April 29, 1998
IT skills drought worsens
Commentary by Bertrand Meyer
The story
We noticed this article somewhat belatedly, since it
appeared in the January 19th issue of Information Week,
but it is still highly relevant.
The article is one of several recent ones covering
an issue that may turn out to be even more portentous
for the industry than massively visible problems such
as Year 2000 updates and the Euro conversion: the growing
shortage of skilled Information Technology workers.
Article excerpts
(These are only excerpts; read the magazine for full details.)
The already-critical shortage of IT skills is getting worse --
and it's driving IT executives to seek new solutions.
A [recent survey released last week by]
the Information Technology Association of America [indicates]
there are 346,000 vacant positions for programmers, systems analysts,
and computer scientists and engineers -- 10% of the 3.4 million
IT jobs in the country.
"The problem has been getting much, much worse over the last
years", says Honorio Padron, CIO at computer retailer CompUSA
in Dallas. "It's harder to find people, and when you get them,
they stay for much shorter periods". CompUSA is shortening
deadlines on internal IT projects. "You have to try to finish
them before people leave and before the remaining people
get burned out", Padron says.
[The article goes on to describe difficulties encountered
by other companies and steps being considered to limit
poaching of developers.]
Our comments
The figures given appear believable in light of earlier
estimates in the 300,000 range. The situation in other
industrialized countries is comparable, so that the
worldwide shortage is probably getting close to one
million people.
With some notable exceptions, the Information Technology
industry as a whole has not yet paid much attention to the
issue of software quality. Partly (but only partly)
for good reasons,
executives have had focused their energy on releasing
something -- anything, provided it comes early, or not
too late. The success of phrases like "good enough
software" (Ed Yourdon) is representative of this state
of mind.
The situation will not last forever and the IT skill
shortage may be the wake-up call. Although the managerial
techniques touted in the press help -- resorting to
various devices to increase employee loyalty, good-citizenship
pacts between large companies to limit poaching, improving
university links, retraining non-IT workers --, their
effect is limited, and not immediate. All companies must
face the prospect of key developers leaving, or of
not finding enough good people to staff a project.
Here quality becomes crucial, especially such quality
factors as software extendibility (ease of change),
readability, self-documentation, reusability,
robustness of the design.
These are some of the areas where Eiffel shines,
and indeed the method, language and development
environment can make a key difference.
ISE clients routinely tell us that thanks to Eiffel they
are able to complete successful projects using one half
or less of the human resources that they would otherwise
need. Some of the principal contributions here include:
The clarity and simplicity of the syntax.
Syntax matters! When someone has just left a project
hanging out to dry, it makes a difference whether you
have a clear, readable notation, or heaps of ampersands
and squiggly brackets.
Structure. Eiffel is draconian about modularity: no
global variables, no sneaky behind-the-scenes dependencies
between modules. This can make a world of difference
when you are asked to take over someone else's project
in a bind -- and need to limit the amount of
code you need to understand right away, and the
potential impact of any change you attempt.
Design by Contract. This brings tremendous benefits
for the reliability of the software and the ease with
which a new team member can find out what's happening
and, if something goes wrong, why.
Self-documenting software. In the Eiffel approach,
powerful tools are available to extract high-level documentation
from the code itself, both at the class level through the
notion of short form (the class interfaces, produced
automatically from the software text by tools of
the environment), and at the system level through
the impressive high-level architectural
diagrames produced by EiffelCase..
The power of the debugging mechanism, to trace through
the execution of a system.
The portability of ISE Eiffel, allowing easy adaptation
to new platforms as they come along or as your company's
IT strategy changes.
The openness of the environment, facilitating reuse of
other software, Eiffel or non-Eiffel.
The list goes on. Beyond individual techniques and tools,
Eiffel's contribution is a constant focus on quality for
long-term project development. Eiffel is not meant for
fly-by-night software hacked by a few developers who may
be here today and gone tomorrow. It places a constant
emphasis on producing software that will be able to
grow harmoniously with the company that produced it,
without having to rely on geniuses, wizards or gurus.
In short, Eiffel is software engineering in the proper
sense of the term.
If you can afford to hire only geniuses -- enough
of them -- and retain them forever, you may not need
Eiffel. If, however, you are concerned about the
productivity and safety of your software investment
in an industry where everything is volatile --
hardware, software, trends and, most of all, people --
then Eiffel can transform the IT skill drought from
a crisis to an opportunity.
Reference
IT Skills Drought Worsens, in
Information Week, January 19, 1998, page 30.
The magazine's Web site is at
http://www.informationweek.com.
Of related interest: an article in Information Week's
"Secret CIO" column, available on the Web:
IT People Aren't
Plumbers, by Herbert W. Lovelace (a pseudonym).
To other "news stories of the week".
| 
