Associative Arrays

[xXziroXx]

Will we ever get them?

[Codein]

Oooh, yeah. Good idea.

[WhiteDragon]

We already have hashes, associative arrays would just be the same thing with different syntax.

[xXziroXx]

Originally Posted by WhiteDragon

We already have hashes, associative arrays would just be the same thing with different syntax.

There's a difference between manually scripting things and to have them implemented straight in the engine, though.

[WhiteDragon]

Originally Posted by xXziroXx

There's a difference between manually scripting things and to have them implemented straight in the engine, though.

I don't really get what you mean.

Hashes are the most common way associative arrays are implemented, so we have associative arrays, just without the conventional syntax one would see in for example in PHP.

[Loriel]

At some point you would think people would just want to plob javascript into the engine.

[Crow]

Originally Posted by Loriel

At some point you would think people would just want to plob javascript into the engine.

Lua please.

[Loriel]

Originally Posted by Crow

Lua please.

Basically a toss-up between one-based arrays and the icky closure semantics, but, yeah, okay, lua.

[WhiteDragon]

Originally Posted by Loriel

Basically a toss-up between one-based arrays and the icky closure semantics, but, yeah, okay, lua.

Well, now that we're at it... Haskell?

[Loriel]

Originally Posted by WhiteDragon

Well, now that we're at it... Haskell?

I have never felt a stronger urge to throw my chair at some guy through the internet than I am experiencing right now.

I would not even know where to begin to write an associate array in haskell that has any better than O(n) lookups.

[WhiteDragon]

Originally Posted by Loriel

I have never felt a stronger urge to throw my chair at some guy through the internet than I am experiencing right now.

I would not even know where to begin to write an associate array in haskell that has any better than O(n) lookups.

Haskell provides the lookup function which works differently on different data structures.

If you do it on a list like [("a", "1"), ("b", "2"), ("c", "3")] it'll return in O(n).

However, if you do it on a Data.Map which are represented as immutable balanced binary trees, it'll give you a O(log(n)).

Those would be the preferred methods since they both provide a purely functional interface, but you can use hash tables as well with an imperative interface in the IO Monad (with a data HashTable) giving you the O(1).

[Inverness]

Keep your inferior languages to yourself, obviously Python would be the better choice, its popularity speaks for itself. Utilizing Boost::Python makes it much easier to wrap C++.

Python implements hash tables (associative array) as the 'dict' object. You can specify dict literals using angle brackets.

e = {'key1': 'value1', 'key2': 'value2'}
print(e['key2'])

Python allows you to define many functions called by the interpreter, in this case, e['key2'] = 3 maps to e.__setitem__('key2', 3). You can quite easily take control of item or attribute lookup or other things.

One of the major points in favor of Python is readability, which is obviously important in Graal considering the programming experience and amount of people that would be looking at one script.

Thanks for thread URL Ziro.

[Loriel]

Originally Posted by Inverness

Keep your inferior languages to yourself, obviously Python would be the better choice, its popularity speaks for itself. Utilizing Boost::Python makes it much easier to wrap C++.

If we are going down that road, keep your python to yourself and give me C++.

Lua is designed for embedding to begin with and there are plenty of ways to wrap C++ into lua. What makes boost.python superior to third-party lua binding libraries?

Python implements hash tables (associative array) as the 'dict' object. You can specify dict literals using angle brackets.

e = {'key1': 'value1', 'key2': 'value2'}
print(e['key2'])

Python allows you to define many functions called by the interpreter, in this case, e['key2'] = 3 maps to e.__setitem__('key2', 3). You can quite easily take control of item or attribute lookup or other things.

lua, and I am pretty sure javascript as well, does these too, for what it is worth, and the associative-array data type is way more integrated into the core languages of lua and javascript as far as I can tell.

One of the major points in favor of Python is readability, which is obviously important in Graal considering the programming experience and
amount of people that would be looking at one script.

This is way subjective. Show me how Python is more readable than lua. Lua's syntax is really hard to beat as far as elegance and straightfowardness go.

[Inverness]

Originally Posted by Loriel

If we are going down that road, keep your python to yourself and give me C++.

Why have one when you can have both? It's quite easy to write C/C++ extensions for Python.

Originally Posted by Loriel

Lua is designed for embedding to begin with and there are plenty of ways to wrap C++ into lua. What makes boost.python superior to third-party lua binding libraries?

Python is designed for both embedding and extending. Boost::Python is merely the connection between C++ and Python's C API, and is designed to wrap your C++ code without changing it. There's also the fact that it is part of Boost, I'd think the name carries some weight.

Originally Posted by Loriel

lua, and I am pretty sure javascript as well, does these too, for what it is worth, and the associative-array data type is way more integrated into the core languages of lua and javascript as far as I can tell.

Python uses the dict (dictionary/associative-array) data type for storing all local, module (global), and object variables. It is a vital part of the language, so naturally it is extremely optimized same with Lua and Javascript.

Originally Posted by Loriel

This is way subjective. Show me how Python is more readable than lua. Lua's syntax is really hard to beat as far as elegance and straightfowardness go.

I'd like a specific situation to apply it to.

[WhiteDragon]

Hey, let's find a thread where Inverness DOESN'T start arguing that Python is better than all other languages when he actually hasn't used most of the other ones people are talking about!

[Inverness]

Originally Posted by WhiteDragon

Hey, let's find a thread where Inverness DOESN'T start arguing that Python is better than all other languages when he actually hasn't used most of the other ones people are talking about!

If you get the impression that I'm saying Python is "better than all other languages" then you have the wrong one.

I'm saying Python is better for Graal's needs, which are: clear and readable syntax, extensive documentation, and rich object-oriented capabilities, among other things.

[WhiteDragon]

Originally Posted by Inverness

Keep your inferior languages to yourself, obviously Python would be the better choice

Originally Posted by Inverness

If you get the impression that I'm saying Python is "better than all other languages" then you have the wrong one.

Either that or I'm illiterate.

[Inverness]

Originally Posted by WhiteDragon

Either that or I'm illiterate.

I think its the latter.

And I await your response to the second half of my previous post.

[WhiteDragon]

Originally Posted by Inverness

I'm saying Python is better for Graal's needs, which are: clear and readable syntax, extensive documentation, and rich object-oriented capabilities, among other things.

1. Lua has clear and readable syntax.
2. Lua has extensive documentation.
3. Lua is richly MULTI-paradigm! Check it out, you can be imperative, functional, prototypal, and object-oriented! Amazing!


1. Haskell has concise syntax.
2. Haskell has extensive documentation.
3. Haskell is a functional programming language! Sweet! Check it out, everything is a function!

[Inverness]

Originally Posted by WhiteDragon

<post>

I see what you did there.

Of course, the better means of comparison would be to solve Graal problems in the language. And I think I'm bored enough to do that.

[Loriel]

Originally Posted by Inverness

Of course, the better means of comparison would be to solve Graal problems in the language. And I think I'm bored enough to do that.

Ah, let us just hack together a quick Graal clone and cram any scripting language we can get out hand on into it!

[Inverness]

Originally Posted by Loriel

Ah, let us just hack together a quick Graal clone and cram any scripting language we can get out hand on into it!

Already working on it.