com.users.c_cpp   
   .os2.programmer.misc:1868 openwatcom.users.c_cpp:1967   
   From: Will Honea    
      
   Jonathan de Boyne Pollard wrote:   
      
   >> If this gets started, I should be able to supply some resources for it   
   >> - I have (had?) a good bit of documentation and some of the source   
   >> code for SOM 3 before that died. It will take some digging.   
   >>   
   >    
   > You have IBM's actual source code for the SOM Runtime?  That's   
   > impressive.  It's also valuable.  I've read that even IBM doesn't have   
   > the source code any more.  Keep hold of that!  Don't throw it away.   
      
   The operative term here is "some" - I doing OS/2 support back when MCI was    
   using it so there was some leverage - but most of what I got my hands on was    
   specific pieces to isolate problems as they popped up.  Basically, I had    
   enough to debug the problems we had.  I didn't give much thought about the    
   extent of the code base then as the whole arrangement was just to speed up    
   our development.   
       
   >> The better plan for building would be to use VACPP 3 rather than   
   >> Watcom until you get a build environment and working code base.   
   >>   
   >    
   > I still use MetaWare High C/C++ for OS/2 primarily.  The problem with   
   > switching to Watcom C/C++ is that my code base requires several things   
   > that Watcom C/C++ doesn't have:   
   >    
   > * anybase numeric literals (e.g. 0x2x0010_0100_1101_1001)   
   >    
   > * local functions   
   >    
   > * for() iterators that use local functions   
   >    
   > * DirectToSOM C++   
      
   The reason I brought up Visual Age C++ was simply that it was what IBM was    
   using at the time.  I fought the VACPP -> Watcom battles enough to know that    
   you wound up doing pretty much a total port of the code.  Even going from    
   the old (circa 1990) CSet compiler to Visual Age was a fight!   
       
   >> Talk about old/moldy: those archives appear to date from 1999 - 2002!   
   >> I still think PM/WPS/SOM was best model I ever worked with. Even DSOM   
   >> was more usable than most current desktop systems.   
   >>   
   >    
   > Which brings us back to the discussion of realistic short-term goals   
   > from elsewhere in this thread.  The whole DSOM mechanism isn't necessary   
   > to start with.  Just enough of a runtime for an ordinary DirectToSOM   
   > C/C++ program to work would be enough.   
      
   I don't recall if the later versions of Watcom still had the DirectToSOM    
   functionality or not.  I have an earlier version around that has it but I    
   have this nagging notion that SOM disappeared in the later versions of    
   Watcom.  To me, DSOM was important because 1. the company wanted to use it    
   and had it incorporated in some of the DB/2 related programs and 2. DSOM    
   identified and caused a lot of SOM2.x bugs to get fixed.   
      
   --    
   Will Honea   
      
      
   --- Internet Rex 2.31   
    * Origin: TeraNews.com (1:261/20.999)