Re: Danger: Derrida at work
From: Edward G. Nilges (spinoza1111_at_yahoo.com)
Date: 12/15/03
- Next message: Bill Godfrey: "Re: Any experience with "The Last One"?"
- Previous message: gswork: "Re: Topicality of the programming profession."
- In reply to: Brian {Hamilton Kelly}: "Re: Danger: Derrida at work"
- Next in thread: Edward G. Nilges: "Re: Danger: Derrida at work"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: 15 Dec 2003 01:25:50 -0800
bhk@dsl.co.uk (Brian {Hamilton Kelly}) wrote in message news:<1071418903snz@dsl.co.uk>...
> In article <f5dda427.0312121944.4fc182a3@posting.google.com>
> spinoza1111@yahoo.com "Edward G. Nilges" writes:
>
> > And, brains in vats may think they can debug a port without a status
> > report.
> >
> > I did so working ninety hours a week since no time had been allowed in
> > the schedule because I knew that the system is inescapably its
> > manifestation to the reader of the status report, and this in turn was
> > a lesson I learned, not only from Derrida but also from the humanities
> > in general.
> >
> > I'd learned never to "privilege" the "hard" or male communication with
> > the serial port as somehow "more important" than its presentation, and
>
> DEK recognized these failings of ordinary programming at least twenty
> years ago, which is why his original paper[1] on "Literate Programming"
> used as one of its examples the concept of a routine to capture data from
> a human operating some input device. As he stated there, if one does a
> "proper job" of all the validity checking that _needs_ to be present,
> then that routine comes out looking as if its sole purpose is error
> checking, and _not_ the input of data. Hence most programmers will tend
> to skimp upon the code for the validity checks (with potentially serious
> consequences). By splitting off the section @<Check the data input for
> validity> into a separate LP section, the programmer can do a good job of
> _both_ tasks.
"Separation of concerns" isn't a "typically male" strategy, and it
therefore is not "feminist" to write a single main() routine :-).
EMOTIONAL separation of concerns is a male issue.
I agree that readTheInput() and checkTheInput() could very well be
separate routines in a well-structured program independent of feminist
theory. In fact, if the program is reading a communications line it
should "layer" readTheInput() inside checkTheInput() such that
readTheInput "covenants" (in Mormon terms) to provide bytes, while
checkTheInput "covenants" to provide valid characters only.
But what a "typically male" programmer would do, would be to create
ersatz meeting-of-the-covenant by claiming somehow that readTheInput
was "enough" and that checkTheInput was a "frill".
Whereas in my experience women programmers don't do this.
>
> > I note that commercially viable systems, such as National Instruments
> > LabView, are always shipped with comprehensive views of communication.
>
> Hmm; YMOV. When I used LabView with a piece of IEE-488-based test
> equipment costing hundreds of thousands of pounds, it seemed to be very
> cavalier about bothering to check whether it was actually talking to the
> DuT via the 488-bus.
>
Pretty scarey since I have to rely upon it. It works for me now and
I've added code to view the operation in detail.
I'm not a comm specialist but more than once have been responsible for
low level comm. What success I've had was constituted in a principled
"phenomenological" refusal to accept that the communications was
working without a text that shows how it works, because as we see in
Derrida, there is nothing but the text.
I remain less impressed by hardware that "works" and more by texts
associated with visibly working hardware that show when and how.
"Phenomenology" is BTW a philosophical movement with which Derrida was
associated early in his career; Derrida's graduate thesis was on
Husserl's analysis of the truth of Euclidian geometry. The Euclidian
arguments, as I think both Derrida and Husserl realized, depend in an
unspoken way on the material circumstances of proof.
Derrida saw mathematicians as claiming that they "know" geometry to be
true strictly by virtue of abstract and formal arguments only
supplemented by demonstrations (whether Euclid drawing triangles in
the sand, or Mathematica demonstrations) and Derrida showed how this
knowledge DOES NOT EXIST outside the context of presentation.
He in fact demonstrated how in practical applied mathematics,
excessive abstraction constituted in the belief, inherited from the
very different subject of pure mathematics, that we would be better
MEN if only we could be ever more abstract and thereby freed from the
mode-of-demonstration.
In computing this is the subconscious bias against presentation logic
and its isolation in a hypostatized GUI interface, for which "low
level" Visual Basic programmers are retained to create crude
interfaces (which conceal more than they reveal) because the
supplemental issue of presentation is not the male term...for the same
reason that mathematicians can't teach geometry to real students,
having lost Euclid's connection to the way in which the geometry can't
live outside of its presentation.
Neither Derrida nor Husserl ever once argued that there is "no such
thing" as the abstraction. Instead their ontology insists I think that
the abstract level is like a virus as opposed to a bacteria.
A virus like the AIDs virus needs to find a host system whereas
bacteria don't, and by analogy, the abstractions are for Derrida and
for Husserl always, and by necessity over and above science (a
PHILOSOPHICAL necessity) encompassed in a physical GUI...whether a
Visual Basic screen or marks in the sand.
I believe now that the subconscious bias makes MIS managers
over-isolate "problems" as in the case where the original developer is
in fact the best person to create the GUI but cannot do so because
it's not considered his "specialty". Perhaps this is the LabView issue
you've identified: I don't know. I do know that despite the high
quality in Visual Studio .Net, its debug facility are less useful than
they might be whenever you use multiple threads and on my theory, this
would be due to the separation, not of concerns, but of
teams...between .Net developers aware of how central multithreading
has become in "real" .Net, and GUI developers who might still be in VB
6 thinking, in part.
Long term I would like very much (once my existing publishing
committments are under better control) to write a bio of Edsger
Dijkstra modeled on A Beautiful Mind. It is probably too much to hope
that Ron Howard will make a movie out of such a book, with Paul
Neumann as Dijkstra and even if he does the writer doesn't get the
chicks anyway, in Hollywood.
Seriously, I think Dijkstra brought an entire world view based on
Continental philosophy that was completely strange to ACM honchos
whose education in American philosophy, not to mention any other form,
was nonexistent and whose philosophical high-flights seem to consist
in (1) John McCarthy's neocon ravings and (2) Knuth's gentler and more
attractive Lutheran obsession with the Bible. I don't think Dijkstra
himself realized the cultural gap and was taken in by America's rather
self-serving narrative that we welcome all sorts of different points
of view.
Comments welcome.
> [1] I've always been intrigued as to why that paper was published in the
> British Computer Society's /Computer Journal/, whereas he had hitherto
> published in various ACM journals. Were the ACM peer review panels too
> blinkered to recognize the worth of Literate Programming?
- Next message: Bill Godfrey: "Re: Any experience with "The Last One"?"
- Previous message: gswork: "Re: Topicality of the programming profession."
- In reply to: Brian {Hamilton Kelly}: "Re: Danger: Derrida at work"
- Next in thread: Edward G. Nilges: "Re: Danger: Derrida at work"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
