I am converting a large project from Paradox to IBO (using the TDataset equivalents).

The software imports and processes large quantities of ASCII data.

So far, the IBO version seems to run at about 1/5 the speed of the original Paradox. I am sure that this isn't how it was supposed to be!

Does anyone know of any obvious pitfalls that people usually drop into?

Neville Richards


[Non-text portions of this message have been removed]