More evidence against Sepples

Fuck Alioth's shootout. Check it:
http://www.cs.bell-labs.com/cm/cs/pearls/sec151.html
This C program takes about 2.4 seconds to read its input (the same as the C++ version), but only 0.5 seconds for the insertions (down from 4.9) and only 0.06 seconds to write the output (down from 0.3). The complete run time is 3.0 seconds (down from 7.6), and the processing time is 0.55 seconds (down from 5.2). Our custom-made hash table (in 30 lines of C) is an order of magnitude faster than the maps from the C++ Standard Template Library.

>>1
Thanks for stating the obvious.

Comparisons like these always piss me off. The author is missing using the library components, and he is not using the correct ones for the task at hand.

Furthermore, no one uses the standard iostreams for anything serious. In fact, no one uses scanf for anything serious either.

We use POSIX file functions (open, close, read, write, etc.) or Windows File API calls and do everything at the lowest level using asynchronous I/O.

The comparison is not comparing the languages. It is comparing the libraries.

>>3
This, std::map isn't a hash-table, it's a red-black tree or AVL tree depending on the implementation.

The author should have used std::unordered_map if he wanted a closer comparison.

>>1
You left out most important part:

This little exercise illustrates the two main ways to represent sets of words. Balanced search trees operate on strings as indivisible objects; these structures are used in most implementations of the STL's sets and maps. They always keep the elements in sorted order, so they can efficiently perform operations such as finding a predecessor or reporting the elements in order. Hashing, on the other hand, peeks inside the characters to compute a hash function, and then scatters keys across a big table. It is very fast on the average, but it does not offer the worst-case performance guarantees of balanced trees, or support other operations involving order.

If author didn't live in
Wed 18 Oct 2000
I'd suggest him to redo the test using unordered_map

>>4
unordered_map is "recent" addition. Also each member of committee should be shot in the face for using unoredered_map name instead of hash_map/hash_table

Also each member of committee should be shot in the face for using unoredered_map name instead of hash_map/hash_table
Because that's the most important feature of a programming language, that single-handedly determines its suitability (or unsuitability as in this case) to a wide range of real-world problems. Moron.

>>7
Can you name at least 10 situation, where Java(or language of your choice) has failed because it used HashMap name instead of UnorderedMapUnsuitabilityForFaggots name?

IHBT

>>8
UnorderedMapUnsuitabilityForFaggotsFactoryFactoryFactorySingleton.

>>8
That was the point, this shit is irrelevant and is not worth a mention and especially not worth shooting someone in the face over it.

People who hate C++ to the point where they create threads mocking it have autism and OCD. Every time they try to program with the language, they can't help but nit pick over every tiny non-issue.

Real pragmatic engineers don't have any problem with C++.

>>11
Real pragmatic engineers use C.

>>12
...and if they have to use C++, they use a minimum of C++ (eg library classes/COM on windows etc) and code their stuff in C

>>13
...which is a different subset of C++ used by, for example, a colleague, therefore becoming a pain to maintain.

>>12
Sure, using "char *" instead of strings and "type *" instead of vectors is a pragmatic choice. No shit.

>>14
I hear this a lot but have never experienced it myself (and I work with morons). I'm afraid it's a typical autistic fairytale: it seems entirely possible that the problem would arise unless there are other uncounted for factors (such as that maybe sane people tend to pick the mostly the same sane subset of C++?). But the little autistic shit can't imagine uncounted for factors, his logical argument is absolutely correct dammit, so he never tries to see if his ideas work IRL, or even ignores contradictory evidence gathered accidentally.

Just like autistic people in general and especially high-functioning asspies tend to ignore the mountains of evidence against their belief that they are geniuses who don't need to check their conclusions against reality in the first place.

I mean, everyone who had to deal with asspies should agree that their most annoying quality is the irrational belief that they are masterminds who can make people do what they want. Against the obvious fact that it only works with their parents, and not because they are fooled by the cunning schemes, but out of pity.

>>15
If you want to see C++ being used correctly, look at either the Clang/LLVM source, OpenCog soure, or at Microsoft's .NET Common Language Runtime implementation (which is implemented in regular C++).

It's a shame that a lot of people shun in the open-source world shun C++, they never learn enough to use the language effectively, and fall into a catch 22 trap.

>>1

Poor Alioth! - free project hosting blamed for a project they host.