Re: Programming skills required by mathematicians?

On 4 Jul, 20:34, Marshall <marshall.spi...@[EMAIL PROTECTED]
> wrote:
> On Jul 4, 11:55 am, goto <g...@[EMAIL PROTECTED]
> wrote:
>
> > Frederick Williams wrote:
>
> > > > I think nowadays nobody, except Herman, will argue in favour of
got=
os.
>
> > > There are common enough situations in which the goto is the clearest
> > > thing to use, so why wouldn't one use it?
>
> > The principal objectioneers always were the academic airheads, those
wh=
o
> > never wrote or imagined a real life program larger than a dozen lines.
> > Other than that,
>
> > =A0 =A0 =A0 =A0 "If you want to go somewhere, goto is the best way to
g=
et there."
>
> Snort.
>
> This was a fierce argument in the 1970s. Your side lost.


That it lost doesn't necessarily mean it was wrong. Statistics on
software projects' failures anyone?

We might need make a distinction between LLL and HLL as different
degrees from machine language up to -say- scripting, and LLL and HLL
as two fundamentally distinct domains, where the higher LLL just
happens to be near to the lowest HLL, maybe due a common formal
background.

For instance, structured analysis was indeed a great step ahead (and
still it is the best reference, IMO): in the process of mapping the
problem to the conceptual domain and then down to the software modules
and concrete data structures. Still, at a certain point you need get
your hands dirty in coding: there is a discontinuity.

IMO, today's higher level programming languages, the ones actually
used in the software industry I mean, mainly suffer of being hybrids
between the conceptual and the technical domains. Not even like a RISC
vs. CISC debate, which is rather around the number of primitives as to
the implied architecture.

-LV


>
> Marshall