X-Authentication-Warning: delorie.com: mail set sender to djgpp-bounces using -f From: Hans-Bernhard Broeker Newsgroups: comp.os.msdos.djgpp Subject: Re: strange error Supersedes: <2m9jprFk1h11U1 AT uni-berlin DOT de> Date: 22 Jul 2004 10:52:03 GMT Lines: 56 Message-ID: <2m9kijFk1h11U4@uni-berlin.de> References: <20040722025105 DOT 05775 DOT 00002121 AT mb-m05 DOT aol DOT com> <20040722031649 DOT 06719 DOT 00001841 AT mb-m13 DOT aol DOT com> X-Trace: news.uni-berlin.de VTngbdZSSamStaeHKBP9UAztqdIeULZKq9Oj+9jycfOmb7INITL2hvHKBd X-Orig-Path: not-for-mail To: djgpp AT delorie DOT com DJ-Gateway: from newsgroup comp.os.msdos.djgpp Reply-To: djgpp AT delorie DOT com Sterten wrote: > I guess this is the usual way how programmers (and other scientists) > proceed. trial and error. You guess way incorrectly. Scientists use trial and error, because out there on the cutting edge, that's all they can do. But programmers worth their salt don't. Trial and error can never be a substitute for knowing how to do the job properly. Programming is, essentially a craft (ein "Handwerk"), with the primary focus on knowing your tools, and how to use them well. > Usually it's easier to try out a doubtful command by writing a > test-program rather than consulting the manual/book, which often > even has some errors. So you consider the likelihood that a test program you never showed anyone, its design based mainly on pure speculation, will have errors to be significantly smaller than that of the information entered in the manuals being correct, after having been read and used by at least dozens of people? That's hybris. Yes, errors happen, but that's no grounds to conclude that all manuals are always wrong, which is what you're essentially doing up there. Even bad manuals are still much more likely to be correct than your method. > the disadvantage is, that books are usually _huge_. Not the one I think I recommended to you a lot earlier: K&R2. It's short, to the point, and will teach you more about C in a day of earnest work than your current method will ever manage to in a year. And it's available in a very usable German edition, too, which may make things easier for you. > Basically, I only need a dozen or so C-commands. > And for each of these the corresponding assembly routine. Nonsense. You need the corresponding assembly routines for exactly *none* of them. If you want to program in assembler, fine, do so, but don't do it in C. > you have to choose where being ignorant and on what to concentrate. > Only very few people will specialize in C. We're not talking about specializing in C here. We're talking about getting to know the tools you'll be using to at least the level where you don't endanger yourself and innocent bystanders. And about the willingness to accept the fact that if you don't specialize in a given tool, it's your moralic duty to accept advice from those that do at face value, not question it endlessly. -- Hans-Bernhard Broeker (broeker AT physik DOT rwth-aachen DOT de) Even if all the snow were burnt, ashes would remain.