Non-declared variables

I would like to hear people’s opinions on the ability to use variables
without declaring them in dynamic languages like Ruby.

It would seem to me that an argument for non-declared local variables
that typically occupy a small scope could be made. But, class
variables? What is the benefit in allowing for runtime definition of
class variables?

Are there real tangible benefits that non-declared, dynamically typed
(at binding time) variables provide? Or do dynamic variables simply
create less compile time errors and more (harder to catch) runtime
errors?

Thoughts?

On Mon, 16 Oct 2006 04:38:09 -0700, Avatar wrote:

create less compile time errors and more (harder to catch) runtime
errors?

Thoughts?

I was programming in Java last week, and writing typed code. I found
that
after about an hour of coding, my code would have worked correctly (it
would have been less time if Java supported operator overloading the way
ruby does), but I still had to spend another hour getting the types
appropriately declared, and typecasting things correctly as I pulled
objects out of their various collections.

C++ can be worse. I wanted to use the Boost Spirit library for CFG
parsing, and found that I couldn’t read the compile errors I was
getting because 90% of the text was multiply nested template types. The
C++
Standard Template Library uses templates so much to get close to Ruby’s
flexibility, but the result is still more complicated than Ruby,
impossible to debug, and less flexible. At that point, why not go all
the
way to a dynamically typed language?

–Ken B.

C++ can be worse. I wanted to use the Boost Spirit library for CFG
parsing, and found that I couldn’t read the compile errors I was
getting because 90% of the text was multiply nested template types. The C++
Standard Template Library uses templates so much to get close to Ruby’s
flexibility, but the result is still more complicated than Ruby,
impossible to debug, and less flexible. At that point, why not go all the
way to a dynamically typed language?

Because then you’re a different language with different use cases with
different benefits and drawbacks.

Boost is an incredibly good set of libraries, with an extremely high
level of QA. The great tragedy is that facilities like this were not
part of a C++ standard library a decade ago.

Still, the C++ Standards Committee are well aware of the problems that
templates cause with respect to indecipherable error messages, which
is why the new system of concepts has been proposed for C++0x, so the
compiler can flag straight away when you’ve tried to instantiate a
template parameter that won’t work.

Also, if you’re a genius, you can use template metaprogramming to do
some really spanky compile-time shenanigans.

Martin

On 10/16/06, Avatar [email protected] wrote:

I would like to hear people’s opinions on the ability to use variables
without declaring them in dynamic languages like Ruby.

You mean without declaring their type, right?

It’s more accurate to say simply that variables have no type in Ruby;
Rather the variable’s contents have a type.

In any case, Ruby requires that variables be declared before use just
as any decent language worth the name does.

Martin

Avatar wrote:

I would like to hear people’s opinions on the ability to use variables
without declaring them in dynamic languages like Ruby.

Could you clarify what you mean by that? Are you concerned about
methods like #instance_variable_set and #class_variable_set? Or more
generally about the fact that code like:

class Foo
@bar = 12
end

creates an instance variable on the fly at runtime, without requiring
some sort of
instance @bar
line that ‘declares’ it before use?

This forum is not affiliated to the Ruby language, Ruby on Rails framework, nor any Ruby applications discussed here.

| Privacy Policy | Terms of Service | Remote Ruby Jobs