"0= if" or simply "if" ??? - Programming Languages

Duke Normandin wrote:[color=blue]
> I'm back at doing tutorial examples. I wrote these 2 words:
>
> : ?metric ( ? -- ? )
> 10 mod
> if ." not metric..."
> else ." multiple of ten!"
> then
> cr ;
>
> : ?metric2 ( ? -- ? )
> 10 mod 0=
> if ." multiple of ten!"
> else ." not metric..."
> then
> cr ;
>
> Which is the better form? Which is least expensive? Personally I find
> the latter more readable and self-explanatory.[/color]

It depends on what you're doing. Usually you're much more interested in
one case or the other, so that should be the 'true' side even if it
takes a 0= to do it. It's also poor style to have an empty IF clause
and something happening in the ELSE; better to do 0= and omit the ELSE.

Personally, I like to see the test in the same 'phrase' as the IF, and
the consequences indented below:

: ?metric ( ? -- ? )
10 mod if ." not metric..."
else ." multiple of ten!"
then cr ;

or

: ?metric ( ? -- ? )
10 mod 0= if
." multiple of ten!"
else ." not metric..."
then cr ;
[color=blue]
> Also, what should the "stack effect" include? Should it reflect *just*
> user input, or user input and the "10" above? On the return side, should
> the conditional flag be taken into consideration? TIA...[/color]

The stack effect is intended to show what the word expects and leaves on
the data stack. Since the 10 is supplied internally, that isn't an
expectation of the word itself. And these definitions don't leave
anything at all on the data stack. So, I'd show it as ( n -- ).

Cheers,
Elizabeth

--
==================================================
Elizabeth D. Rather (US & Canada) 800-55-FORTH
FORTH Inc. +1 310-491-3356
5155 W. Rosecrans Ave. #1018 Fax: +1 310-978-9454
Hawthorne, CA 90250
[url]http://www.forth.com[/url]

"Forth-based products and Services for real-time
applications since 1973."
==================================================

On Mar 22, 2:42 pm, Duke Normandin <merr...@telus.net> wrote:
....[color=blue]
> Which is the better form? Which is least expensive? Personally I find
> the latter more readable and self-explanatory.[/color]

Of course you should use:

10 mod 0if
." multiple of ten!"
else
." not metric..."
then
cr ;

which is why Reva has "0if". But that is all a digression. You
should, at this point, not worry too much about "expense" but rather
be concerned with readability and maintainability.

Generally, the more words the more code is generated (that is not
always true, but it's a pretty good rule-of-thumb. But again, just
aim for code you can understand now and that you'll be able to
understand in a month or six.

On 2007-03-22, Ron Aaron <rambamist@gmail.com> wrote:[color=blue]
> On Mar 22, 2:42 pm, Duke Normandin <merr...@telus.net> wrote:
> ...[color=green]
>> Which is the better form? Which is least expensive? Personally I find
>> the latter more readable and self-explanatory.[/color]
>
> Of course you should use:
>
> 10 mod 0if
> ." multiple of ten!"
> else
> ." not metric..."
> then
> cr ;
>
> which is why Reva has "0if". But that is all a digression. You
> should, at this point, not worry too much about "expense" but rather
> be concerned with readability and maintainability.
>
> Generally, the more words the more code is generated (that is not
> always true, but it's a pretty good rule-of-thumb. But again, just
> aim for code you can understand now and that you'll be able to
> understand in a month or six.
>
>[/color]

Nice trick, Ron! ) ...as if I'm not confused enough as it is

Seriously (Elizabeth and Ron) I'm currently knee-deep in Brodie's
"Starting Forth" at:

"...the fact that an arithmetic zero is identical in meaning to a "false" flag
leads to some interesting results."

That was what my first example was supposed to show.

and

"...All comparison operators return well-formed flags"

Would my ?metric2 be an example of using a "well-formed" flag? If it is,
then IMHO that's the way to go, because "well-formed" == "easily
groked". Thanks!
--
duke | A: Yes. |
| >Q: Are you sure? |
| >>A: Because it reverses the logical flow of conversation. |
| >>>Q: Why is top posting frowned upon? |

Duke Normandin wrote:[color=blue]
> I'm back at doing tutorial examples. I wrote these 2 words:
>
> : ?metric ( ? -- ? )
> 10 mod
> if ." not metric..."
> else ." multiple of ten!"
> then
> cr ;
>
> : ?metric2 ( ? -- ? )
> 10 mod 0=
> if ." multiple of ten!"
> else ." not metric..."
> then
> cr ;
>
> Which is the better form? Which is least expensive? Personally I find
> the latter more readable and self-explanatory.
>
> Also, what should the "stack effect" include? Should it reflect *just*
> user input, or user input and the "10" above? On the return side, should
> the conditional flag be taken into consideration? TIA...[/color]

I think it's a beginner hangup to write "if x .ne. 0" in BASIC, C, or
Fortran instead of "if x". They mean the same thing and should be seen
that way. Same for Forth.

Jerry
--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

Duke Normandin wrote:
....[color=blue]
> Seriously (Elizabeth and Ron) I'm currently knee-deep in Brodie's
> "Starting Forth" at:
>
> "...the fact that an arithmetic zero is identical in meaning to a "false" flag
> leads to some interesting results."
>
> That was what my first example was supposed to show.
>
> and
>
> "...All comparison operators return well-formed flags"
>
> Would my ?metric2 be an example of using a "well-formed" flag? If it is,
> then IMHO that's the way to go, because "well-formed" == "easily
> groked". Thanks![/color]

I agree with Jerry. You need to imprint on your brain that IF (and also
WHILE and UNTIL, when you get to them) care only about zero=false and
non-zero=true. Don't clutter your code with meaningless tests.

