Agreed. This is why .NET developers tend to develop apps in a different manner than Fox developers. A good developer always codes to the strengths of their chosen platforms.

A few comments though -

1. We've made good strides with the scalability of ADO.NET datasets in VS2005. Not there yet, but you'll see the start of it (thousands of local rows will work fine now).
2. Memory/disk spanning is one of the top things on my personal list for ADO.NET.

Also - I personally appreciate you calling this memory/disk spanning. Cursor is way too overloaded a term. For old time Fox developers like me, you may remember when the SQL folks kept saying "but that's not really a cursor" about Fox (they did use that name first). I find it funny to see folks saying essentially the same thing about .NET not calling it a cursor. <g>

yag

>Rod,
>
>>How is automatic memory/disk spanning different than a memory swap file handled by the OS ?
>
>It is very different, because the swap file is not as scalable it is used to swap process memory in and out. The automatic disk spanning feature found in cursor in VFP does not eat up process memory. So a small VFP executable that handles large cursors still processes in a small amount of memory. With ADO.NET, the process eats up all assigned physical memory and after that all the virtual memory until your Windows because non-responsive. IOW, you can't scale applications using ADO.NET this way as you can do with the VFP equivalent using cursors.
>
>Walter,