Revolution speed sucks?

[Alex Rice]

On Thursday, August 7, 2003, at 08:16  PM, Björnke von Gierke wrote:
>
>> repeat for each line theLine in theWinners --note: NOT "each line 
>> theLine in fld winners"
>
> This is wrong. Actually when you do a "repeat for each" loop, then you 
> generate automatically a variable at the beginning of the loop and 
> then use that. Hence it is impossible to change a container directly 
> with that loop (putting something into "theLine" does not change the 
> appropiate line of "theWinners"). When you generate another variable 
> before the loop then that will only use more memory. (I tested that 
> once and speed is equal for both cases)
>
Björnke, I don't understand what's wrong about it. Don't understand 
what you mean about using more memory. "repeat for each line" is 
supposed to be faster than any other loop form in transcript. Maybe you 
can write a counter example to show what I'm doing badly?

If one needed to modify theWinners while inside the loop, and maybe it 
wouldn't be necessary, but supposing one needed to, wouldn't one just 
maintain the line count in parallel, like this?  I think I've seen this 
idiom used before:

put fld "winners" into theWinners
put 0 into tLineNo
repeat for each line theLine in theWinners
  add 1 to tLineNo
  if "target" is among the items of theLine then
     put "found target here" into line tLineNo of tWinner
  end if
end repeat

I would be glad to be corrected- I'm still pretty new to this 
optimizing transcript stuff.

Alex Rice, Software Developer
Architectural Research Consultants, Inc.
http://ARCplanning.com