Thread: Is program salvageable?

In 1990, I designed a small calculator-like program written for DOS. A friend, who has passed away, coded it in C++. I have the complete source code for it with these exceptions. Let me add I know absolutely nothing about C++ other than I can spell "compilor" (just kidding). I do know the source code contains only one *.h (header?) file which begins with

#include <stdio.h>
#include <cwuser.h>
#include <cwdos.h>
#include <cwform.h>
#include <cwarch.h>

The five files I do not have. I know the stdio.h file is available but I can find no reference to any of the "cw" files.

I want to upgrade my program to Windows interface (I hope you understand what I want to do since I don't know the correct terminology) and make some minor "label" like changes and perhaps, rearrange or remove certain modules, such as remove the "Length Calculator module" and the like.

Am I out of luck without those four files?

Thank you very very much for your expertise.

Maybe.....

By chance are the initials of the original author cw?

Not that cw is of any significance in finding or recreating them. Just curious.

It may be possible to reconstruct the contents of those headers by a bit of deduction. If one commented these includes out of the compilation, the compiler would complain about functions without a declaration, access to member variables without declaration, and it's probably that these could be used to reconstruct these declarations. Part of this depends on how much the author used inline function in those files, or if there are any template classes in those files.

This would, however, require a fairly knowledgeable C++ programmer - not necessarily an expert, just someone experienced enough to interpret the compiler complains and generate the required declarations.

You could get a fair headstart, too. Depending on the complexity of the application, it may be workable to skim through the source to identify class names obviously referenced as member functions for which there are no class declarations, and fill that out first. This would go a long ways toward the reconstruction. A compiler pass would then pick up any member variables referenced that are missing, completing the major portion of the work. After that, depending on what complaints remain, various judgment calls to make about possibly missing inline functions.

This could be obviated, however, by your intent to translate the work into a GUI application. If you could deduce, by reading the code, that important features were implemented as member functions, and are therefore present in the CPP files you have, you could treat that a body of functions to be applied to a new design. Redesign the user interface, interpreted as a GUI interface, and apply these functions to that design - a two for one kind of approach (reconstruction and translation into GUI). This would require a more expert programmer.

You are all wonderful! Thank you very much. I didn't expect such a quick and thorough reply. It seems most of what was said is way over my head (I did get a grasp of the concepts, though).

It would seem better that I try another approach. Now, if I had all the *.c files...and a working copy of a relatively small program that's really comprised of about 7 very small routines (modules), would anyone care to speculate what it'd cost to accomplish my goal. I realize that's impossible to really answer so let me ask: Under $1500 (I"m hoping I hear under 1k)?

The next step, depending on your input, would be to post my request in the "hire" forum.

Thanks, again.

I was almost afraiid to check back after realizing what a stupid question I asked. MacGyver, your reply was quite encouraging. I've been in contact with an outsourcing company in India and they want to charge $1300 to replicate a single module of the program based on 240 hours of programming. That seemed outrageous, but since I'm not a programmer, I couldn't tell for sure.

I'm baffled by the missing *.h files I mentioned in the beginning. My friend who coded the program told me he gave me everything I needed, except a compiler, to modify it in the future.

I gotta check what "child widget" is...

I agree the rate is good. The total hours seem excessive. Salem and swgh mentioned stdio.h is a C header, not C++. Gulp...I thought the program was written in C++ but I don't have a clue. Here's a small piece of a file:

Code:

#include <dir.h>
#include <stdio.h>
#include "ac.h"


//#include <sys$msg.mlh>
//#include <sys$err.elh>


char *buffer;

extern LIBRARY_INFO *library;


char *ReadFile(char *fileName)
{
   FILE *fp;
   int blocksRead;
   char *workBuffer;
   struct ffblk rfile;

   if (findfirst(fileName, &rfile, 0xff)) {
      DisplayErrorText(E_MISSING_FILE, WARNING, fileName);
      return(NULL);
   }

   fp = fopen(fileName, "r");
   if (fp == NULL) {
      DisplayErrorText(E_BAD_OPEN_READ, WARNING, fileName);
      return(NULL);
   }

   if (rfile.ff_fsize >= (long)(MAXBUFFER_SIZE)) {
      DisplayErrorText(E_2BIG2VIEW, WARNING, fileName);
      if (fclose(fp)==EOF) 
         DisplayErrorText(E_FILE_CLOSE, WARNING, fileName);
      return(NULL);
   }

C or C++??

OOPS. Now I know why the two dudes with the Fedoras and trenchcoats just showed up at my front door.

Aside from me moving this out of this forum, does that change the overall picture of my original objective?

Thank you for point out my studipity. But, I have an excuse: I was born confused....and it's only gotten worse...