From: muta...@gmail.com   
      
   The more I look at the 8086, the more I am happy   
   that segmentation was the correct technical   
   solution to cope with machines that still had a   
   16-bit processor and registers, but had more than   
   64k memory available to them. I just would have   
   made the segment shift flexible instead of telling   
   everyone that "4" was being set in stone.   
      
   I believe they made machines with a 32-bit x86 but   
   more than 4 GiB of memory too. Or there was at   
   least the theoretical option to do so. And I think it   
   would be good to set up the required infrastructure   
   for this generic problem in computer science   
   regardless of actual processor or actual number of   
   bits in registers or shift value.   
      
   It seems to me that the tiny/small/compact/medium/large/huge   
   would be retained.   
      
   The C language doesn't need to change in order to   
   accommodate this. The onus is on the C programmer   
   to select the appropriate memory model.   
      
   It seems that only a fairly slight adjustment to   
   MSDOS is required to accommodate this.   
      
   And then a slight adjustment to Windows to support   
   the extra API calls required by 32-bit huge memory   
   model programs who need to know how to adjust   
   the segment register. Which may be a selector   
   instead???   
      
   BFN. Paul.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)