Scratch

I was thinking you could hake a hat block that will do something, then reports a boolean value. The hat block will check the condition every step of the project (unless it has a

wait () secs

or

wait until <>

block is pending) and, if it evaluates to true, the script starts. For example, you could make a

when key [space v] released
script

block that only fires when you let go of the key, defined as

when key (letter = space) released
if <key (letter) pressed?>
  wait until <not <key (letter) pressed?>>
  report <true>
else
 report <false>
end

Last edited by joefarebrother (2012-07-08 04:15:32)

Hey bharvey, my Scheme interpreter is running wonderfully well! But I've run into a bit of a problem. When writing the Factorial function, the if n=0 should only evaluate the clause it needs to, right? Because if it evaluates both the if-true clause and the if-false clause (and then picks the appropriate result) it would fall into an infinite loop if one clause is recursive. So, my question is, is "if" the only special case like this, or do I need to consider each function I add to see if all arguments should be evaluated right away or only if needed? Thanks! 

Hardmath123 wrote:

So, my question is, is "if" the only special case like this, or do I need to consider each function I add to see if all arguments should be evaluated right away or only if needed?

IF is one of maybe a dozen special forms in Scheme:  COND, DEFINE, SET!, AND and OR for lazy evaluation, QUOTE, QUASIQUOTE, LAMBDA, LET, LET*, LETREC, and DEFINE-MACRO or DEFINE-SYNTAX or whatever you're using for macros.  And of course the user-defined macros are special forms; that's the whole point of macros.

joefarebrother wrote:

I think macros should be in snap!!

Yes of course.  It's the only big missing piece of Scheme left, since we added call⁄cc.  I'm not promising hygienic macros, though; plain old FEXPR-style macros should do.

Hi asampal,

Snap follows Scratch's example in treating the ANSWER block as a sensor value, not as a user-settable variable (I've followed that thread on Scratch-Ed, btw). Since you never know what the last answer given previously actually was, I'd recommend you to not directly query the ANSWER value in any loops but to always assign it to a variable first (as you've found out yourself).

I was a little astonished by the corresponding thread on Scratch-Ed. OTOH I realize that Scratch does have a RESET TIMER block, so do you think a RESET ANSWER block would be important?

bharvey wrote:

Hardmath123 wrote:

So, my question is, is "if" the only special case like this, or do I need to consider each function I add to see if all arguments should be evaluated right away or only if needed?

IF is one of maybe a dozen special forms in Scheme:  COND, DEFINE, SET!, AND and OR for lazy evaluation, QUOTE, QUASIQUOTE, LAMBDA, LET, LET*, LETREC, and DEFINE-MACRO or DEFINE-SYNTAX or whatever you're using for macros.  And of course the user-defined macros are special forms; that's the whole point of macros.

I'm not dealing with quoting and macros for now, I just want something that works. 

Hardmath123 wrote:

I'd say the best way for the ask block to work would be an (ask for ()) reporter which reported the value typed in by the user.

+1, I think.  I'd certainly do it that way if we were designing a language from scratch rather than from Scratch.  As it is, I'm not sure it's important enough to change.

Last edited by bharvey (2012-07-09 07:00:14)

bharvey wrote:

joefarebrother wrote:

I think macros should be in snap!!

Yes of course.  It's the only big missing piece of Scheme left, since we added call⁄cc.  I'm not promising hygienic macros, though; plain old FEXPR-style macros should do.

I don't really understand those, I only understand the hygienic ones. Maybe it could be like logo macros?

joefarebrother wrote:

I don't really understand those, I only understand the hygienic ones. Maybe it could be like logo macros?

You understand hygienic macros??  I'm impressed; you are a true KLC.

FEXPRs are just like (Berkeley) Logo macros: the macro returns a new expression that then gets EVALed in place of the macro call itself.  The only difference is that in a FEXPR none of the inputs are evaluated; the entire macro call is used as the sole input to the FEXPR.

The issue with macro in a block-based language is that there's (so far) no way to build a block with desired input expressions (as opposed to input values, which CALL and RUN allow).

Well... Wikipedia says, "Hygienic macros are macros whose expansion is guaranteed not to cause the accidental capture of identifiers."

In this case:

#define INCI(i) {int a=0; ++i;}
int main(void)
{
    int a = 0, b = 0;
    INCI(a);
    INCI(b);
    printf("a is now %d, b is now %d\n", a, b);
    return 0;
}

Running the above through the C preprocessor produces:

int main(void)
{
    int a = 0, b = 0;
    {int a=0; ++a;};
    {int a=0; ++b;};
    printf("a is now %d, b is now %d\n", a, b);
    return 0;
}

These are not hygienic macros because the variable 'a' is redefined in the macro causing it not to increment, unlike 'b'. A hygienic macro would not do this, it would change the macro to avoid conflict. If the variable 'a' were defined, the macro might internally change to "{int a1 = 0; ++a};" so that 'a1' and 'a' don't conflict.

Right Brian? Did I screw this up? 

(credit to Wikipedia for the example lol)

Last edited by MathWizz (2012-07-09 21:34:46)

MathWizz wrote:

These are not hygienic macros because the variable 'a' is redefined in the macro causing it not to increment, unlike 'b'. A hygienic macro would not do this, it would change the macro to avoid conflict. If the variable 'a' were defined, the macro might internally change to "{int a1 = 0; ++a};" so that 'a1' and 'a' don't conflict.

Right Brian? Did I screw this up? 

Right, except you make it sound as if some C macros are hygienic and others aren't.  It's a language's whole macro system that is or isn't hygienic.  If a language uses hygienic macros, then the programmer doesn't have to worry about name captures.

It's the details that are complicated.  Sometimes a macro does want to refer to a particular variable in the caller, and that's the hairy part.  I think.

bharvey wrote:

Hardmath123 wrote:

Just numbers, lists, and lambda for now.

And how do you get a data list into your program without quoting it?  I mean, how do you say the equivalent of

(car (quote (1 2 3) ) )

?  (You can use the ' notation, as in

(car '(1 2 3) )

but that's just an abbreviation for the QUOTE special form.)

What's the difference between "quote" and "list"? Or have I got the wrong LISP?

Hardmath123 wrote:

Press the blue check mark button next to it.

I dont see any buttons next to it.