The Business of Software
million. Given that today’s Xbox
can hit 6,400 MIPS it should
therefore cost around $93 million.
Clearly, Moore’s Law has trumped
Grosch’s Law.
Readability. In a study on the
readability of major computing
periodicals, CACM scored in the
“Difficult” range of the Flesch
Reading Ease Index. It scored
higher (more readable) than any of
its major competitors at the time—
the most difficult to read being the
IBM Systems Journal, which was
rated as solidly in the “Very Difficult” category. Of course, some elements in our profession believe that
if something is relatively easy to
read it cannot be worthwhile.
However, when the job of a magazine is to communicate it would
seem that any effort put into making that communication more efficient would be valuable. I think
CACM continues to invest heavily
in this area. Certainly these days it
is much easier to read than some of
the 1985 benchmarks such as-then-number-one in circulation
Infosys-tems, number two in circulation
Datamation, 1 number three in circulation Computer Decisions and
the much-missed number four
Mini-Micro Systems. Clearly there
has been a Darwinian process at
work.
Niklaus Wirth won the ACM
A.M. Turing Award in 1984. In
his address published in the February 1985 issue, he discusses his
pursuit of an “appropriate formalism” in programming languages
1
Datamation still exists, though not in paper form; see
www.datamation.com.
and describes some of the common characteristics of the projects
that informed his view of languages at the time. They included
the idea of a bootstrap tool that
provided capability but more
importantly acted as the basis for
the next generation or iteration.
He identified the need to separate
requirements and capabilities into
the essential and the “nice to
have.” He thought the choice of
tools important but that they must
not require more effort to learn
than they save. And, most importantly, he viewed each project primarily as a learning experience
supported by the most essential
(though “elusive and subtle”) element of an enthusiastic team that
collectively believes in the worth
of the endeavor. Well that hasn’t
changed.
Selecting Decision Support
System (DSS) Software was a
cogent and useful article on how
to select DSS including, if I perceived correctly, screen shots
from Lotus 1-2-3. The striking
thing is that none of these systems exist now, though some
have probably evolved into
today’s offerings. The evaluation
criteria with a few deletions (such
as the ability to produce “basic
plots and charts”) could serve as
a starting point for an assessment
capability today. Curiously, none
of them mention service-oriented
architecture or even Web access.
The article really addresses a
process for selection rather than
what is being selected. While the
target software has changed considerably, much of the process for
selecting it could still apply. This
just goes to show that software
and systems come and go, but
process lives forever.
Programmer Productivity.
Starts off with a complaint about
the dearth of empirical research
in the kinds of tools that really
help programmers. The list of
software tools desired by programmers in 1985 shows how far
we’ve come in this area (screen
editor anyone?)
Event Simulation Modeling
Language. With 56 cited articles,
this article on a condition specification (CS) language as part of a
model specification (MS) language could have been written
yesterday. It doesn’t seem that this
aspect of software engineering and
computer science has advanced as
it should given such a start.
Efficiency of List Update and
Paging Rules. This article alone
must have pulled the magazine’s
Flesch Reading Ease Index down
by a good 20 points. Well, it was
in the “Research Contributions”
section, so that would explain the
six theorems and their associated
proofs embedded in the article.
While not everyone’s cup of ∑ti
it showed the breadth of topics in
this magazine and might well
have appealed to those also interested in the…
...Positions Vacant in Computer Science. I counted 186
positions in computer science
and software engineering in universities and colleges around the
world (but particularly around
the U.S.), 31 calls for papers just
for February 1985–May 1985,
14