>Mike,
>
>>No, my problem was not just with the names. Maybe I should have capitalized it or something when I said operations based on record number or position are contrary to the core principles of SQL. That whirring sound you hear is Dr. Codd spinning in his grave....
>
>Whatever... I guess Codd would have sucked down the entire data table every time just so he could page it in code isntead of using a positional subselect with top n based on where and order by? Maybe purer, but not very efficient (or real-world). Or he could have used Fox tables with moving the record pointer and spin even more. :-)


Of course there are sometimes valid reasons to stray from relational rules for performance reasons. This one just seems especially crude. The unaware may think this is the first way to do things.

Not sure what you mean by sucking down the entire table every time. Why would you want to do that? Read it once and then loop through the data in an interface with the GUI layer.