Information Technology Reference
In-Depth Information
theoretical foundations and disciplines of daily practice that one found in traditional fields
of engineering."
Software, it was argued, should be created under rubrics as rigidly defined as those by
which civil engineers designed and built bridges, and architects skyscrapers. Likewise, pro-
grammers should be officially qualified and certified through special training and testing.
Nevertheless, as we shall see, all the best software would inevitably be created by program-
mers and designers thinking out of the box, adhering only to the most fundamental para-
meters enunciated by the likes of Knuth. Over time, the formal software engineering pro-
cess was to result in only the most mundane, all-be-they necessary, application solutions.
Throughout most of the sixties, software (and therefore software development) was
largely machine (and provider) specific. For example, IBM developed programs meant ex-
plicitly for the 700/7000 series which shipped with the machines as part of IBM's sales/
service package. This changed in December 1968 when, in response to antitrust concerns,
IBM announced its intention to unbundle its software - in other words, to sell/license soft-
ware applications separate from hardware sales, rather than forcing the sales of both hard-
ware and software as a unit. The most popular of these applications proved to be IBM's
Customer Information Control System (CICS), which IBM offered as a license at $7,200
per year.
Now, for the first time, there crystallized the idea of "software" as a separate marketable
item (and as a sales opportunity ripe for exploitation by organizations not involved in the
design and manufacture of hardware). Thus, from the very late 1960s and through the
1970s, third-party mainframe software and systems developers were to play a pivotal role
in shaping the business computing environment. Mainframes produced by IBM, especially
the 360, held vast potential for applications - in fact far more potential than IBM's in-house
programming personnel could effectively exploit. This fact created an enormous opportun-
ity for third-party contractors. Indeed, much of the very best software, systems and imple-
mentations for these and other machines would henceforth be developed by outside parties.
In this environment, such firms as American Management Systems (AMS), H. Ross
Perot's Electronic Data Systems (EDS), Systems Development Corporation (SDC) and
Thompson-Ramo-Woldridge (TRW), not to mention the federally funded SDC and MITRE
(both having their roots in the development of the SAGE air defense system in collabora-
tion with the RAND Corporation) rose meteorically in prominence.
Fortuitously, this same time period saw the creation of elegant new programming lan-
guages with which software for the new pipeline could be created.
Save for FORTRAN and COBOL - each of which remain in use today after many re-
visions - most other languages developed as of 1968 had proven failures in the creation
of business applications. Conspicuous among these failures was ALGOL (short for Al-
gorithmic Language), pushed by the Burroughs Corporation, which in most people's view
was overly complicated in its final iteration. Too lean with its first release in 1960, the lan-
