home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.org.eff.talk
- Path: sparky!uunet!cs.utexas.edu!swrinde!emory!rsiatl!jgd
- From: jgd@dixie.com (John De Armond)
- Subject: Re: Software as PE
- Message-ID: <ftsr94#@dixie.com>
- Date: Sat, 02 Jan 93 00:10:35 GMT
- Organization: Dixie Communications Public Access. The Mouth of the South.
- Distribution: na
- References: <46.2b2f8ada@ivgate> <bhayden.724605662@teal> <889520164DN5.61R@tanda.isis.org> <1992Dec30.125324.27900@mksol.dseg.ti.com> <522322457DN5.61R@tanda.isis.org>
- Lines: 104
-
- marc@tanda.isis.org (Marc Thibault) writes:
-
- > Fred J asked me to define a real engineer. Here's my shot at
-
- Very excellent analysis. Couple of additional thoughts.
-
-
- > Software development methods are anything but time-tested;
- > they are changing as we speak and none has been consistently
- > successful at turning out software that is both reliable and
- > efficient. The (relatively) rare program that meets those
- > criteria is inevitably the result of intensive hacking by a
- > master craftsman whose respect for the paradigm-du-jour would
- > be infinitesimal if he or she cared what it was. Nor is the
- > behavior of the components well understood. That's why we have
- > bugs while engineers make design errors. We can't engineer
- > software because we haven't yet figured out _how_. Calling
- > programmers software engineers is at best wishful thinking,
- > and wishing only makes it so in Disneyland.
-
- I don't believe software creation will ever become an engineering
- discipline any more than the creation of oil paintings will.
- Both are inextricably linked to the craftsman's artistic intellect.
-
- I'm one of those people who is educated in a conventional engineering
- discipline and got into software after about 15 years of conventional
- practice. I was able to watch the evolution of my brand of engineering
- (nuclear instrumentation) move from the sliderule era through the
- calculator and into the computer age. At the same time I am a (budding)
- artist, working with hot glass. I can therefore observe both sides.
-
- Consider a few things. First, languages. It is intrinsicly difficult to
- move between programming languages. Some people are gifted in this
- area, of course, but for most people, learning a new programming language
- is quite a chore. That is because each language embodies a different
- way of approaching the creative process. Assembler programmers tend
- to do things quite differently than SQL programmers. This is at odds with
- the conventional engineering disciplines where the same methodologies
- are used throughout. I found little difficulty, for example, on
- picking up on designing hydraulic systems. A little bit of reading
- and some study to find out what the standards were and away I went.
-
- Consider artistic media, on the other hand. An oil painting artist does
- NOT easily move into, say, stained glass. A wood carver cannot easily
- become a painter. This is quite in parallel with programming. While
- a central artistic ability has to be present in both cases, one's skill set
- may only match certain media. I personally have tried for years to find
- a media in which to express some of my artistic ideas. Hot glass appears
- to be it. I tried drawing and cartooning and painting and sucked bilgewater
- at them all. Woodcrafts are not much better. Yet I find working with
- glass easy and people tell me my work is quite good.
-
- Same with software. As a development manager, I've found most programmers
- excel in one or two areas. I have little patience with end-users and
- therefore do poorly with applications. Communications, on the other hand,
- seemed a natural.
-
- > I don't think this is forever. Wide adoption of standard
- > objects will, in time, give the craft of programming a leg up
- > on the way to engineering.
-
- I don't much think so. I suspect it will further the development of a
- technician/craftsman class of programmers analogous to the carpenter
- or electrician. These people will bolt together variations of standardized
- applications just as carpenters hammer togethers pretty standard structures
- from "objects" of wood (the plank.) In this niche may be the place for
- the licensed engineer to oversee the design of these things when the end
- result has public safety implications.
-
- This brings up another difference between engineers and software development.
- Most any engineered system can be proved correct or else the weaknesses
- can be comprehensively identified. Provably correct designs have
- been a big thing in the nuclear biz for the last 20 years or so. It is
- this concept that has kept computers out of the reactor control and
- protection loop until recently and even then, the impossibility of provably
- correct software systems is addressed by massively redundant systems AND
- by keeping the human operator in the loop.
-
- > No electronic engineer has ever advertised himself as expert
- > only with the "Apex Model 5 Logic Analyser". Anyone who
- > studied to be an electrical engineer with me can, in
- > addition to engineering electrical systems, design a logic
- > circuit or machine tool, program any computer, build a barn or
- > a cannon, distinguish between a terminal moraine and a
- > land-fill site, survey a meadow, and navigate a ship.
-
- Yup, and that is in stark contrast with many "CS" programs that
- seldom venture out to anything outside the computer.
-
- I think the titles "science" and "engineering" get tagged onto what
- programmers because a) academians realize the engineering schools get
- all the money and thus want the same legitimacy as, say, the school
- of nuclear engineering, b) because corporate managers realize the same
- thing applies in business and c) upper management could not stand the
- idea of paying big bucks to a bunch of people who are identified as
- artists.
-
- John
- --
- John De Armond, WD4OQC |Interested in high performance mobility?
- Performance Engineering Magazine(TM) | Interested in high tech and computers?
- Marietta, Ga | Send ur snail-mail address to
- jgd@dixie.com | perform@dixie.com for a free sample mag
- Need Usenet public Access in Atlanta? Write Me for info on Dixie.com.
-