Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

SOFTWARE PROCESS-Improvement and Practice, Vol.

2, 281-289 (1996)

CMM and I S 0 9001


Martyn A. Ould
Quality 6 Technical Director, Praxis plc, 20 Manvers Street, Bath
BAl l P X , UK

Practice Section

In their own ways, both I S 0 9001 and the Capability Maturity Model (CMM) have provided
guidance and urgency to the improvement of software processes. However, there is a point
where each calls for sfatistical analysis of the performance of the processes in use, treating
software development as a manufacturing activity. This paper argues that this is an unhelpful
demand that should be reconsidered by the writers of standards. 0 1996 by John Wiley &
Sons Ltd and Gauthier-Villars
Softw. Process Improve. Pract., 2, 281-289 (1996)

No. of Figures: 0 No. of Tables: 0 No. of References: 8

KEY WORDS: IS0 9001;CMM; SPC; software process improvement

INTRODUCTION IS0 9001 and the CMM brings combined benefits.


I also argue that the tendency of both schemes
A great deal - perhaps too much - effort is to push towards a statistical view of software
expended by organizations trying to answer the development wdl be a fruitless one which should
question ’which should we use: IS0 9001 or the be replaced with approaches appropriate to the
SEI CMM?’. One might as well ask ’which should intellectual nature of software development as a
we use: code inspections or good programming collaborative design activity, and that responsive-
practices?’. We have had a Quality Management ness to change should be promoted at least as
System registered under IS0 9001 (or its precursor strongly as the ability to hone a stable process.
British Standard 5750 Part 1) since 1986, and we
have a TickIT+ certificate. We are looking at the
relevance of the CMM (Paulk et al. 1993)framework USING I S 0 9001 AND CMM TOGETHER
to further improving our software engineering
practices. Drawing particularly on our own experi-
The spirit and the letter of I S 0 9001
ence, I argue in this paper that intelligent use of
~ ~ ~~~ ~~ ~~
There are no guarantees in this world. Buy the car
Correspondence to: M.A. O d d , Quality & Technical Director, with the best safety record on the market and you
Praxis plc, 20 Manvers Street, Bath, BAl IPX, UK. can still drive it like an idiot. A bad driver can
*Origin-ally a UK certification scheme speafic to software pass a driving test.
development that amplified I S 0 9001 and included IS0 9OOO-
3 ( I S 0 9000-3: 1991 - Guide to the application of I S 0 9001 to the ISO9001 is no different. An Iso 9001 certificated
dtwelopment, and maintolance of software), TickIT is Quality Management System (QMS) does not
increasingly being adopted internationally. TickIT-registered guarantee the quality of the software developed
auditors are required to be well versed in software development.
ccc 1077-48661%1020281-1 7.50 under it. Bad software engineers can operate a
0 1996 by John Wiley & Sons, Ltd. and Gauthier-Villars good QMS. In our consultancy work we are used
Practice Section
M. A. Ould

