It's just a different method to coding. Having to check for types isn't worse than having to convert between types.

Typeless variables aren't necessarily better but they're not necessarily worse either. You can mess them up with bad code just as you can strict types.

Always gets me to smile when the newness of of style and syntax troupe gets a walk around the block.

Technically true since 1966 is a whole eight years after 1958.

Couldn't disagree more.. i found python more confusing than C++ in attempting to learn them. (not that i'm very fluent in either)

This is what really annoys me in this whole discussion. How in the world did python get the reputation of being so much easier to learn than any other language? I found it horribly complex in comparison to other languages i've learned and gave up on it. Whats the criteria for an "easy to learn" scripting language? Cause i Personally don't see it in python.

TBH i'm not worried about what language they choose.. if its something i'm comfortable with i'll mod if not.. i won't. I just want to know where all the python is easy to learn rhetoric came from.

In the case of preforming an undefined operation, it is big. Having it not fail would be like accepting XML with a tag that does not exist (likely misspelled) and then not preforming as intended. In such a case it would be better for the loading of the XML to fail right then and tell you where the failure occured. While it is not perfect protection it can save lots of bother when it does catch a mistake.

Even if they do not, Elemental does use multi-threading and may (will) call multiple scripts at once. In this case the scripting engine must be able to deal with it without too much overhead.

And there is no reason for a script that just preforms calculations based on data passed* to it should require a second process or not be able to run at the same time as other such scripts.

* As in it gets data in away that will not let it change anything. Changes will be done elsewhere based on the result.

Doesn't Python have similar sandboxing issues? I don't know the specific, but I seem to remember that Wesnoth moved away from Python due to the problem of untrusted code. For example, is it possible to prevent code from calling "import" and the like without directly hacking the interpreter?

As for Lua, it is possible to have it call untrusted code while only giving the functions that you want. I have seen a situation where Lua code was called with zero functions (not even print), which is obviously what you want to do if the file is supposed to be pure data and is not trusted at all.

There's a big, big difference between being annoying for a minute while you get your types straightened out, and being annoying for months and months because you accidently misspelled something and put a garbage value deep in your code that causes it to fail one in twenty times! Strong typing goes a long way to eliminate some of the most insidious errors by catching it in the compiler.

None of the languages discussed have especially strong typing (they have far from weak systems but there has been much more research done). Catching errors in compilation in languages without much of type inference doesn't catch any noticable percentage of insidious errors.

The non-programmers will still have access to the XML modifications. I think that this level needs to be for modders who are beyond the beginner level as far as programming is concerned.

Just to check, by type inference do you mean it as defined at http://en.wikipedia.org/wiki/Type_inference?

 More or less yes. Haskell witch I regularly use goes beyond H-M type inference to Lambda Cube and System F territory. One of many fun things possible in second order type systems are tools that given a function signature gives a suggested function implementation for that signature.

I'd go with whatever language would offer the most power and ease of access to a hypothetical Elemental fan with little to no coding experience and offer C++ (or whatever) to the experienced scripters. The reason being is that the easier it is for Average Joe Gamer to mess with stuff, the easier it is for the Average Joe community to get excited about any and everything. Few things are more powerful than word-of-mouth advertisement combined with people being able to pimp out their own stuff.

I'd rather see 10,000 would-be designers getting excited over a bunch of rocks and leaving the headaches to the 100 experienced scripters who have the knowledge to handle problems over 10,000 would be designers getting frustrated and waiting for the 100 to do something slightly faster/better in 10 days instead of 14. The 100 experienced scripters can handle themselves and will figure things out on their own. The others, not so much.