# Cute open range?

Discussion in 'Ruby' started by Clifford Heath, Feb 15, 2008.

1. ### Clifford HeathGuest

So I wanted to allow folk using my DSL to say "0..N",
"1..N", etc, to describe the cardinality of a relationship,
and I needed to find a useful way to define N without
making it a fixed integer maximum value.

Well, I came up with this, which I thought was cute:

class N
def self.coerce(n)
[ n, n+1 ] # Anything you can do, I can do better
end
end

Now when I ask (0..N) === m, for any m, it says "true",
and when I say the following, it never ends:

(0..N).each do |n|
puts n
end

But more importantly, I can say:

do_something if range.last == N

Ruby is sweeet .

Clifford Heath.

Clifford Heath, Feb 15, 2008

2. ### Joel VanderWerfGuest

Clifford Heath wrote:
> So I wanted to allow folk using my DSL to say "0..N",
> "1..N", etc, to describe the cardinality of a relationship,
> and I needed to find a useful way to define N without
> making it a fixed integer maximum value.
>
> Well, I came up with this, which I thought was cute:
>
> class N
> def self.coerce(n)
> [ n, n+1 ] # Anything you can do, I can do better
> end
> end
>
> Now when I ask (0..N) === m, for any m, it says "true",
> and when I say the following, it never ends:
>
> (0..N).each do |n|
> puts n
> end
>
> But more importantly, I can say:
>
> do_something if range.last == N
>
> Ruby is sweeet .

Does that do anything that Infinity doesn't do?

irb(main):001:0> Infinity = 1/0.0
=> Infinity
irb(main):002:0> (0..Infinity) === 3
=> true
irb(main):003:0> (0..Infinity) === -1
=> false
irb(main):004:0> (0..Infinity).each do |n|
irb(main):005:1* p n
irb(main):006:1> break if n > 3
irb(main):007:1> end
0
1
2
3
4
=> nil
irb(main):008:0> puts "do_something" if (0..Infinity).last == Infinity
do_something
=> nil

Also, there's this....

class N
def self.coerce(n)
[ n, n+1 ]
end
end

x = 3 + N
p x # ==> 7 w.t.f.?

Infinity = 1/0.0
x = 3 + Infinity
p x # ==> Infinity

--
vjoel : Joel VanderWerf : path berkeley edu : 510 665 3407

Joel VanderWerf, Feb 15, 2008

3. ### Clifford HeathGuest

Joel VanderWerf wrote:
> Does that do anything that Infinity doesn't do?

Hey, I forgot about Infinity, much better idea,
and you get -Infinity too . But this does do
one thing that Infinity doesn't: it works correctly
with Bignums that are too big to be coerced to a
float. I don't think that should concern me though .

Is there a standard place where Infinity is defined,
or should I risk re-defining it? Or worse, test and
define only if needed...

Clifford Heath.

Clifford Heath, Feb 15, 2008
4. ### Robert KlemmeGuest

On 15.02.2008 07:34, Clifford Heath wrote:
> Joel VanderWerf wrote:
>> Does that do anything that Infinity doesn't do?

>
> Hey, I forgot about Infinity, much better idea,
> and you get -Infinity too . But this does do
> one thing that Infinity doesn't: it works correctly
> with Bignums that are too big to be coerced to a
> float. I don't think that should concern me though .
>
> Is there a standard place where Infinity is defined,
> or should I risk re-defining it? Or worse, test and
> define only if needed...

Why define or redefine? Why not just

irb(main):001:0> N = 1/0.0
=> Infinity
irb(main):002:0> (1..N) === 2
=> true
irb(main):003:0> (1..N) === -2
=> false

etc? I believe this is what Joel wanted to suggest.

Kind regards

robert

Robert Klemme, Feb 15, 2008