>

>use http://TerryServer.com:Customer alias Cust
>go top
>scan for active
>   ? Cust.name
>endscan
>

>
>...is not optimizable over an HTTP connection. The fact that each scan iteration have to make a separate trip to the server means you are on a bad track. No matter how fast your connection is, the HTTP connection and requirement is not make for this kind of job. But get all the customers in a single request (select * from customer), and your closer to a real solution.
>
>Do you see my point?

I see your point! But when we started - (you're looking younger these days :) - it would take more time to copy a 100KB to a 180KB floppy than it does to download 20MB off the net today - even dialup would seem speed of light compared to the transfer machinery we had then.

Back then - backing up a 10MB XT, formatting (with /S to make space for SYS.COM:), and restoring the backup (the original,but ugly, defrag method) was an all day job!

Ken is still young - he's smart - he can do it - everything you say makes sense - but the issues are not a barrier - only a challenge. Perhaps NET USE DBFs would be more like XMLs and SEEK would translate into some king of getNode query - and perhaps the return would be an XML string - I don't know - and I don't need to understand the physics of the "transaction", but I sure would happy if I could do it!