Re: google "top coder" contest = stacked against C++ coders
From: White Wolf (wolof_at_freemail.hu)
Date: 10/03/04
- Next message: White Wolf: "Re: ostream like interface to a string?"
- Previous message: assaarpa: "Re: google "top coder" contest = stacked against C++ coders"
- In reply to: assaarpa: "Re: google "top coder" contest = stacked against C++ coders"
- Next in thread: Francis Glassborow: "Re: google "top coder" contest = stacked against C++ coders"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Sun, 3 Oct 2004 23:53:22 +0300
assaarpa wrote:
>> In a well design OS you do not need all that.
>
> True, but not all very widely used OS are well designed. ;-)
>
>> Why do you take this as an attack against you?
>
> Don't worry: I didn't, I appreciate your opinion and views but don't
> agree some of the conclusions.
>
>> Did you design how DLLs work in Windows?
>
> The point I am driving at is that it is well and good to know the C++ and
> how it works and how things *should* behave, but those are just
> guidelines which should be followed to have a shot at correctly behaving
> software. Guidelines which are very, very important as not following
> them has dramatic side effects like demons flying out of the nose.. but
> at end of the day, you still have to know the platform and architechture
> you are compiling for, there is no substitute for competence.
IOW: C++ does not define how dynamic linking OR multithreading works or does
not with C++. So there aren't even guidelines. Some companies choose to
make systems which makes it simple to work, some made systems where it is
harder.
> It is possible to write 100% conforming C++ sourcecode, which doesn't
> work on a very widely deployed platform: Windows.
??? As soon as you are using dynamic linking or MT it is NOT 100%
conforming code.
> If this is good or bad
> I don't care to take a stance on, just stating the fact: nothing more
> nothing less. I am not taking this personally, but facts are facts,
> sometimes the reality just manages to elude some of the very optimistic
> people here. ;-)
But let's ask yourself honestly. If there are many OSes where it is easy to
do... is it a problem with C++ or is it a problem with Windows? And is it a
problem Microsoft tries to make into your problem instead of solving, or is
it primarily your problem to solve?
> It is different to write code that invokes undefined behaviour, and code
> that doesn't but still doesn't work. The former is poor software
> engineering and, well, so is the later -- but -- for the context of this
> NG, irrelevant, thus, I admit that I been ranting about inconsequential
> off-topic platform specific implementation detail (in otherwords, I'm
> wrong).
The trouble is, that as soon as you do dynamic linking or MT you are not
doing standard C++ and all bets are off. Then again: some companies made
those things easy on the people, some did not.
-- WW aka Attila ::: The things that come to those who wait may be the things left by those that got there first.
- Next message: White Wolf: "Re: ostream like interface to a string?"
- Previous message: assaarpa: "Re: google "top coder" contest = stacked against C++ coders"
- In reply to: assaarpa: "Re: google "top coder" contest = stacked against C++ coders"
- Next in thread: Francis Glassborow: "Re: google "top coder" contest = stacked against C++ coders"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
