XPost: alt.os.development   
   From: Ekrem.Saban@utanet.at   
      
   > I think that having good knowledge and practical skills of tool use is a   
   > very   
   > major requirement.   
   >   
   Unfortunately, those days are over.   
      
   I can remember the days where when you had spoken the sentence "I know C",   
   the director would embrace you with the words "Welcome to our company!!!"   
   Today, to say that you know C++ is just like you were saying that you can   
   are affluent in Russian. It is necessary and good, but it is not enough to   
   be employed. They would ask a lot of questions after that, like do you know   
   MFC? Have you ever programmed under AIX V.5 with the network such and such   
   for a bank? What did you do (if yes), otherwise a lot of specialised   
   questions...   
   >> An unrealistic way would be to throw away all the "coders" & employ   
   >> specialists.   
   >   
   > Realistic. In some projects, the coder's stupidity is tolerable to some   
   > degree.   
   > In some - it is _not_ so at all, so even the coders must be specialists.   
   > An OS   
   > and embedded software is second thing, the website is first one (it is   
   > developed using the tools which are very forgiving in terms of the coder's   
   > stupidity, C is not).   
   >   
   >>Then, the program would be well, even if it is programmed in C,   
   >> even if you use machine language! But then either   
   >> (a) nobody will buy it either, as it is too expensive, or   
   >   
   > Then how Microsoft, Oracle, Symantec, Google and others do live? If   
   > Siemens's   
   > software department could not become one of these names - then this is   
   > definitely a management issue in Siemens (possibly they are unwilling to   
   > pay   
   > proper salaries and want to make good things using cheap labour - then   
   > sorry,   
   > this can easily fail in some scopes of activity).   
   >   
   >> Nobody will accept that a plain programmer must be an expert.   
   >   
   > He must not be sub-qualified for his exact task, this is the rule. For   
   > some   
   > tasks, experts are needed. For some, they are not. Coding embedded systems   
   > in C   
   > is nearly and expert task, a task for a guy which will be qualified enough   
   > to   
   > understand pointers well, and not forgetting to call free() after   
   > malloc().   
   >   
   > You cannot do complex tasks using cheap labour, period. And no Ada or   
   > other   
   > similar complex languages will help. The cheap labour force will just   
   > misunderstand them :-)   
   >   
   >> expensive for me. I prefer to communicate with humans over a high-level   
   >> protocol.   
   >   
   > Me too :-) especially at work.   
   >   
   >> Maxim, if you want to, you can see the problem from this way: stop to   
   >> think   
   >> that you can CHANGE the way people are thinking!   
   >   
   > What I see from your description is sub-qualified (for this task)   
   > developers   
   > making constant errors in their C (or C++-near-to-C) code, errors related   
   > to   
   > pointer use. Sorry, but how Ada or Eiffel can help them? Such guys will   
   > just   
   > not be able to comprehend these languages properly.   
   >   
   >>YOU have to find the proper language for the people.   
   >   
   > ...or proper people, if we are about work.   
   >   
   >> way they approach the problem can be described with one word the best:   
   >> CHAOS.   
   >>   
   >> :)))))   
   >>   
   >> It is not true for ALL of the SW designers, but it is unfortunately true   
   >> for   
   >> MOST of them...   
   >   
   > Maybe. This is one of the cause the software professionals are well-paid -   
   > the   
   > company usually can hardly afford the personnel flow.   
   >   
   >> Oh no! You will get NO memory problems with a real language, like JAVA,   
   >> say.   
   >   
   > You will get another problems, like "this framework-provided class throws   
   > this   
   > stupid exception sometimes, why is it?".   
   >   
   > But for many tasks Java and .NET are fine and better then C, this is for   
   > sure.   
   >   
   >> of lines, written in a low- or intermediate-level language, you have   
   >> something "too good to throw away, but not good enough to be used   
   >> properly"!   
   >> This is worse than having a solution that do not apply at all. Because in   
   >> such a case, you find someone else who will rewrite the whole again.   
   >   
   > Depends on whether the code is comprehendable or not so. Sometimes it is   
   > really   
   > easier to re-implement something instead of comprehending the existing   
   > implementation.   
   >   
   > And now - the paradox - C++ code is often lesser comprehendable then C.   
   > For   
   > instance, the "a = b" for objects is lesser comprehendable then "Copy(&a,   
   > &b)".   
   > The reason is that with C++ "a = b" you have no copy constructor code at   
   > hand,   
   > and will need "grep" to find it somewhere in the header files, possibly of   
   > some   
   > base class.   
   >   
   >> They are not so evil as pointers, as to my knowledge - please correct me   
   >> if   
   >> I am wrong -, references never refer to NOWHERE.   
   >   
   > So what? The issue with deref of the NULL pointer is very, very rare, it   
   > is not   
   > an issue in real world.   
   >   
   >> So what is the realistic answer to THIS problem?? Quitting the SW   
   >> industry?   
   >   
   > For individual - educating yourself. At least you must comprehend your   
   > tool   
   > very well, also you must comprehend your platform very well. You can have   
   > bad   
   > comprehension of some feature group of your tool, but then - sorry - do   
   > not use   
   > them in your code at all, or take an effort and study them. First -   
   > comprehend.   
   > Then - design or code. This must be the law.   
   >   
   > For a company - proper hiring policy.   
   >   
   >> cost-efficient, then tell it to us! OK, getting personnel from China at a   
   >> cost for 1/10th.   
   >   
   > They are usually very poor specialists. Indians are often better, but a   
   > good   
   > Indian guy will cost the same as good Austrian guy (well, at least the   
   > same as   
   > good Slovenian guy).   
   >   
   >> that. So, why not use the so much increased and now levelled CPU power to   
   >> define high-level languages?   
   >   
   > They are more complex, and will require the good brains just to comprehend   
   > them. Otherwise, people will do some other stupid errors, just due to   
   > mis-using   
   > the complex language features.   
   >   
   >> By the way: do you ever used PURIFY under UNIX to see the memory leaks?   
   >   
   > I use PREfast on Windows, the static code analyzer for C provided by MS.   
   > The   
   > memory leak detector is embedded into Windows kernel, just must be   
   > activated by   
   > a registry flag.   
   >   
   >> Please, do not misunderstand me! I do not say that   
   >> low-/intermediate-level   
   >> languages are rubbish! They are good, but good for experts!   
   >   
   > Correct. For many other tasks, PHP, Java and C# are adequate - due to them   
   > being _forgiving_, more forgiving then C. Again no room for, say, Ada.   
   >   
   > Note: you say about "please be more forgiving to bad developers". Then   
   > Dmitry,   
   > who holds your side in this argument, is speaking about deep theoretical   
   > things   
   > like the "multiple dispatch" which even average-to-good developer will not   
   > understand, and looks like comprehension of such things are necessary to   
   > study,   
   > say, Eiffel. A paradox.   
   >   
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)