>In DOS days I could capture on-screen text by reading "video memory" locations.
>
>Does anyone know of a way to accomplish such a feat in a Windows screen? In other words, capture text from a running program (Non-VFP, Non-HTML).
>
>I know screen readers can convert windows text to speech, but I don't need the speech, just to be able to capture the text itself from a known screen layout. I will then dump it in data tables.
>
>Anyone have any leads on techniques?

I would check out the http://www.news2news.com/vfp/ site for an excellent resource on Windows API routines. Subscribe for the small amount to get actual VFP examples showing how the calls work.