Previously, Armin Steinhoff wrote in qdn.public.qnxrtp.advocacy:
Tony Mantler wrote:
There is no reason to assume that a ‘shorter’ program is inherintly
better than a ‘longer’ program.
I didn’t say that a shorter program is a better program … I said
that ‘less code means less probabilities to make coding failures’.
Poor logic densities, either too much code for too little logic, or too little code for too much logic, will encourage programming errors. Either extreme will kill your project.
I know of very few proper languages (that is to say, excluding brainfuck and the like) if any that will forcibly push you into poor logic densities without the programmer first adding a nice heaping pile of incompetance to the mix.
Certainly some languages are more condusive to being fucked up by incompetance - I’d personally put C++ and Perl on the short list because of their ad-hoc fundamental designs.
But then, the trick is to not hire incompetant programmers, isn’t it?
Go read some IOCCC entries or perl one-liners and tell me that less code
makes a lower probability of coding failure. >
Hmm … with a failure rate of e.g. 10% it would mean for 100 lines
of code there are 10 errornous lines and for 1000 lines of code
there are probably 100 errornous lines. Something wrong with it ?
If you have a 10% failure rate per line, you’re probably in the wrong business.
I would say per-line failure rates are irrelevant, though, they almost always shake out pretty quickly - it’s easy to see where logic and code don’t match. The real killer bugs are the logic bugs, and the logical complexity of a program cannot be reduced by changing languages.
Anyways, it all boils down to this: no programming language can reduce the
complexity of the problem set that the program is being written to solve.
Correct … but a program language can reduce the complexity of the
coding. Do list processing with Python and do it with plain C >
A convinience of implementation.
It is, at most, a minor implementation convinience.
Saying different languages make better programs is like saying different
calculators make better mathematical theorems.
OK … then do all of your theorem proofs with a TURING machine …
BTW… here some convenient languages … have fun:
http://koeln.ccc.de/projekte/brainfuck/index-e.html
[…]
I seem to recall writing a caeser cipher in brainfuck once, dunno what I did with it though. Certainly it wouldn’t be quite on the level of the brainfuck quine in one page you linked to, but it was a neat trick to pull off none the less.
Cheers - Tony ‘Nicoya’ Mantler
–
Tony Mantler | Proud ---- Days since the last
QNX Consulting | of our | 27 |
tony@astra.mb.ca | Record ---- “Gerbil Incident”