>Our company has a 15 year old, very large DOS based system which consists of a main module and many add-ons. We want to convert it fully to VFP (i.e. DBC, report forms, etc.) but the reality of what that will take is starting to sink in. What we're wondering is how feasible is it to do a phased migration?
>
>For example, if we leave our existing tables as free tables and use local views to access/update them, then we can leave much of the DOS system in tact while we take a more gradual approach to the upgrade.
>
>Is there anyone who has any experience with this sort of thing that can give me some advice before I get too far along with the project?
>
>TIA

I have been in this position several times in the past. Incremental upgrading is possible. The way I recommend you do it is this. First, do an OOA/D process. Analyze what you want to do with your windows product. In 15 years, requirements may have changed. Then, design your classes. When you are done with this process, you will have a map of where you want to be with the system in the long run.

At the same time, do the minimum to get ur application running under VFP. That might be as simple as recompiling the code and maybe defining the _screen font to something that works with the DOS coordinate scheme. Then, peice by peice, you move the application into the new OO model.

To be honest, I tend to recommend a full rewrite... I think that works best. I don't know if you and your business can afford the time, though.

Be careful, though. I have seen rushed upgrades to VFP that have had major problems because of a lack of planning. The more you plan for the upgrade, the better of you will be.

One quick tech note... you can leave ur tables as free tables up front and kabash the .DBC to start with. You don't *have* to use local views.

HTH