Engineering I.E. Quality
James Richard Tyrer
tyrerj at acm.org
Wed Apr 7 23:08:20 CEST 2004
Perhaps my previous posting was not clear or was not really a good example:
http://dot.kde.org/1081204077/1081219687/1081261910/1081317231/1081328080/#1081369703
<<
"Proper engineering criteria" is such a vague thing when it comes to
programming. The perfect "abstract" design doesn't exist, only
implementations can be seriously evaluated, and programmers have the last word.
>>
To an engineer, it seems hard to imagine that he could be more wrong about
so many thing is such a short posting. But, I presume that this is what
naturally comes from a programing centric Weltanschauung.
First, for those that don't know, the idea of a "perfect" design has
nothing to do with engineering. In fact the application of engineering
criteria makes the a priori assumption that there is no perfect design.
Engineering is always about optimization -- adjusting conflicting
parameters to produce a design that best meets the criteria. Which is why
you MUST have criteria to design something.
To say that a design can only be evaluated after it is implemented is
absurd. Try applying that notion to a bridge. Yes, you will have issues
when you implement a design, but that does not mean that you shouldn't
start with the design and then implement it.
But it is really the attitude that: "programmers have the last word" --
offered as a decree with no supporting argument -- that is the real issue.
They just don't get it. It is the design of an application that is
important -- at least as important as writing the actual code. Writing the
actual code is just the implementation. Until all of the software
engineers (developers) in this project come to realize this, the ultimate
progress of this project is limited.
With this attitude prevalent among developers, it is going to be us against
them, and I don't want any more of that.
--
JRT
More information about the kde-quality
mailing list