Re ... In fact when I start out a new app, I: - write myself a brief [what the app is, what is should do, with a few ideas on how] - draw the interface and pricipal windows - write the manual [or sketch it out in such a way that I know what I should be implementing and how] - conceive the data structures [this part allows me to determine the sort of records that I'll be needing internally, for saving data, for cancelling, for passing messages internally...] - coding Erm.. thanks for the gnomic utterances... Whilst I agree with you in principal, _I_ don't know everything when I set off. In fact I still don't know when I have arrived?. (I don't know everything that I can do with FB, or the toolbox, or what 'improvements' sic will occur to me in process.) I think a lot of people using FB are autodidacts, like me, and the scope of what we can achieve is constantly changing. For me program design is similar to many other design problems in that it involves successive generation/evaluation loops while a solution space is defined. I call stickonstack a couple of times, and HandleMangle dozens of times. Somehow, you've got to set the parameters that you send to a function, whether they are part of a Parameter/Record Block, or not. For me this method is actually tidier, but I would welcome alternates. Top Down or Bottom Up Ah well Bottoms Up Ian PS I have not had any trouble with Fn stickonstack. I get the warning, but it works fine. PPS I sure am glad I threw some petrol on this thread.. PPPS I'll stick your methodology on the wall. Its a good guide.