to finding organizations for whom their QMS is IT systems to the organization, something to bring
an irritating appendage. cohesion and hence lasting benefit.
All this is obvious, but it tells us that if we This 'repository' aspect of a QMS is the first that
aspire to an IS0 9001 certificate for our QMS and I want to draw attention to. Our QMS is the major
we want to get some real benefit from the achievement repository of our process, at least in so far as we
it represents then we must live the spirit of the want to capture that process in terms of process
standard as much as the wording. descriptions, as opposed to, say, training courses
In the rest of this paper, you can assume that or in-house or purchased tools. It is also 'our'
when I talk about an IS0 9001 QMS, I mean one repository; it was originally developed by our
that is used in the spirit in which I S 0 9001 is engineers and it continues to be maintained by
intended. In particular, that means not treating it our engineers, with an internal quality assurance
as something that is simply tacked onto daily life, team having editorial responsibility for it.
or as a bureaucratic ordeal. The sort of QMS I am
thinking about is a QMS that is effective in
Building a QMS guided by the CMM
promoting the business's effectiveness.
Interestingly, when we look back at the order in
which we developed our QMS we find it ran
The QMS as repository
roughly as follows:
Some years ago I visited a large organization that
0 have a standard format for documents and a
was well into a major Total Quality Management
standard approach to version control;
(TQM) programme. I had been invited to spend a
have a standard way of peer reviewing anything
day addressing about 30 of their IT delivery
and of closing defects, and start applying it
managers on the topic of I S 0 9001 QMSs. I took
to everything;
on the assignment with some trepidation since, at
institute a project-oriented approach under
the time, I felt rather apologetic about our rather
which all work is managed;
traditional QMS when in the company of people
have a standard approach to the planning and
in the midst of exciting TQM programmes, even
reporting of projects;
though I had real evidence that 'quality' had (and
0 start to define common technical processes and
still has) a real meaning in the company and with
put in place standards for the products of
our staff. I S 0 9001 has this reputation of being a
those processes;
recipe for bureaucracy and the stifling of inno-
0 concentrate on managing requirements;
vation - allegedly quite the reverse of TQM
...
with its emphasis on innovation. Nevertheless, I
prepared my one-day seminar. Does this look familiar? It bears a very close
At the end of the day, I was interested by the resemblance to the order in which one ought to
comments from the people who came. The gist of be doing things to lift oneself from level 1 in the
their feedback was that a QMS was just the thing CMM to (round about) level 3. It was about getting
they needed in their TQM programme. So far they projects under basic control before worrying about
had only heard a great deal of exhortations to technical issues. I once had discussions with a
measure, to control, to improve; they had received client who wanted to introduce quality-oriented
training in team building, process analysis, cause- practices into his IS department, was aware of the
effect graphing and all the paraphernalia of TQM, requirements of I s 0 9001, and had been reading
but they did not know what to do next (or even a lot about measurement. Basically his brain was
first). Some people had indeed started measuring spinning from all the ideas that had been presented
things and putting graphs up at their desks, but to him. I led him through CMM levels 2 and 3
when it came down to it, they had no foundations and showed him how we had, without having the
on which to build. CMM to hand (back in 1983-1986), gone through
They recognized that the real point of a QMS that same foundation laying and then up through
was that it could act as the focus they lacked; a the building floor by floor. In particular, I could
repository for things they agreed as a group were demonstrate to him why his uncertainty about
essential to achieving quality in their delivery of measurement in an organization with an ad hoc
S o b . Process Improve. Pract., Vol. 2, 281-289 (19%) 0 1996 by John Wiley L Sons, Ltd. and Gauthier-ViIlars
282
& Practice Section CMM and I s 0 9001

process was making his brain spin; it was premature 9001 does not require us to do things that do not
to think about measurement without significant make sense for our business.
foundations in place.
What we see is that the CMM gives us a good
plan of action for achieving IS0 9001 certification BEYOND LEVEL 3: GREAT
starting from an ad hoc process, and IS0 9001 EXPECTATIONS
provides us with the notion of a Quality Manage-
ment System, which reinforces the notion of the So, we can see a helpful degree of convergence of
process asset repository and gives us a framework the two schemes at the point which I might
for building our process capability. The two notions characterize by I S 0 9001 certification. There are
neatly support each other (I rather wish we had differences of emphasis in the two schemes, but
had the progression defined by the CMM when these are of no great import to us here. Neither
we started our QMS development). Establishing says ’what is not required is forbidden’.
good software engineering practice by building Beyond this point of rough convergence things
through levels 2 and 3 will generate a QMS get harder.
(whether or not we choose to call it that) and
should, I assert, lead to an I s 0 9001 certificate for
Measurement in I S 0 9001
that QMS (pace Mark Paulk’s (1995) detailed
comparison). There are some pluses and minuses IS0 9001 leaves some openings in two areas. First,
in each scheme but my point here is that the CMM it calls for ‘preventive action‘ which will include
and IS0 9001 are not alternatives from which one ways of detecting, analysing and eliminating
must be chosen. ’potential causes of non-conformities’. I s 0 9000-3
does not address this topic, based as it was on the
earlier version of Is0 9001 which did not deal
The effective QMS with it specifically. So we are left rather in the air
currently. Secondly, I S 0 9001 requires the supplier
IS0 9001‘s requirement that there should be an to ‘identify the need for statistical techniques
auditable trail showing that the plan-do-check-act required for establishing, controlling and venfying
cycle has been carried out is often turned inside process capability and product characteristics’ and
out and portrayed as a demand for unnecessary then to have ’procedures to implement and control
bureaucracy. Those who run an I s 0 9001 QMS the application’ of those techniques. ‘Process capa-
properly - in other words in a way that is fit for bility‘ is used here as a technical term defined
their purpose - know that the only records you (Ishikawa 1990) as ’the performance of a process
keep are those you need anyway for good quality over a certain period of time while in the statistically
management. It is hard to convince people who controlled state’; one way of putting it might be
have seen or experienced an ineffective QMS that ‘the degree of variation in the output of the process
bureaucracy is not a must-have of IS0 9001. In that is due to the process itself‘.
our own QMS, because we have a wide variety of IS0 9000-3 elaborates this by requiring the choice
work for different clients on different architectures and use of quantitative measures of quality as a
for different sorts of systems, we devolve engineer- trigger for ‘remedial action’ should performance
ing decisions about process to individual develop- not meet targets defined for those measures, and
ment projects and this extends to the choice of as a way of defining ‘specific improvement goals’.
quality controls and records for the project. (This This is an open-ended definition because, as it
is not a carte blanche to say ’we are not bothering’ admits, ’there are currently no universally accepted
since the quality management plan for the project measures of software quality‘. This (properly) lets
must be produced and must be approved by IS0 off the hook when it comes to trying to define
management as appropriate for the job in hand.) any ’universal’ measures, but it leaves the more
The rule for project managers is then simple: ‘if practical question to be a’mwered by the individual
you are doing something that you think does not organization: ’what will our measures be and how
add value to the engineering of quality into this will we use them to assess statistical control or
system then you have only yourself to blame’. I s 0 lack of it?‘.
8 1996 by JohnWiley & Sons, Ltd. and Gauthier-Villan Softw. Process Improve. Pract., VoI. 2,281-289 (1996)

