Re: unhappy with loops
Posted by: jconkey <jconkey@...>
I must disagree with you about the FOR loop. I believe it is standard to
not execute the loop at all if the first value is > the last value. I also
think that the implementation of FOR in FutureBasic should be changed to be
this way.
0.02 from Jason 🙂
>>The FOR loop unhappiness
>>------------------------
>>ZBasic and FB made a poor choice in implementing FOR loops, such that the
>>loop body is _always_ executed at least once, regardless of the start, end
>>and increment values. To allow the body to be skipped we must bracket the
>>loop as shown:-
>>
>>LONG IF last%>=first% ' else we want to skip the loop
>> FOR j% = first% TO last%
>> .. do stuff
>> NEXT j%
>>END IF
>>
>>FOR loops in BASIC are officially supposed to be implemented such that the
>>body can be executed zero times if appropriate. Some variants of BASIC got
>>unhappy loops in the days of tiny interpreters, which were too stupid to
>>follow the language definition. The only language with an officially
>>unhappy loop was FORTRAN. The late lamented MS QuickBasic did it right. How
>>about FBIII?
>>
>>The LONG IF nonsense
>>--------------------
>>Ever felt embarrassed showing your source code to another programmer? The
>>LONG IF, inherited from ZBasic, curls my toes by its ineptness:-
>>
>>LONG IF flagSet
>> .. do stuff
>>XELSE
>> .. do other stuff
>>END IF
>>
>>Those other programmers snicker at the silly syntax and ask "Why LONG? Why
>>no THEN after LONG IF? Why XELSE? Why no ELSEIF?..."
>>
>>In MS QuickBasic you could write, with no perceptible strain or discomfort:-
>>IF flagSet THEN
>> .. do stuff
>>ELSEIF tr%=2 THEN
>> .. do other
>>ELSE
>> .. yet different
>>END IF
>>
>>How about FBIII? Is it going to cast off the ugly and inflexible LONG IF?
>>
>
>I find nothing wrong with FB's implementation of the for loop. If this were
>"C" it would be a different story. A for loop is based on the
>implementation on a loop based on an initiating expression, an ending
>expression and the conditional use of a STEP clause. By definition the
>initiating expression exists causing the loop to executed atleast once.
>
>EX:
>
>WINDOW 1
>
>FOR loop = 12 TO 11
> PRINT "this is a test"
> PRINT loop
>NEXT
>
>DO
> HANDLEEVENTS
>UNTIL 0
>
>
>With this example, even though the step value is not negative one, loop
>initially equalled 12 and is executed one time. I dont know the ANSI
>definition for a FOR loop but I would bet that is not only within standard
>but is the correct implementation. If we were looking at a FOR loop in C
>then the loop would not execute unless the {condition} were true.
>
>I also see nothing wrong with LONG IF. Every version of BASIC I've seen has
>it own dialectical spin. In all honesty, it took me a bit of getting used
>to but that was just a bump in the learning curve. If you want to test
>multiple conditions your could always use the SELECT statement. It's not as
>pretty but just as effective.
>
>If you want an easier conversion of other languages such as C to FB(III)
>then its reasonable and desirable. It just sounds more like you have a
>gripe than a suggestion.
>
>
>W.
>
>--
>To unsubscribe, send ANY message to <futurebasic-unsubscribe@welovegod.org>