>Interesting info on the future of MS 10 but how does it solve the SMB 1 vs SMB 2 issues?

it does not. But as there are enough cost-free DB backends, depending on dbf as multiuser common datastore is not really necessary (except for cases where control of installed programs is handled by a group you have no control over and security is high, leaving you no way no fiddle with the backend)

> Converting to XAML - not so easy?

When I tried to convince that porting vfp to Python is not too costly I slapped together a converter from vcx to either Winform or WPF via Ironpython, moving the info from vcx into Ironpython dict and having a vfp - similar subclass of the controls in a week. The list / grid controls were given Ado.Net datatable "cursors" to bind to, which were fed either at runtime where not possible to build at compile time in code (mostly piping over vfp cursors, tested/timed for up to 1000 rows).

Worked remarkebly well for such a short amount of effort - could convert a couple of test screens into a corresponding Dotnet screens. High probability of edge cases and small behaviour discrepancies not found as I could not feed it to a seasoned test team, but from well layered architecture nothing to hide from IMO. Vfp itself has enough inconsistencies you run into if you attempt to max out every customer wish.

Writing out to XAML not attempted, but not expected to be problematic. But or screens are writen with "V" perspective in mind: just display, nothing fancy involving timing or instatiating order, discrete code in separate classes if dependencies of more than a line exist

regards

thomas