283
Practice Section M. A. O d d

Measurement in CMM levels 4 and 5 The CMM and SPC


The CMM explicitly imports the work of Shewhart,
For its part, the CMM also takes us into unknown Deming and Juran on statistical process control
waters in levels 4 and 5. I say 'unknown' because and calls for its use on the software development
it seems clear that they are, at present, largely process. It requires that projects actively bring
speculative. In his first book Humphrey's (1989) process variation within acceptable quantitative
coverage of levels 4 and 5 was comparatively weak, boundaries, using notions such as mean and vari-
reflecting the fact that the role of measurement for ance. It talks about 'idenhfying special causes of
process improvement was still poorly understood. variation within a measurably stable process and
However it, and the subsequent authoritative text correcting ... the circumstances that drove the
from the SEI (1995), proposed the translation of transient variation to occur' (my italics). It follows
the techniques of statistical process control (SPC) traditional SPC principles: 'without controlling the
(e.g. from Deming (1986) and Ishikawa (1990, process within statistically narrmu boundaries (small
sections 2.8 and 4.7.7), and the use of defect variations in process measures), there is too much
analysis for process improvement. I believe the ice noise in the data to determine objectively whether
becomes very thin at these two levels. a specific process improvement has an effect' (my
Humphrey says that 'Levels 4 and 5 are relatively italics), and it deduces that level 4's aim of bringing
unknown territory for the software industry. There the process under statistical control must be
are only a few examples of Level 4 and 5 software achieved before level 5's aim of process improve-
projects and organizations. There are too few to ment can be attempted. Carnegie Mellon's book
draw general conclusions about the characteristics The Capability Maturity Model: Guidelines for Improv-
of Level 4 and 5 organizations. The characteristics ing the Sofhoare Process (1995) is expressly clear
of these levels have been defined by analogy with about this use of SPC (pp. 19-20).
other industries and the few examples in the We need to look closely at the appropriateness
software industry exhibiting this level of process of SPC to software development, since both IS0
capabilitf. Little appears to have changed since 9001 and levels 4 and 5 of the CMM rely on it. In
1989 to provide any more examples to validate the particular, we need to ask 'can we give meaning
definitions of levels 4 and 5. I know one organiza- to the notions of SPC and process capability when
tion in the UK that could reasonably lay claim to applied to the software development process?'.
having a level 5 operation as far as measurement
and process improvement are concerned. It is a
SPC AND SOFTWARE DEVELOPMENT
maintenance group dedicated to handling change
requests to a single system some decades old. Not First we need to understand a little more about
quite the average situation of an IT development the style of SPC called for in the Ch4M and IS09001.
department moving new and legacy systems into A glance at Ishikawa's book (1990, section 2.1)
client/server with RAD, or of a group developing shows us that of the Seven Tools of Quality Control
new systems in avionics, or air traffic control, or (which include Pareto charts and histograms) the
medical instrumentation. main one is the 'control chart' which involves
There is the danger that, whilst we might find 'statistically calculated control limit lines'. These
a few organizations that meet the current level 4 control limits bound the statistical variation
and 5 requirements, they only validate the model inherent in the process; the wider they are apart
for organizations like them. We will only get a useful the less the capability of the process (Humphrey
validation if we perceive that those organizations 1989, chapter 15). Note that they are statistically
are representative of the types of situations prevalent calculated, on the assumption that the feature we
in our industry and that we can, by induction, are measuring about our process conforms to a
demonstrate that levels 4 and 5 offer an effective normal distribution. It is the control chart style of
route for the generality of organizations. I see no SPC that makes possible what the CMM calls for
evidence that this has happened or will happen. in levels 4 and 5. (My investigations suggest,
however, that most people in software development
interpret SIT as Simply People Counting.)
Softw. Procress Improve. Pract., Vol. 2,281-289 (19%) 0 19% by JohnWiley & Sons, Ltd. and Gauthier-Villars

