Re: misconceptions on computer science

From: Shayne Wissler (thalesNOSPAM000_at_yahoo.com)
Date: 08/06/04 

Date: Fri, 06 Aug 2004 15:50:31 GMT

"Alan Gauld" <alan.gauld@btinternet.com> wrote in message
news:gf85h01vdpo5vd8ko88bnpmc267glo5rdj@4ax.com...
> On Thu, 05 Aug 2004 15:28:01 GMT, "Shayne Wissler"
> <thalesNOSPAM000@yahoo.com> wrote:
> > > Again I disagree. For example I would argue that CS probably
> > > doesn't need to spend much time on testing, but an SE degree
> > > would have, I hope, at least an entire semester devoted to it,
> > > if not more.
> >
> > And this furthers your position how?
>
> Basically what I'm saying is that testing is an essential skill
> for someone who needs to write good quality code. It is not an
> essential skill for someone who only needs to write code that
> works most of the time. CS students often write quite buggy
> code(*) but it is adequate to demonstrate some aspect of
> computing that they are interested in. An SE is expected to write
> code that works correctly at least most of the time! Thus a CS
> course is not merely about writing code. (Although granted the
> aspect oif computing should ultimately result in something from
> which another mayt write some code - but that other could be an
> SE...)

I never said CS is merely about writing code. I said that CS exists in order
that code be written at some point. That does not imply that the particular
individual doing the research about a particular algorithm necessarily needs
to do the coding for it.

<snip>

> > > Of course there is virtue in algorithms for their own sake, just
> > > for the art and beauty of their form.
> >
> > Show me an algorithm that is beautiful but useless.
>
> I'm not a scientist, I'm an engineer. All the algorithms I know
> off the top of my head are useful! But I do recall seeing work
> to produce a sorting algorithm that would be as inefficient as
> possible but still guarantee a result. Now maybe somebody can
> find a use for such a thing but the researchers weren't expecting
> a queue to form anytime soon. They were just intrigued to see how
> inefficient they could go - and to see if anything interesting
> came out of their efforts. I don't know the outcome but I know
> they spent over 6 months working on it (a team of 3 I think?)

Show me an algorithm that is beautiful but useless.

> > > No they are doing pure science. How they get it funded is another
> > > matter, but given a suitable benefactor the research is perfectly
> > > legitimate science.
> >
> > If you want to further your position you're going to have to come up
with
> > examples.
>
> I thought I'd given several.

You haven't given a single example that furthers your position.

> > > I'm thinking about people like Turing and his contemporaries
> > > who were trying to think about what would happen *if* they hasd
> > > such a machine. (Turing of course was a mathematician and would
> > > never have claimed to be a scientist, but his work falls into wat
> > > today would be considered CS).
> >
> > What specific thing that Turing did are you talking about? And why do
you
> > think that engineers can't think about the implications of a machine
they
> > might construct?
>
> The obvious one is the Turing test. It is a thought experiment.
> Turing never seriously tried to build a machine that would pass
> the test. I'm not even sure he seriously expected one to be
> built.

Oh, you mean the test where if you can fool a computer scientist into
thinking that a program is a real person, then you've succeeded in building
AI. Well by that standard I don't think it's very hard to build AI.

> > > And although Turing et al eventually got to try to build their
> > > machine it took a war to provide the funding. They had been
> > > working on the concepts long before that.
> >
> > Non sequitor. Babbage worked on his machine's design long before he got
the
> > funding too.
>
> But he was always intending to build it. Turing and his
> colleagues never really expected to build their dream machnines.
> Even in the war what they built was very specific and not very
> much like the machines they were dreaming of.

Of course not. That's because Turing was quite confused regarding AI.

> > > No, they are researching the future. Maybe someday we will have
> > > such machines. Indeed much of what they envisaged in the 70's has
> > > come to pass. This is the nature of science.
> >
> > Examples please.
>
> Fingerprint/Iris recognition systems, natural speech recognition
> real-time image manipulation, virtual reality systems.

How exactly do these systems connect to computer science as a science not as
engineering? Because those sound like engineering projects to me. And just
because someone with the title "computer scientist" thought them up does not
indicate that he was doing science not engineering. A lot of CS guys are
doing proper SE--that furthers my case not yours.

Shayne Wissler

Relevant Pages

  • Re: No Set Contains Every Computable Natural (was Church-Turing compared to Zuse-Fredkin thesis)
    ... Russell is right about Turing saying this, ... You might not want to call this an algorithm which has ... decimal representation of x digit by digit. ...
    (comp.theory)
  • Re: Quantum computer using using artificial atoms.
    ... >> get a quantum polynomial time algorithm to run in polynomial time on ... > Factoring has never been proven to be a hard problem, ... Every important factoring algorithm ... > others who are trying to push the idea that quantum Turing is different ...
    (sci.crypt)
  • Re: Turing Machines and Physical Computation
    ... rather than adumbrated in that 1936 Turing paper On Computable ... ... any actual computer has finite memory, ... computer will run out of memory for some huge calculation. ... It is generally assumed that an algorithm must satisfy the following ...
    (sci.math)
  • Re: Turing Machines and Physical Computation
    ... rather than adumbrated in that 1936 Turing paper On Computable ... ... any actual computer has finite memory, ... computer will run out of memory for some huge calculation. ... It is generally assumed that an algorithm must satisfy the following ...
    (comp.theory)
  • ALENEX04 accepted papers
    ... Engineering a Cache-Oblivious Sorting Algorithm ... Faisal N. Abu-Khzam, Rebecca L. Collins, Michael R. Fellows and Michael ... A New Approach to Shortest Path Algorithms Optimized ...
    (comp.theory)