Hi Jim,

Basically I agree with you. However, I am a perfectionist by nature. Right now I am working at home on the project in OH, while the working version of the application is in Boston. However, while I was documenting this project, I found couple of points, where this application could be improved. For instance, I had a range criterion control. In the current version of the application I save values in both textboxes, even if one of them is left empty. Obviously, it's an overhead and I can reduce number of records in the metafile, if I would save only non empty value. I'm wondering, why I didn't think about it before... And some other minor changes in the code logic...

I'm testing my changes as I make them, but not thoroughly... Basically, once I finish, I'm going to send the changes, I made. My former manager would decide, if they need to be incorporated or not. In the second case it would mean, that I worked into the "table" (Russian expression).

>Nadya,
>
>Personally, I would put this item at the end of the potential optimizations list. Even then, after other things might have been done, I'd look again at if these things need changing.
>
>But I also want to point out that "documenting" is documenting and "optimizing" is optimizing.
>My practise is to keep it simple, so when I optimize I opitimize and when I document I document.
>
>In my opinion, if what you want/need to do is to document, then changing **ANY** of the (active, executing) code should be a no-no. It would be fine to add/revise COMMENTS in code as you document the thing (though even that needs, of course, testing after doing to ensure continued correct execution).
>
>If you feel the need to optimize then my suggetion is to optimize FIRST, get that out of the way (completed, tested) and then do the documentation.
>
>Just one opinion.
>
>
> >Hi everyone,
>>
>>I am documenting one old project of mine and while doing it, I am going through the code and making bunch of optimization changes (wondering, when I finish, would it be worth anything? :)) Anyway, here is the situation [I'm using VFP6]: I have a form with lots of multiselect listboxes. All these listboxes have this.btcValArray as their rowsource. This array usually has two columns, the first is description and the second is a code. The number of rows in these arrays is ~10-20. I need to restore selections.
>>
>>Here is the current piece of code, which does it:
>>

>>for n = 1 to alen(.btcValArray, 1)	&& cycle over rows
>>   if at(.btcValArray[m.n, 2], m.list_arg)>0   && this item is to be selected
>>     .Container1.List1.selected(m.n) = .t. && flag list item as Selected
>>   endif
>>endfor

>>
>>list_arg is a comma-delimited list of codes.
>>
>>As you see, I'm looping through the array to set selected flag. This seems inefficient to me, if, say, I have only 2 items to mark as selected. There should be another way around... I'm wondering, if the optimization does make sense in this scenario and if yes, how would I optimize this code?
>>
>>
>>Thanks in advance.