284
Practice Section CMM and I S 0 9001

Control charts are important to SPC because put in place. On a manufacturing line, if I give
they allow the separation of what are called special the green knob half a turn I know - because there
causes and common causes. ‘Special causes’ are those is a physical or chemical causal model available
that are due to a specific situation: things that are to me - that the widgets will be 0.01 nun thinner.
the responsibility of the individual worker, or due In software engineering we simply do not have
to faulty inputs to the process, or to machinery that strong causal model. If we have causal models
going out of alignment. ’Common causes’ are those at all they will be very complex, and, as a
that are a property of the process itself things result, our ability to use measurement in the
whose prevention is in the hands of management manufacturing sense will be correspondinglyweak-
as process owners. The control chart allows us ened. Why otherwise does COCOMO have 15 cost
to discriminate between these; given a set of drivers and Mark I1 Function Point Analysis
measurements, statistical formulae give us upper even more?
and lower control limits which bound the inherent Moreover, we have no a priori reason to suppose
variability of the process. A value falling outside that our software engineering processes behave in
those limits is thus due to a special cause. A a normally distributed way. If we are to use
process which is ’under statistical control’ is one statistics we will need other forms that do not
that is free from variations other than those inherent make this assumption.
in the process itself - the measurements consistently Is this a counsel of despair? No, but it is a
fall between the upper and lower control limits. counsel of caution, especially to the definers of
Only such a process, the argument goes, can international standards. Am I against measure-
sensibly be the subject of process improvement, ment? No, but I believe we have to be clear about
for, if the process is not under statistical control what is a valid role for measurement in a sociological
(i.e. free from special causes), we might f b d design activity carried out in an environment of
ourselves changing the process in response to learning. In particular we need statistical techniques
effects generated by a special cause, with the result that only make assumptions that are true of the
that we actually make the process worse. Hence software engineering environment and that use
the assumption underlying the two-stage approach more complex causal models than SPC.
of levels 4 and 5: first get the process under control
by removing special causes (level 41, and then
How can we use measurement?
improve the process by removing common, i.e.
systemic, causes. I know that in everyday rough-and-ready terms I
can use measurement in order to spot a potential
problem: the module that seems to be taking more
Software development is not a manufacturing
time to code than one would expect for its size;
process
the module that has undergone more than its
However, in everyday software engineering situ- quota of changes given its size; the area of the
ations we cannot expect to be able to apply specification that has generated the largest number
manufacturing-style SPC; the software process - of change requests from the users (what we might
or any small part of it - is not a manufacturing call Spotting Potential Chaos). However, this is
process which we can adjust the knobs on. It is a not the same as using SPC with its assumption of
design activity. As such it is largely an intellectual an underlying statistical distribution model. By all
and a sociological activity, prone to many influences means let us make simple measurements of defect
including changes in the type of work being done, densities and let us spot potential troublemakers,
subtle differences in the complexity of solutions, but this is not an issue for means, or subgroup
changes in the technologies which the development ranges, or standard deviations, or upper and lower
group might be forced to use, changes in staff or control limits, or control charts, or statistical
recruitment policy, changes in general morale due analysis.
to external factors, a new project manager on the I also know that I can keep my measurements
client’s side, and, perhaps above all, the learning over an extended period of time and perhaps, only
that goes on inside the heads of the individuals - perhaps, spot a trend. How do I respond to that
let alone any planned process improvements being trend? Suppose the trend represents an improve-
8 1996 by John Wiley & Sons, Ltd. and Gauthier-Viars Softw. Process Improve. Pract., Vol. 2, 281-289 (1996)

