>Hi Harry,
> I suggest you rethink your strategy in converting the 2x app to VFP. It would probably be a lot easier in both the short term and the long term to do a better conversion up front rather than what I gather you're doing--a small conversion (by small I mean sticking to 2x concepts and ways of doing things) and then planning to go back to make it more VFP-like.
> I've done a few of these conversions and IMO you will save time and money by doing it closer to properly (if not completely properly) the first time.
> Regards, Chaim

yes, I understand the theory!

However, I wasted the best part of a year trying to do it that way and it is utterly impractical when dealing with a large application already in use by about 150 users. In particular no one ever managed to point me in the direction of using a data dictionary where the data is in a user selected path (which can and does change depending on the task they are performing) and although I managed a bodge or two to get it almost working, it was never satisfactory. I got absolutely nowhere and pissed off a number of my clients in the process.

By doing it the way I am doing it, I've achieved in 3 or 4 weeks more than I managed in the previous 52. About 30% of the application is now in vfp form and working pretty satisfactorily. So I'll stick with this path until the app is completely in vfp - then I'll worry about taking advantage of the frilly bits and restructuring from the ground up. At least I'll then have my clients working in a 2000 proof environment and able to use the program almost exactly as they were used to doing in dos - which means minimal new learning curve.

Most important, by doing it this way, when the system crashes, because the vast majority of the code is my own, I have a good chance of figuring out what to do about it - in the process learning a little bit more about vfp. Trying to start from square one requires learning to many new tricks at one time - at least for my limited brainpower!

Regards

Harry