RE: [WDSCI-L] Performance - perception is King

[Vern Hamberg <vhamberg@xxxxxxxxxxxxxxxxxxxxxxxxx>]

Good point, Buck. Yet I can see how you could get the count without running 
through the entire list. At least for member count in source files. Now I 
don't think there is an object count for libraries, much less by object 
type, so this'd probably require the whole list.

About the scroll bar, yes, that's true. However, OpsNav seems to set the 
scroll bar min/max according to what it knows already, then when you scroll 
down, it changes these limits. Other products can do this, too - MS Access, 
e.g., IIRC, does not guarantee that the number of records is determined 
until you get to the bottom of the record display.

You do seem more live with this than I in this case.

Later

Vern

At 08:26 AM 6/2/2003 -0400, you wrote:
> >Is it possible that you (Toronto) are
> >populating the entire list before returning
> >control to the user? Is it possible to return
> >the first visible entity ASAP (first
> >screenful, e.g.)?
>
> This is one of those places the 5250 world differs from the GUI world.  If
> one doesn't load the entire list, how big should I make the scroll bar
> 'cursor'?  Normally, GUI apps have a big fat cursor when there are only a
> few items scrollable, and a teeny-tiny one when there are many items in the
> list.  Sure seems like a pain, but the time it takes to load that list is
> one reason the GUI people use progress bars for 'long running' processes.
>
> I think it's just something I need to get used to.  In Code, I create a
> project which has fairly tight filters.  Also, I have a strong tendency to
> check members out into a very small 'development' library, where the time to
> load every member is negligible.  So now I have several small libraries,
> each containing work in progress.  I have no idea if the concept is
> reasonable to keep editor/workbench performance high, or just another case
> of me having low standards and being willing to work with... less than
> optimum conditions.
>   --buck (at Code V5R1M0 for now...)