books, articles and talks coming up this fall, agile processes do
not have any of the characteristics that you state above. Those
characteristics are only those of heavyweight methodologies and processes.
As stated in process control theory, defined processes have the methodology
characteristics you stated. These are referred to as defined process models.
They are used when everything can be well enough defined that the noise from
the processes doesn't interfere with their predictable operation.
The other process control model is the empirical model, where it is
understood that the activities are not predictable, are non-linear, and are
noisy. Constant inspection and subsequent adaptation are used as a control
The agile alliance states that the failures of systems projects over the
last twenty years can be directly laid to the feet of using an inappropriate
process to control projects ... that the defined process model (based on the
Taylor manufacturing model) is just wrong for software projects and won't
work no matter how hard you tweak it. When the Standish group bemoans that
over 2/3 of all projects fail or vastly overrun, they mean that thee
projects fail to live up to the predictions and definitions provided by the
methodology. The people don't fail, the definition of the project fails. The
work doesn't fail, the expectation set by the methodology fails. Such an
The agile alliance uses processes that inherently rely on the empirical
model of process control. They allow requirements and architectures to
emerge, and teams of professionals adapt to what emerges. Inspections occur
when checking the results of each iteration's increment, as well as through
daily status meetings.
I know you won't be the last, and you certainly aren't the first to propose
that the difference between RUP and XP, heavyweight and agile, and defined
and empirical are simply semantic and the "rup"ture can be resolve. Not
true. This is a basic model difference of how to manage and control
development projects where the requirements, technology, and people iteract
to create complex, difficult to predict patterns of activities.
I hope this helps.
From: owner-otug@... [mailto:owner-otug@...]On Behalf
Of Sinan Si Alhir
Sent: Sunday, August 12, 2001 1:13 AM
Cc: Sinan Si Alhir
Subject: (OTUG) XP, the Agile Alliance, and RUP
The following paper may be of interest:
* "XP, the Agile Alliance, and RUP"
This paper elucidates the reality beyond the surface of the debate between
heavyweight approaches and lightweight or "agile" approaches by delineating
between the Unified Modeling Language (UML) and computer programming
languages, deciphering the concepts of "weight" and "agility" relative to
methodology and process, and exploring and bridging the chasm between XP,
the Agile Alliance, and RUP to conclude that *RUP is a more massive but more
agile methodology than XP, from which appropriately weighted and more agile
processes than XP may be derived, with which more agile projects than XP may
be executed*, and given RUP's broad scope, breadth and depth, and
flexibility as a process framework and given XP's collection of practices,
*the chasm is bridged via applying RUP, as a process framework, within which
XP, as a collection of practices, may be leveraged*.
I welcome your comments.
Sinan Si Alhir