kmagel@plains.UUCP (ken magel) (06/13/90)
AN earlier article mentioned the inadequacies of present CASE tools for
handling object-oriented systems. Several other correspondents have supported this view and argued that the needed new CASE tools to support object oriented
systems clearly will be very different from present tools. They argue that
structure charts, data flow diagrams, etc. will not have much use in object
oriented development methodologies.
I think these arguements are not correct. The CASE tools needed to support
object oriented systems and development methodologies will be only modestly
different from those needed to support structured analysis and design. Many
of the diagrams such as entity-relationship, data flow, and so on will need
only minor modifications to be extremely useful in designing, implementing and
understanding object oriented systems. The minor modifications will be needed,
but they should be doable well within the framework of existing CASE tools.
Some correspondents are arguing that object orientation will cause a
paradigm shift, a revolution in software development. I think object
orientation is merely an evolutionary step along the path of improved
software manipulation. It is not a radical change althoug it is a significant
change. The notion of fifty year revolutions overthrowing the old order
seems to overstate the real situation.cox@stpstn.UUCP (Brad Cox) (06/16/90)
In article <5012@plains.UUCP> kmagel@plains.UUCP (ken magel) writes: > Some correspondents are arguing that object orientation will cause a >paradigm shift, a revolution in software development. I think object >orientation is merely an evolutionary step along the path of improved >software manipulation. It is not a radical change althoug it is a significant >change. The notion of fifty year revolutions overthrowing the old order >seems to overstate the real situation. This misrepresents what I was trying to say. OO will not 'cause' a paradigm shift. The software crisis will do that. Technologies will meekly follow. The paradigm shift has nothing to do with the change from structured to OO languages. The change in technologies is only a symtom of a deeper cause, a paradigm shift from one in which the programmer's relationship with the machine is at the center, to one centered on the producer-consumer relationship; a marketplace in interchangeable (reusable) components. To express the change in different terms, the paradigm shift involves getting away from todays enfatuation with *languages* to begin concentrating on building a robust marketplace in *components*. There *is* a silver bullet that can solve the software crisis, in precisely the sense that the Ptolemaic astronomy crisis was solved by Copernicus' silver bullet. His silver bullet was none other than a paradigm shift; a shift in focus that put the sun at the center rather than the earth. With no offense to Dr. Brooks intended (which whom I intensely agree on most issues), I can readily imagine a Ptolemaic astronomer, frustrated at the difficulty of continually recomputing epicycles, expressing this frustration in an article titled "No Silver Bullet; Essence vs Accidents in Astrophysics". Had computers existed back then, I could even imagine exuberant discussion of how this new technology will "revolutionize" astrophysics, thus confusing a new *technology*, such as the computer in their case and OO in ours, with the *true* revolution, the Copernican paradigm shift. -- Brad Cox; cox@stepstone.com; CI$ 71230,647; 203 426 1875 The Stepstone Corporation; 75 Glen Road; Sandy Hook CT 06482