>We are building a medium sized product with VMP. This is both my client's and my first venture into VMP, and we are seriously disappointed so far with form load speed. Initially, our forms were built on tables using the native filtering techniques built into the framework. I was convinced that converting everything to parameterized views would help things, but after converting some forms, we have found that the difference was minimal.
>
>I am hoping to get some encouragement and speed tips, as we are afraid we might have to eventually abandon the framework because of speed issues. We can tolerate the longer time that it takes for the app to load, but some of these forms are pretty high volume and will be opened and closed all day long, and we are afraid that the speed will be unacceptable to customers. Hopefully this is an issue of technique, and not a fact of life with VMP.
>
>I know Barbara and a few others here have built sizeable apps with VMP, can you please offer some insight? Are you happy with VMP performance? Thanks.

Erik, how fast are your computers? While VMP is noticeably slower than a 'base class' version of the same form our users aren't complaining too much. We've cut some of their time by making sure that each form loads data matching the previous form, so the user can follow the same data across multiple forms. These forms will also update when the SHOW() method is activated, so users just minimize a form or let them go to the back of the 'stack'. When they need it, it's automatically updated to the 'real-time' data.

One of my apps has 6-7 page pageframes, and I increased the loading speed by using VMP controls (text boxes, combos) but not labels, page frames, option groups or anything else elaborate except Grids. I was demo-ing it Friday on my 486 laptop and the only problem was where I was using a parameterized view - that took a couple seconds to refresh with the correct record (2-3 out of about 150). I haven't double checked my indexes on the base table - that may be part of the problem.

Our big app has 200MH Pentiums as the minimum acceptable machine. We expect our testers/trainees will find this is fast enough once the app loads. The complaints have been minimal and they're all working on 133s or 100s at this point. Network speed is much more a factor here.

If you're working on a network, be sure your EXE is on the user's hard disk and so are their temp files.

HTH