285
Practice Section M. A. Ould

ment. Can I congratulate myself on the process prevention and process improvement but that
improvements I have recently put in place, or is characterizes them in ways that are meaningful
it more likely that any improvement due to them for the software development process and the
is swamped by the fact that this is the second such generality of software engineering groups.
system that this team has built?’ Suppose the trend
represents a degradation. What can I tell from it
except that, for some reason, things have become DEFECT PREVENTION AND PROCESS
worse? IMPROVEMENT WITHOUT FORMAL SPC
So, all that my measurements are telling me is
that things are better or worse, but I get no more For defect prevention, our use of measurement
help. (I can probably tell that things are going must be restricted to its power as an indicator in
wrong, albeit a bit later, through my (CMM level a very crude sense, that of pointing to potential
2) measurements of planned and actual effort use trouble spots (Spotting Potential Chaos). We might
and schedule.) To get that help I must look to use simple and robust data exploration (for
defect analysis: how did that defect get inserted instance, ‘box plots’ described by Hoaglin et al.
and why did it escape detection until now? This (1983)) to help us decide what is an ’outlier’, i.e.
tells us clearly what could be done to improve our what we wish to regard as anomalous and worthy
production process and our verification process. of investigation, and we must then use our
Again we will look for trends so that we can tackle engineering judgement to decide if there was a
the big and/or expensive ones first, but this is still genuine reason for the outlier or if it was an outlier
not a statistical question, simply one of recognizing only in terms of our crude discriminator. No
patterns and thinking through what we do. This algorithms or statistical formulae will replace that
for me is a more constructive aspect of levels 4 engineering judgement. In other words, we do not
and 5 - the institutionalized use of defect analysis need to use a statistical algorithm to tell us which
to seek common causes and remove them - and are the potential anomalies. We will look at anything
it is an aspect that is consonant with IS0 9001. that looks anomalous and make a decision. I
Deming would not, I suspect, have liked my believe this is both practical and methodologically
rough-and-ready approach, but I believe that we sound (enough).
are in danger of being seduced by the lure of SPC. For process improvement we must still rely on
We are dealing with a sociological system, not a defect analysis, but our use of defect analysis
physical one; a soft system, not a hard one. If we should be restricted to the search for patterns, not
demanded equivalent use of SPC by physicists we for statistical significance. We must use engineering
would be demanding from them the measurement judgement to spot those patterns. No pattern
and statistical analysis of their processes for matching algorithm or spurious classification
developing hypotheses, designing experiments, car- scheme wdl replace that engineering judgement.
rying out experiments, and updating hypotheses. We have now disconnected anomaly detection
They do not do any of that; should we then accuse based on measurement from process improvement
them of being unscientific? based on defect analysis, and thereby have effec-
So, for the everyday software development tively removed the ordering of levels 4 and 5. We
group’s environment, I do not believe we can are now using measurement in a realistic way, but
achieve a process which would in Deming‘s terms it is not the way that either I s 0 9001 or the CMM
be recognizably under statistical control: the local want us to use it.2
influencing factors are too many and too strong, A level 5 organization in my definition would
and too unavoidable. SPC in its traditional manufac- therefore be one that was routinely exploiting the
turing sense is inappropriate for the software measurement of quality in the refinement of its processes.
engineering design process. However, all is not This would mean dropping the requirement for
lost. I believe we can find a formulation of levels SPC and distribution-based statistical control and
4 and 5 that still retains the notion of defect
* I S 0 9001 might of course be e x 4 as it was not defined
’ This would not be a process improvement since we have not expressly for Quality Management Systems for software
it - it is only in the heads of our engineers.
instituti~~lized development.
Softw. Pmcess Improve. F’ract., Vol. 2, 281-289 (1%) 0 1996 by John Wiley & Sons, Ltd. and G a u t h i e r - V i

