Why ... Sucks

Why C sucks

Quite easy:

    * Primitive type system
    * No overloading
    * Limited possibility of data abstraction, polymorphism, subtyping and code reuse
    * No metaprogramming except preprocessor macros
    * No exceptions

Why C++ sucks

    * No garbage collection
    * No reflection
    * Type system excessively complicated
    * No support for variadic functions and variadic templates
    * Really hard to write efficient, object oriented mathematical code
    * Not even compile time reflection (ie in a template you cannot inspect the parameter type)

Why C# sucks

    * Way too prolix
    * It has many special purpose constructs (such as yield/foreach)
    * Not so powerful generic system
    * No compile time metaprogramming facility

Why Java sucks

    * No data pointers nor function pointers
    * No templates (java 1.5 templates are a fake)
    * No operator overloading
    * No anonymous methods (C++ operator())
    * Extremely verbose to write - and slow to run

Why Haskell sucks

    * Indentation dependent syntax
    * Types and values live in completely different worlds
    * No metaprogramming (solved by Template Haskell)
    * It is quite difficult to learn and to use for simple things, expecially if you come from an imperative/OO background. Sometimes, language features, extensions and design patterns are best documented by research articles, instead of tutorials written in an informal style.

Why Nemerle sucks

Some reasons as C#, except first and last.

Why Ocaml sucks

    * Too many keywords
    * Ugly syntax
    * No templates and oveloading (unless you are using a still a bit limited and incompilete gcaml)
    * Strange behavior of '_a and 'a types.

Why Ruby sucks

    * Slow
    * Complex
    * Surprising
    * Inconsistent
    * Bad embedding
    * No native threads
    * No m17n

>>40
From "The Unix Haters Handbook", pp241-242

>>39
No, you do it because you want to. You're too scared to admit to yourself, family and friends (do those in /prog/ actually have these?) that you are a faggot.

>>40

How do you justify saying that objects in C++ aren't objects when the code is running?  Sure they're mostly structs with mangling and other magic, but.. what do you think most of these other languages are?  Do you honestly think most of these languages at the lowest levels actually have objects?  It's all "Hexadecimal sludge" behind the scenes, and really.. there are working C/C++ garbage collectors out there.. Eg. libgc.  Many do not use this approach because speed matters above all else.

Many do not use this approach because speed matters above all else.

Except that it doesn't have much to do with speed anymore. If you use a language like C++, you're pretty much stuck with a conservative stop-the-world garbage collector. This is unfortunate, because it can cause delays.

Oh, wait, freeing something isn't deterministic either, unless you're doing custom memory management. But wait! You can do custom memory management with many GC'd languages too.

If you're writing for speed, you avoid allocating and deallocating to the heap. That applies whether or not you use a GC.

Now, add the wrinkle that modern GCs in higher-level languages can be incremental copying generational collectors. They give you O(1) allocation, and because they're incremental the delay is very low -- some are so fast their delay is on the order of that of a context switch.

An example of the latter is IBM's Metronome GC: http://domino.research.ibm.com/comm/research_projects.nsf/pages/metronome.metronomegc.html

EXPERT PROGRAMMERS call malloc one time during execution of their programs. Garbage collection languages are for people whose head is full of landfill.

>>45
You can't even say EXPERT PROGRAMMERS right. Why would I listen to you?

>>46
Umm… Did you mean ``EXPERT PROGRAMMERS''?

>>47
can't even spell EXPERT PROGRAMMERS

^{Back to /b/, ``GNAA Faggot''}