mpeuser at web.de
Sat Sep 6 17:01:34 CEST 2003
"John J. Lee" <jjl at pobox.com> schrieb im Newsbeitrag
news:87vfs6unsx.fsf at pobox.com...
> "Michael Peuser" <mpeuser at web.de> writes:
> Windows is an esoteric OS?
Counted by numbers, the widest spread OS is an endless loop, interrupted by
hardware timer interrupts. About 50 billion I should say, but can be much
more. There is one on your video card and in your hard disk and even inside
your Compact Flash (CF) card. Then come some real proprietary OS. And -
granted - there are also some Linux and Windows boxes running. This shifts
to more standardisation (VxWorks, QNX embedded Linux,...) as electronic
equipment becomes more intelligent (DVD Player e.g) A modern limousine
contains about 50 microcontrolles with independant operating systems.
> Still, you make an interesting point here: is it the case that more
> code is written for hard real-time systems than not? That would
> certainly surprise me. Where do you get your data?
I am still looking for something traceable. My information is from some
diagrams in seminare hand-outs from end of the 90. So they probably come
from data of the early 90. They say there are twice as many commercial
programmers working in technical (embedded systems) than in administrative
programming. Since then there has been have a lot of web programming which
was non-existent 10 years ago, but we also have a lot of telecommunication.
> > May be they will be programmed in Python in 20 years, though I do not
> > it. You need a certified compiler, a fail proof and predictable garbage
> > collector, and a little bit more.
> Yes. And? We seem not to be communicating here...
That is my fault. You are proably thinking of the language as such. I was
thinking of its standard implementation as open system with hundreds of more
or less mature add-ons
> Sorry, I don't quite follow that. How does that relate to the rest of
> the paragraph?
(There were some mixed threads here, I was ansering as well to some Linux
> > > Python is an order of magnitude easier to learn than C++, but it
> > > brings really significant programmer-efficiency benefits. As a
> > > result, the argument about training simply doesn't stand up: the costs
> > > are lower than the benefits even on a short timescale.
> > I know this argument well and have repeated it myself in the days of
> > and Algol68. Few have listened.....
> Well, they'd be right not to listen if you argued it about Pascal.
> There's a real difference in productivity between Pascal and Python.
> Pascal doesn't even have built-in garbage collection, for heaven's
> sake, which places it much closer to C than to Python in this respect.
I said "in the time of Pascal and Algol 68". If you kindly would consult a
calender you will find out that had been 35 years ago. Pascal had been
designed to not need a garbage collector which was a great issue for Algol68
with its (explicit) heap concept. Some of the first C++ compilers 20 years
later were notorious for their never working GC. There are companies still
not using C++ for embedded aplication because they mistrust even modern
> > Programmers are generally not trained in their company, they are hired
> > including their special language skills.
> You're begging the question: arguably, that is the mistake of those
> hiring. Training (for Python, anyway) will frequently be cheaper than
> sticking to, say, C++. Other skills are more important than
> pre-existing ability to use a language like Python (though that's
> certainly not irrelevant!).
> > [..]
> > > > This is an important factor. However all investigations show that
> > > > programmer's productivity is an unmeasurable quantity.
> > >
> > > How would an empirical investigation show such a thing?
> > Alex gave a hint to a German study. I think he was refering to:
> > http://www.ipd.uka.de/~prechelt/Biblio/jccpprtTR.pdf
> (That was probably the one I was thinking of in another paragraph in
> my reply) I don't see how such a study could show that "programmer's
> productivity is an unmeasurable quantity". It's obvious that
> variations between individuals make differences due to languages hard
> to measure, but that doesn't make it impossible.
The study was meant (and shows) that using scripting languages for special
projects is advantegeous over using standard languages. It also shows
the"productivity" of the (wide range of ) programmers. I will have to look
into it again ....
... yes take page 22: Total times of programming: The JAVA programers needed
between 3 and 63 hours. This is nothing you can work with when it comes to
estimating the costs for your next project. On the other hand it shows that
there are some Python programmers who delivered their task much faster. So
what? Everyone knows that Python programmers are the better, faster, busier
This is not a general distrust for statistics (I studied mathematics) but
also my personal experience from 30 years of programming. I do not remember
who (Dijkstra?) once said: "When a programmer is good he is generally very
good, and when he is bad he is generally very bad". But all of them want to
earn their money in the software business.
More information about the Python-list