286
Practice Section Ch4M and I s 0 9001

replacing it with the use of the 'simpler' approaches to use the next version of Oracle until they have
of histograms, box plots, Pareto charts, etc., as proven their capability of using it with a pilot?
indicators. As a result, for example: Pilots are a natural risk management strategy, but
can we always be so risk-averse in the face of an
an individual project would measure defect
ever-increasing rate of product change beneath our
densities for some or all of its products; feet? Standing back, perhaps what is needed is the
the project would use coarse criteria based on capability of adopting new technologies, new products,
those densities to identdy potential trouble and new infrastructures, quickly and effectively, rather
spots; than of analysing them slowly.
the organization would consolidate defect den-
sity data at the organization level for compari-
son by projects and for identification by manage- The boring organization?
ment of processes needing attention; One might guess that only an organization that is
the project would carry out causal analysis of well insulated from its environment and that
defects in order to make local process improve-
internally has a high degree of stability of work
ments; area, staff, infrastructure, and technology has any
the organization would consolidate project-
chance of attaining level 5. The organization is
level improvements at the organisational level
probably a high volume organization, repeating
as improvements to the 'corporate process'.
its process hundreds or thousands of times.
The definition of level 5, as it stands, seems to
be a blind alley in that it defines a target that is
WHO IS CMM LEVEL 5 AIMED AT? irrelevant to all but a tiny minority of organizations
that find themselves in one, very unusual situation.
The timid organization? I feel we should be looking for a definition that is
wider, not because it will let more people 'in', but
There is a very strong message in the CMM about because a wider definition could more usefully
the importance of organizational focus and the characterize the sorts of organization that buyers
institutionalization of proven practice. At the higher of software will be looking for; those able to take
levels the degree of caution called for seems to be on new work, to adapt to changing circumstances,
increased and increased; the Technology Change to deal with the large forces at work in a social
Management KPA requires 'pilot efforts' to be activity like software development, and to reliably
performed where appropriate 'to assess new and adapt to new technologies whilst improving its
unproven technologies before they are incorporated processes as far as stability permits. A wider
into normal practice'. What is a new technology? definition would also represent a realistic targef
Presumably one that is new to the organization. for more organizations.
What is an unproven technology? One that is This all suggests a level 5 that promotes sound
unproven by the organization or unproven by continuous process improvement at the organiza-
anyone anywhere? What level of proof is required tional level whilst recognizing at the same time
for a technology to be proven? Praxis has been an the sociological and intellectual content of software
innovative user of mathematically formal methods development and the reality of change that software
for system specification because, as engineers, we developers actually have to manage.
believed that that was a proper way in which to The call for strict SPC could breed only inward-
increase the quality of specifications and resulting looking, ossdymg, risk-averse, dull, and increas-
systems. The problem of finding a convincing pilot ingly outdated software development organiza-
for new technology is a familiar one, and I would tions. The search for stability of process called for
hazard that the majority of innovations come in the CMM is only there to make the separation
because someone knew in their engineering bones of special from common causes possible with SPC -
that 'this is the right way to do things'. In the but is not this a case of the tail wagging the dog?
eyes of the CMM, this sort of 'gut-feel' decision- (Remember that I am talking here about level 5
making puts at risk the established process and competence for an organization undertaking a number
lacks organizational focus. Must a company refuse of projects. The situation for the individual project
0 19% by JohnWiley & Sons, Ltd. and Gauthier-Villars Softw. Procffs Improve. Pract., Vol. 2,281-289 (19%)

