The LOLCODE forum

I agree with most of the amendments, but have some issues with it:
- There's no reason to eliminate the option for any unary function to be used as the operation. Just because increment and decrement are most common doesn't mean that no other operation is useful. For example, a loop where the index is doubled each time could be useful for iterating through powers of two. And, of course, anyone can define their own unary function that does whatever they want it to.
- There's also no reason to make the loop variable read only. While messing with it is unsightly, I don't see any reason to prohibit it.
- On the other hand, _encouraging_ messing with it by having a STEPPIN operator also seems like a dumb idea to me.

risser wrote:

A couple of things.  First, labeled GTFOs.  I think they are useful, but they are also being used in WTFs, aren't they?  What if I have a switch in a loop and use the labeled GTFO inside the switch?  I assume it drops out of everything, but I'm worried it might be difficult to implement.

Sorry for being so long replying. I got distracted by your own proposal. The more I read it, the more I realised I was carrying over some habits from my Forth days. LOLCode is very different, as i'm slowly starting to appreciate.

I don't see how you could exit a loop with an unlabeled GTFO, since it has to be enclosed by a conditional terminated by OIC. At that point, because of the label, the compiler already knows just what structures have to be dropped by the run-time jump.

Example:

IM IN YR HUNTIN
    GOT ANIMAL, WTF?
     OMG “Hedgehog� 
        VISIBLE “Prickly!!�  GTFO
     OMG “Dog� 
        VISIBLE “Halp!!!�  GTFO HUNTIN
     OMG “Maus� 
        VISIBLE “NOM NOM NOM�  GTFO
      OMGWTF
        VISIBLE “Wat iz Dat?� 
      OIC
      VISIBLE “Still Huntin!!� 
IM OUTA YR HUNTIN

Second, I don't like the way that normal loops and counting loops are different, when they don't have to be.  The simple addition of a "FROM" allows the ability to assign a starting value.  There are other issues with it as well:

i've addressed most of these issues, but you have also addressed the difficulty I had with your version, and I've understood it better. I'm particularly glad you got rid of the need for a boolean termination test - LOLCode comparisons are soo verbose!

- What happens with STEPPIN in a nested loop?  Will it always step the inner-most loop variable?

Yes, trust me.

- STEPPIN is interesting, in that it's the increment and the test all in one, and that you can put it in different places and give it different values.  However...

I've fixed it now. It makes more sense as a continuation than as a break. Interesting, yes, but i have been unable to come up with a convincing use for it.

In short, I can't think of any reason now to reccommend my proposal above yours.

Not that example, because the 'exit loop value' is raised by 1 for every 1 lice, this means that it will loop ad infinitive.

another way to write it is: keep ageing until you are older than your dad.
Also, you didn't use IT anywhere... wtf?

JoshSuereth wrote:

I have a random idea, what if loops check the value in IT for WIN before continuing (always).


Example:

Code:

I HAS  A LICE ITZ 1
IM IN YR BED
   GIMMEH bedfellow
   WTF?
        OMG "mouse"
              VISIBLE "nom nom nom"
              GTFO
        OMG "lice"
              LICE R SUM OF LICE AN 1
              GTFO
        OMGWTF
              VISIBLE "WTF?"
   OIC
   BOTH SAEM LICE AN BIGGR OF LICE AN 10
IM OUTA YR BED

This loops until there are 10 LICE in the bed.

That's Java's DO-WHILE which is probably the least common looping construct, because he whole of the loop has to be executed before it can exit. Using my WILE it's the same as:

   OIC
   WILE DIFFRNT LICE AN 10     BTW I think you meant
IM OUTA YR BED

not

    OMG "lice"
              LICE R SUM OF LICE AN 1
              WILE DIFFRNT LICE AN 10
        GTFO
        OMGWTF
              VISIBLE "WTF?"
   OIC
   BTW code here is skipped on 10 lice

IM OUTA YR BED

But there's something odd about the first construction. I reads like I'm coming out of the bed while lice<>10, because that's the next statement. I don't get that with the more complex counted loop, because there I've already clearly stated that I'm repeating the block.

The trouble is, a bare  IM IN YR <label> ... IM OUTTA YR <label> just doesn't read like an infinite loop. 

On the good basi of telling folks what you're going to do before you do it, I'd prefer "IM IN YR <label> FOREVA", knowing that cats are fickle creatures, and are bound to GTFO somewhere. This corresponds with what I've often heard said when teaching beginners - "Ah, but there your in your ... you need to"

What other uses are there for IM IN YR and IM OUTTA YR as named scope markers?

Heliomance wrote:

Proposal for a FOR EACH loop

Code:

IM IN UR <label> PLAYIN WIV UR <var1> <var2>

as equivalent to

Code:

for <var1> in <var2>

I'm not sure that the spec for bukkits is developed well enough to decide yet how to implement a FOR EACH, but I've done some more work on my idea of letting the loop identifier be more than a label.  See the re-written proposal http://lolcode.com/proposals/1.3/loop

In this scheme the syntax would probably be   "IM IN YR <var1> WATCHIN <var2>"

memodude wrote:

a) there's a few typos that need to be fixed ("indentifer"??!!)
b) the "Exiting a Scope" example has a few lines that contain multiple statements not separated by commas
c) calling loops "Named Scopes" is just downright confusing, especially when all "Named Scopes" are loops

I've fixed these now.  Much of what I was thinking could be done with named scopes could perhaps be done better with "O HAI" or in some other way. We'll see. 

d) the second example in "Conditional Exits" seems to use the loop iterator as some kind of implied default object - this is lame.

I was thinking that with objects we're going to get many!!deeply!!nested!!calls, and there should be some like VB's

WITH deeply.nested.object
        .doThis
        .doThat
        .doTother

But I forgot I'd have to end with "IM OUTTA DEEPLY!!NESTED!!OBJECT". Not good

The alterative I've given now should work, I think.  It seems right that the "value" of a object is a pointer to the object.

Norgg wrote:

The ANIMAL example used in this proposal has a couple errors in it (use of GOT and misspelling of GIMME).

Thanks for pointing that out. Changed now.