Whenever I have to register at some bullshit forum, I typically take the name RichardStallman. If it's taken, I use RMSMatthewStallman. Sometimes I wonder how many people think that they are talking to the real rimmisu.
>>10
If you're going to reply to a post, at least quote it. Otherwise your replies are meaningless and quite passive-aggressive.
Name:
Anonymous2010-07-28 9:05
Scalability:7/10; //I interpreted this as in relation to as inbuilt language constructs for concurrency?
Performance:7/10;
Ease of use:10/10;
verbosity:8/10;
void trimCodans(void)
{
int found = 0;
for (;;)
{
int i;
for (i = 0; i < codans.rows; i++)
{
if (codans.table[(i * codans.cols) + codans.emptyLeft] != CODAN_EMPTY)
{
found = 1;
break;
}
}
if (found) break;
else codans.emptyLeft++;
}
//TODO: check right
}
Scalability: 10/10
Performance: 10/10
Ease of use: 3/10
Verbosity: 2/10
C++
Scalability: 7/10
Performance: 9/10
Ease of use: 1/10
Verbosity: 4/10
Java
Scalability: 8/10
Performance: 7/10
Ease of use: 7/10
Verbosity: 5/10
Python
Scalability: 0/10
Performance: 0/10
Ease of use: 10/10
verbosity: 5/10
Is there a language out there with worse performance than CPython? Serious question. It's still my language of choice for all non-performance-critical scripts :(
>>17 nothing performs better than C
Meaningless statement.
Aside: For some reason a lot of C++ programs tend to show better performance than equivalent C programs. I am still trying to work out why that is. The claim is often that C++ can express more efficient algorithms for these cases more naturally than C can, but every time I see a comparison made the C source is not what I would write (and not as fast.)
Name:
Anonymous2010-07-28 14:36
>>19
Make a benchmark:
1. C++ code
2. C code
3. C code which directly emulates C++
4. C code compiled as C++ code.
>>19
That claim is usually made by people who don't really know C.
Name:
Anonymous2010-07-28 14:39
>>19
The vast majority of the time, this is because of function pointers. Compilers aren't smart enough to statically link or inline a call of a function pointer argument. The equivalent C++ code uses a template parameter instead, forcing the compiler to statically link or inline it.
This isn't because it's not possible to do this in C; it's just because it's easier and safer in C++ than in C, so people are more likely to do it. The usual example is qsort() vs std::sort(). This fails of course as an argument for C++, because it's relatively easy to write a macro version of qsort() that does the same thing. For example: http://www.corpit.ru/mjt/qsort.html
What we really need is a language that can do this automatically. You just take a function pointer like normal, and if profile data determines that the function indirection is a bottleneck, the compiler can just duplicate the function for each value of the argument and statically link and inline each version. Unfortunately, modern languages are heading in the opposite direction in terms of performance and optimization, so we might never see this happen :(
Name:
Anonymous2010-07-28 14:50
im not a expert, but isnt typecasting much faster, safer and wastes less memory then templates?
Scalability: able to deliver scalable, modern turnkey implementations to rise to user demand
Performance: provides real-time results for round-the-clock mission critical solutions
Ease of use: patented accessibility platform ensures ease of integration between static computer and dynamic human inputs
Verbosity: colour-coded CamelCase engine creates a platform designed to hold the most information in the least amount of space while still being extensible to meet developer needs