287
@ Practice Section M. A. Ould

is of course quite different; a project can be a design activity. The resulting ‘interpretation’ of
relatively closed world with a single problem to levels 4 and 5 would still characterize organizations
solve, and with fixed infrastructure and techno- committed to continuous improvement but do it
logies. Its chances of achieving stability are greater in a way that makes it applicable to the generality
given the greater likelihood of a stable environment, of development organizations, rather than to a
and, more importantly, it is to its advantage to narrow set whose processes can be more
achieve that stability and to exploit it. The major mechanistic.
example given in The Capability Maturity Model: The CMM’s higher levels recognize an organiza-
Guidelines for Improving the Software Process is for tion’s ability to hone its process over time, but
the Flight Software Project at IBM Houston which perhaps the most important process a software
is just such a project.) development group must acquire is that of taking
on and adapting to change very quickly. If change
is the norm and stability the exception in today’s
T h e learning organization?
software development environment, then perhaps
This surely is what both I S 0 9001 and the upper the industry - and the CMM - should value the
levels of the CMM are trying to get at. We want ability to deal with change above the ability to
to say ‘this organization has got its process act achieve stability.
sorted out and learns’. In this way, level 5 could If our aim is to capture the characteristics that
be used to speclfy an organization that can be make for a learning organization, SPC is not the
trusted to deal with the change that is endemic in foundation we should build our international
our world, not one that has spent its time trying standards on. We need to drop the distinction
to isolate itself to the point where SPC can be between CMM levels 4 and 5, which arises princi-
applied (even if it were applicable). The mistake pally from the distinction between special and
is to spec* a learning mechanism - SPC - which common causes, and to concentrate on the real
requires a priori that software engineering processes learning process at the individual and the organiza-
operated by software engineers exhibit two-dimen- tional level.
sional statistical variation conforming to a normal
distribution. Organizational learning is a large,
complex, multi-dimensional thing not reducible to ACKNOWLEDGEMENTS
the application of SPC. My thanks go to colleagues who have contributed
to the debate in Praxis and helped shape my views,
in particular Dave Deans, Mike Hewson, Tim
SUMMARY Huckvale, George May, Chris Miller, and Chris
Warren, and to others in the wider software
IS0 9001 and the CMM up to level 3 are mutually engineering community.
supportive and complementary. An organization
should not see them as mutually exclusive. It can
aspire to a level 3 rating whilst using an IS0 9001
QMS. It can aspire to an IS0 9001 certificate for REFERENCES
its QMS whilst using the progression in the CMM
Deming, W. E. 1986. Out of the Crisis. Cambridge
to get there. Anyone doing either will want to University Press.
take from each whatever suits their business.
Defect prevention and process improvement, Hoaglin, D. C., Mosteller, F. and Tukey, J. W. (eds) 1983.
called for in general terms in IS09001 and detailed Understanding Robust and Exploratory Data Analysis. Wiley.
in CMh4 levels 4 and 5, are traditionally discussed
in terms of SPC, but control chart based SPC is Humphreys, W. S. 1989.Managing the Software Process.
Addison-Wesley.
neither appropriate to, nor is it necessary for,
defect prevention and process improvement in the Ishikawa, K. 1990. Introduction to Quality Control. 3A
software development process. Cruder approaches Corporation (translated by J. Loftus).
can be used that recognize the intellectual and
sociological context of software development as a ISO.1994. ISO 9001:1994. Quality systems - Model for
S o h . procesS Improve. Pract., Vol. 2,281-289 (1996) 0 1996 by John Wiley & Sons,Ltd. and Gauthier-Villars
288
@ Practice Section CMh4 and I S 0 9001

quality assurance in design/development, production, Paulk, M. C. 1995. How IS0 9001. compares with the
installation and servicing. CMM. IEEE Software, 12, No. 1, 7 M 3 .

Paulk, M. C., Curtis, B., Chrissis, M. B. and Weber, C.


V. 1993. Capability Maturity Model for Software. Version Software Engineering Institute. 1995. The Capability
1 .l. Software Engineering Institute, CMU/SEI-93-TR-24, Maturity Model: Guidelines for Improving the Software
Camegie Mellon University, Pittsburgh, PA. Process. Camegie Mellon University, Addison-Wesley.

EDITOR’S COMMENT

The Capability Maturity Model and the ISO9OOX standards build the basis of software process improvement.
There is some debate how both are related and to what extend they contribute to improved software
development. Ould points out that both are useful for driving software process improvement into the right
direction, but that they cannot guarantee high quality software. He clearly demands that the spirit of the
standard has to be lived (in contrast to just the following the words of its definition). He claims, that the
intelligent use of IS09001 and CMM brings combined benefits. Additionally, he sketches his experiences
with a statistical view on software process improvement. This experience is that software processes
cannot be controlled by statistical means, but that software process should be understood as collaborative
design activity. The experiences of Ould are based on building up a quality management system (which
served as repository of knowledge about the software process) and from several consulting projects. That
is why the practical experiences presented in this paper should be worthwhile for all practitioners who
plan to introduce quality management.
-VG

@ 19% by John Wiley & Sons, Ltd. and Gauthier-Villars S o h . Process Improve. Pract., Vol. 2,281-289 (19%)

289

You might also like