Re: 4GL to ESQL/C

>From: apatanka@magnus.acs.ohio-state.edu (Atul Patankar) >Subject: 4GL to ESQL/C >Date: 8 Oct 1993 17:14:51 GMT >X-Informix-List-Id: <news.4546> > >The organization I work for is investigating the benefits of >migrating the code from 4GL to ESQL/C. I think the two largest gains >would be a 30% to 40% drop in the size of the executables and a 20% >gain in the throughput. Can someone comment on other gains/losses? Your biggest loss will be in modifiability. Modifying ESQL/C is NOT easy. I'd say that is likely to outweigh any gains. You'll also lose the advantages of the screen and I/O handling, the report handling, the Interactive Debugger, p-code for speed of testing. I wouldn't do it, but I can hardly claim to be unbiassed. >This also means that we will have to come up with our own curses >routines, as the 4GL curses routines for forms and windows would not >be available with ESQL/C. Does anyone know of an archive for >notes/hints/code for such ESQL/C routines? For the application we >are working with, I believe we can come up with a better screen >interface than the one that 4GL offers, but hints/pointers from this >group have helped a lot in the past. ESQL/C is independent of screen handling (orthogonal to screen handling). If you really want to do this, then investigate 3rd party screen painters and such like -- don't go inventing your own unless you really have to. And consider your range of platforms very carefully -- when are you going to be changing any of them? And also, are you going to need GUI interfaces? If so, your choice may be very different from what you use if you will never need GUI? But remember, "never" is a dangerous word, and almost never correct. Yours, Jonathan Leffler (johnl@informix.com) #include <disclaimer.h>