The reason the comparison operators return well-formed flags is so you
can use the Boolean combination operators AND OR XOR with them.

: test ( n -- ) dup 100 < swap 10 mod 0= and if
." Yay!"
else ." Boo!"
then ;

Cheers,
Elizabeth

--
==================================================
Elizabeth D. Rather (US & Canada) 800-55-FORTH
FORTH Inc. +1 310-491-3356
5155 W. Rosecrans Ave. #1018 Fax: +1 310-978-9454
Hawthorne, CA 90250
[url]http://www.forth.com[/url]

"Forth-based products and Services for real-time
applications since 1973."
==================================================

Duke Normandin wrote:[color=blue]
> On 2007-03-22, Ron Aaron <rambamist@gmail.com> wrote:[color=green]
>> On Mar 22, 2:42 pm, Duke Normandin <merr...@telus.net> wrote:
>> ...[color=darkred]
>>> Which is the better form? Which is least expensive? Personally I find
>>> the latter more readable and self-explanatory.[/color]
>> Of course you should use:
>>
>> 10 mod 0if
>> ." multiple of ten!"
>> else
>> ." not metric..."
>> then
>> cr ;
>>
>> which is why Reva has "0if". But that is all a digression. You
>> should, at this point, not worry too much about "expense" but rather
>> be concerned with readability and maintainability.
>>
>> Generally, the more words the more code is generated (that is not
>> always true, but it's a pretty good rule-of-thumb. But again, just
>> aim for code you can understand now and that you'll be able to
>> understand in a month or six.
>>
>>[/color]
>
> Nice trick, Ron! ) ...as if I'm not confused enough as it is
>
> Seriously (Elizabeth and Ron) I'm currently knee-deep in Brodie's
> "Starting Forth" at:
>
> "...the fact that an arithmetic zero is identical in meaning to a "false" flag
> leads to some interesting results."
>
> That was what my first example was supposed to show.
>
> and
>
> "...All comparison operators return well-formed flags"
>
> Would my ?metric2 be an example of using a "well-formed" flag? If it is,
> then IMHO that's the way to go, because "well-formed" == "easily
> groked". Thanks![/color]

A well formed flag (WFF for short) in ANS Forth is either 0 or -1. Those
are the values *produced* by tests, and they allow what some think is
abominable and I think is neat and simple. Flags *consumed* by IF (or
UNTIL or ...) needn't be well formed. All that means is that any
non-zero value may stand in for -1 without altering the outcome.

Jerry

P.S. Suppose I want a word that leaves positive numbers unchanged and
sets negative numbers to zero.

: rectifier ( n -- n|0 ) DUP 0> AND ; / abominable, or neat?
--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

Jerry Avins escreveu:
[snipped]>[color=blue]
> I think it's a beginner hangup to write "if x .ne. 0" in BASIC, C, or
> Fortran instead of "if x". They mean the same thing and should be
> seen that way. Same for Forth.
>[/color]
I think is a good practice that begginers are forced to abandon due
'experts' calling it "hangup" until they need to code in a professional
environment and have to 're-learn' to write code that is more readable
(== maintainable).

Cesar Rabak wrote:[color=blue]
> Jerry Avins escreveu:
> [snipped]>[color=green]
>> I think it's a beginner hangup to write "if x .ne. 0" in BASIC, C, or
>> Fortran instead of "if x". They mean the same thing and should be
>> seen that way. Same for Forth.
>>[/color]
> I think is a good practice that begginers are forced to abandon due
> 'experts' calling it "hangup" until they need to code in a professional
> environment and have to 're-learn' to write code that is more readable
> (== maintainable).[/color]

Obviously we differ. To me it is easier to maintain "If X" than "If X
exists" simply because it clearly means the same thing and uses fewer
words. When I need to understand code you write your way, I will suffer
through it in silence.

Jerry
--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

On 2007-03-23, Elizabeth D Rather <eratherXXX@forth.com> wrote:[color=blue]
> Duke Normandin wrote:
> ...[color=green]
>> Seriously (Elizabeth and Ron) I'm currently knee-deep in Brodie's
>> "Starting Forth" at:
>>
>> "...the fact that an arithmetic zero is identical in meaning to a "false" flag
>> leads to some interesting results."
>>
>> That was what my first example was supposed to show.
>>
>> and
>>
>> "...All comparison operators return well-formed flags"
>>
>> Would my ?metric2 be an example of using a "well-formed" flag? If it is,
>> then IMHO that's the way to go, because "well-formed" == "easily
>> groked". Thanks![/color]
>
> I agree with Jerry. You need to imprint on your brain that IF (and also
> WHILE and UNTIL, when you get to them) care only about zero=false and
> non-zero=true. Don't clutter your code with meaningless tests.[/color]

I'm just doing the examples and exercises in Brodie's book. Is there a
better tutorial that I could use IYO?
[color=blue]
> The reason the comparison operators return well-formed flags is so you
> can use the Boolean combination operators AND OR XOR with them.
>
>: test ( n -- ) dup 100 < swap 10 mod 0= and if
> ." Yay!"
> else ." Boo!"
> then ;[/color]

OK!! So am I to conclude that generally it's NOT a good idea to use
arithmetic zero results (eg mod) to perform double duty as flags? Or am
I totally confused?
--
duke | A: Yes. |
| >Q: Are you sure? |
| >>A: Because it reverses the logical flow of conversation. |
| >>>Q: Why is top posting frowned upon? |