Re: RDS vs Compiled?

On 3 Oct 1997 15:04:36 GMT, Douglas Wilson <dgwilson@gte.net> wrote: >Perry Dillard wrote: >> [snip] > >There seems to be a >> point where you really take hit and if memory serves me correctly it >> is when your program file exceeds about 1.5 meg and approaches the 2 >> meg mark. > >Is that 1.5 megs in source code, RDS or compiled code? >Another 'disadvantage' of compiled is that it takes significantly >more disk space. > I meant when the resulting p-code runnable file approaches this size then it begins to take a long time for the runner to load it. Although the compiled code takes more disk space, it takes less machine resources when running the program. The code segments can be shared in memory even though the data segments cannot. With p-code, each user runs his own copy of fglgo and loads his own copy of the p-code file. Since fglgo is obviously an executable itself, it can take advantage of sharing its own code, but not the p-code file it is loading. If you only have a few users this isn't much of an issue. But if you start putting lots of users on the system then you'll be looking at installing more memory. What I've always found to be the best environment is to develop with RDS and deploy compiled code. There are definitely some subtle differences that can bite you when doing this, but they are usually easily discovered if you run the same testing on the compiled code as you do on the RDS code. In some cases RDS is quite a bit more forgiving. For instance, uninitialized variables usually don't cause a problem in RDS, but 4GL will core dump every time. HTH, Perry Dillard Shop Floor Solutions, Inc. > >Just my opionions, >Douglas Wilson