marcos@caus-dp.UUCP (Marcos R. Della) (12/31/87)
In article <378@apr.UUCP>, las@apr.UUCP (Larry Shurr) writes: > In article <431@humu.NOSC.MIL> kent@humu.NOSC.MIL (Kent K. Kuriyama) writes: > >In article <337@caus-dp.UUCP>, marcos@caus-dp.UUCP (Marcos R. Della) writes: > >> I just got my copy of TP 4.0 about three weeks ago and have been busy > >> converting my BBS system from 3.0 to 4.0. Unfortunatly I ran into some > >> problems and after several calls to Borland, ran into a brick wall... > > >The lack of overlay's in TP 4.0 is a very serious shortcomming. We have > >a 3.0 application that has 30 screens that are all overlayed. Under 4.0 > >all of these screens would be memory resident increasing the size of the > >program by 120K. > > It may also be possible to solve the problem by using an overlay link > editor. I have used Phoenix' Plink86 successfully (with Lattice C, this > was much too long ago for Turbo Pascal). The successor to Plink is > Plink86 Plus. I see in a recent PC-Brand advertisement that Plink86 > Plus lists for $495 and is sold by PC-Brand for $326, creating the > spectacle of a link editor whose cost is many times that of the compiler. > Not such a great deal, hmmm... too bad. As I recall, Plink86 was a > traditional link editor - somewhat difficult to use, but compensating > with useful features and also capable of linking extremely large programs > if somewhat slowly. Maybe Borland should consider a new product: Turbo > OVLink? Maybe Phoenix should consider Borland pricing policies. > > regards, Larry After a few more phone calls to my favorite company, I managed to find out some more information. Borland is currently working on another version of TP4.0 (4.0A?) that will include among other things, overlays. This certainly will solve my problems, but the expected due date isn't until around August through December of 88. A little time away. Oh well. The technician recommended that I look into some of the third party software vendors as there are two or three companies that he said had produced TPU's that allowed for overlays. Also this brings up something else that has been bugging me about the new turbo pascal. I like the ideas of the TPU files as it really makes compiling a heck of alot easier. The problem that I have with them is that all of a sudden, there is a influx of TPU files going around in public domain but without the source code. This is really infuriating as I cannot see exactly what I am doing with the code and I can't tell exactly what the code does that might effect something else I wrote. Unfortunatly, the people who distribute these files are not always the best documentors in the world. I have programmed around the method of not bothering to look at the code, just use the call, but there has been massive coordination and documentation going into the usage of these things. You know exactly what the parameters are that go into the procedure/function, the ranges allowed, the effects of going outside the ranges or not including values at all, you know the output and its format along with the error messages/values produced by the procedure. I know that its a pain to write all this out, but when you don't include the source code of what your routine is doing, how in the world is someone else going to debug your mess if it effects something in his mess? An example of this was a set of windowing routines that were/are going around as a TPU file. These provide a really nice set of window routines, but I had to toss them out as I found out later that they messed around with some of the memory that I REALLY needed for my code and that the windows were bombing out everything else. Finally had to re-write the windowing routines myself from scratch, just using the output of the other design as a guideline in order to insure that nothing I needed got messed up... What to do about this all? What can I say, the TPU is the greatest thing since the Toilet Paper Unit, but there are some definate drawbacks. The only thing that I have put into TPU files for distribution are serial numbers and the routines to insure that whatever I'm writing won't work without them. That way I can keep my name on the program, have a date stamp go out with it, a version number, and keep track of the versions floating around and be very indiscrete about it. Only major finished programs get these, and EVERYTHING else is source code so that others may play with it and do their own thing. Who knows, maybe someone will make things better and send me a patch! I never did like restricting others from code... Are there any comments on this subject? I really would like to hear how others feel about this new up and coming problem in the PC field. Oh yea, while I have your attention, does anyone want the source code to a decient BBS system? I'm trying to give it away (shareware type of thing, but so far I've made only $0.00. Oh well, thats life) so that others can modify the code and come up with new and exciting stuff that I can use myself. Its called spread the wealth so that some of it trickles back to you... If your curious, drop me a note and I'll tell you about it. :-) -> On another side note, has anyone created a pascal mailing list yet? Marcos R. Della -- ...!csustan ->!polyslo!caus-dp!marcos | Whatever I said doesn't ...!sdsu ---/ Lt. Marcos R. Della | mean diddly as I forgot ...!csun --/ Smail:PO Box 8104 SLO,CA 93403-8104 | it even before finishing ...!dmsd -/ Tele: (805) 544-4900 | typing it all out!!! :-)