Good Pla

j

Slike

New develop-
meni methodolo-
gies like rapid
- applications de-

executives are under pressure to keep
their core applications running and
maintain existing systems that are criti:
cal to the business. This tends to con- ' ;
sume so much of their time and energy
that they can barely consider a major
change in the IS development culture.
They don't have any spare stafi with

Because IS has slow development cy-
les and because maintenance of the old
systems is difficult, there are large ap-..;
plication backlogs. An IS department
with backloga like those shown in the

figure (at right) severely damages the
ability of a corporation to adopt new:
computerized business procedures. To -

RAD life cycle be introduced. k
Rather than attempt to change the en-
tire IS organization, the best way to in:
troduce RAD technigues is to start
small. A small group should be responsi:
ble for selecting the tools and life-cycle
technigues for fast development. Small,
specialized construction teams should be
established, and their members should
be welktralned to use the new, tools and
technigues.:ivi |: saje
A goal should be to make the
brightest and most capable developers

powerful tools. These special teams of

to work on pilot projects that are non-:

ized to the corpora-

JE eti lake

!. which the new
€ are forced to conform to older,

which to take risks: ie tdi

Si reabat

two, three or four people should be put h
critical but stra t
ga tegically important to,,.4«
The teams should use a cor k
mputerized
methodology customized č
tion and its chosen tool set. As the team

< PO WEEKNAPPLIGATION DEVELOPMENT

APPLIED INTELLIGENGE

"work:to-rule" attitude, in
methodology and tools
much
less productive development rules.
'The low-productivity figures-with
CASE tools in many organizations are

adopting a

. the result of developers not wanting to
; use them. These people are comfortable
| only with existing technigues and don't

want to climb a new learning curve.
A skilled and well-motivated construc-
"tion team can demonstrate the value of
the RAD life cycle and can tune the
methodology to make it as effective as
possible. Once a team has proven the

;- value of the new development process,
). , more teams can be added.

may be an inappropriate
choice of a tool. In this case, it's impor-
tant to find a tool set that does what is
reguired and switch to it. Managers —
should seek out integrated com; uter-aid-
ed software engineering (CASE) tools
that are capable of generating 100 per-
cent of the code for an application from
design specifications.

There are many unfortunate examples
of IS organizations becoming committed
to an inadeguate tool set, such as a non-
integrated CASE tool. These tools re-.
auire hand-generation of the application
from design specifications and should be
avoided. The switch to a different tool
set is painful; but if it's needed, it

'The problem

The Extent
Of Application-Development Backlogs

Use of Rapid Application Development
Technology, Tools, Can mi oš

Cut Backlog

compete effectively, it's critical that a - ;

as skilled as possible with the most ''.-.

; Productivity, cycle time and guali
s V are measured ani the (Dia o di
Sram
promotions

technigues. >: Sometimes N
JO A first attempt

blok despite positive attitudes ubont ba
bank post... projec- When tno happens the cause of
People who. should be identified, the situation.

are firmly wedded to old technigues | Corrected, and another attempt, should be
blorta see inad The developero shouldnt be penal-
ali too one initial failure; on the

demonstrate that new tools don't easily. contrary, they should be rewarded for a.

ee
k
; 1
inder $
tlom
ito$d:
bililon
ver $5
Glinor E

should be done as early as possih
fore many people have climbed tne Le
PRSRH With the inadeguate

Sometimes the reason for failure is in-

: have li M bi Organizations

The comrni
Ment to good training is vital, and is

DPRL 2, 1990

Means Success in Implementing RAD

training should be oriented to fast devel-
opment.

The first applications to be developed
with the RAD methodology should fit
comfortably within the capabilities of
the tool set. They should be applications
that can be accomplished by one small,
skilled construction team in three
months or so. Within these constraints,
the applications should be appropriately
complex—between 600 and 1,000 func-
tion points. A purpose of the first
project is not only to build systems but.
to demonstrate that the methodology
works well.

To develop systems fast, its essential
to bypass bureaucratic delays. As its
victims know, bureaucracy insists on
formal procedures. It emphasizes filling
in the right forms, going through the
right channels and getting the right ap-
provals, rather than finding out how to .
get the job done as guickly and effec-
tively as possible.

Bureaucracy is the enemy of speed.
Indeed, it's the enemy of the three main
goals of RAD—higher guality, lower
cost and rapid development. Many at-
tempts to create new computer applica-
tions have guickly ground to a halt be-
cause of career bureaucrats or managers
motivated by politics rather than effi-
ciency.

Know Your Enemy

In a healthy competitive corporation,
the enemy is on the outside. In a bu-
reaucratic organization, however, the en-
emy tends to be internal, in other de-
partments.

This is one reason why successful en-
trepreneurs have difficulty transferring
their executive skills to large organiza-
tions. 'They expect loyalty from a staff
seeking common goals for the enterprise;
instead, they find that predominant
goals relate to internal politics. Instead
of hostile external forces, they find hos-
tile internal forces.

'The operation of small, specialized
construction teams, with good measure-
ments, can bypass the bureaucratic ten-
Pe in IS organizations. The execu-

dve owner and his IS counte; m
agree that rapid action is SVA, pre
the user community as well.

1£ bureaucracy threatens to interfere

, pvith the process, the executive owner

must be responsible for cu the
tape. As in IS, the players from do cor
K epa should be determined to
lon gui to
s mori ickly to make the new
final article in this seri
les, I will
osti pkoj kata that lead to suc-
ment life cycles. 8 ee?

oe oai iii oza
The concepts embodied t
OO o oje a ke ae

ume in the
Martin Report Series. For pa upaš

mation on this voli
1240. For informatioi on sodo

Lancs, LAS 9BX United Kingdc
(0524) 734 506. m

?