Is UML still a fad? Shouldn't it be going away soon? I don't get all those people that are all "OMG, UML!" about it.
Name:
Anonymous2008-07-01 17:18
>>32
Yes most of the "design patterns" are bullshit. I remember reading the GOF book (which is without a doubt the most boring book I have ever read). A lot of it is useful in *some* cases but again should not be abused.
I think all C++ users should read the meyers books and GOF and realize C++ sucks ass and switch. Seriously fuck C++.
>>42
Trolling aside, what's good besides C++ when you need something fast to do real-time volume rendering, image processing, or generally anything that can't be handled by slow as fuck languages?
FIOC with C++ for the most intensive parts?
Name:
Anonymous2008-07-01 22:48
>>48
Common Lisp, calling out to C if there's something that Lisp is somehow just too slow for (this will not happen a lot).
Duh.
Name:
Anonymous2008-07-01 22:52
C++ is horrible for performance. You get performance from it solely because everything is implemented pretty fast to begin with. But the second any modularity or layer of abstraction is added you immediately lose all those gains.
Other languages can handle this (MLton) gracefully, but C++ can't. The sad fact is that you often will write things in C or fortran and link to C++ to get better performance.
The better question is "If C++ is in this performance no man's land, not good enough when trying to be fast, and horribly slow when being idiomatic, why do people say it performs well at all?"
Name:
Anonymous2008-07-01 22:54
>>50
No, I think the better question is, "If C++, despite not actually being especially fast, is plenty fast for >>48's purposes, what's he fussing about?"
Name:
Anonymous2008-07-01 22:54
C++ isn't even good at virtual dispatch. Java kicks its ass at virtual dispatch benchmarks.
Name:
Anonymous2008-07-01 22:56
>>51
But C++ isn't even fast for that. To do image processing you link to fortran implemented libraries.
Name:
Anonymous2008-07-01 22:59
>>53
I see your point. I wasn't actually paying attention to what his purposes were.
>>55
It's true. C++ can't inline virtual functions, because C++ can only optimize during compilation. The JVM can also apply optimizations during runtime. So if a method gets called a lot, it can inline the method and the only overhead is a check that the object has the correct type.
Name:
Anonymous2008-07-02 8:22
WANKING OFF
Name:
Anonymous2008-07-02 9:27
So if a method gets called a lot
I don't think this is implemented anywhere.
This is not really related to topic, this is just may way of saying "use common sense".
Name:
Anonymous2008-07-02 15:53
>>58
That's your problem. You don't think to even fucking google it.
Name:
Anonymous2008-07-02 17:14
>>48
If you want fast handling of fucktons of data, use Fortran.
>>66
And I suppose LISP is the old gay man of computing. FORTH and Smalltalk are the young guys who for some reason hang around him. Sepples is a successful bug chaser completely unrelated to the above. He's a disgrace to his father, C, who worked hard all his life.
Name:
Anonymous2008-07-03 0:35
>>68
And Objective-C is C's and Smalltalk's little secret?
Name:
Anonymous2008-07-03 0:40
Java is the overweight thirtysomething with the mind-numbing office job who can never seem to get the hang of "business casual".
Name:
Anonymous2008-07-03 0:56
>>69
I guess so. Smalltalk would have to be a woman then.
Name:
Anonymous2008-07-03 1:00
>>71
At least she allows complete introspection...
Name:
Anonymous2008-07-03 1:59
>>72
That's the kind of woman I want. Except I'm not really into her. What kind of person loves Smalltalk?
Name:
Anonymous2008-07-03 8:19
FIOC is the stuck-up guy who's alway nagging the others about their TPS reports and whines if you write in the margin or use a special font.
Name:
Anonymous2008-07-03 8:38
>>59
I googled it and found few texts about how it can be implemented. JVM may do this, JVM may do that.
Is it implemented somewhere already? Would storing how often every function is called not already be a great overhead?
Name:
Anonymous2008-07-03 8:44
>>74
I don't see how that's any different to any other programming language. C is the stuck-up guy who's alway nagging the others about their TPS reports and whines if you write in the margin or use a special font.
>>77
That's not always a good thing. In my experience, I've seen way too many programmers abuse that philosophy to write code in their cryptic and non-standard (to the current project) manner. It is very rare that I see a hack that is truly worth breaking the guide of consistency. In a vast majority of cases, it is better to be consistant*.
* Yes, I do bracket my {
single_statement;
} code in C-style code in the name of consistency
Name:
Anonymous2008-07-03 10:47
>>79 Yes, I do bracket my { single_statement; } code in C-style code in the name of consistency
So do why, but I have issues.