From: eriknoc@gmail.com   
      
   I guess the complexity of how you handle this depends on what the needs of   
   your code is.  You would have to weigh the pros and cons of each method and   
   decide for yourself what's best for each case.   
      
   If your code is completely relocatable and doesn't need direct access to parts   
   of itself (or data), then sure, sticking it between your program end and LOMEM   
   would make sense.  But if your code does need direct access to itself or it's   
   data, then you can'   
   t have that just floating up and down in memory... at least, not without some   
   overhead code.  So if were sticking to wanting to use SAVE/LOAD/RUN filename,   
   here's a couple more ideas I could think of.   
      
   You could have some additional code that's relocatable for initialization that   
   would recalculate and rewrite each place with any kind of absolute addressing   
   reference by using indirect ZP addressing.   
      
   Here's my other idea.  For program execution, you would have your ML code and   
   data located above HIMEM.  Before you SAVE, you would need to temporally   
   adjust LOMEM and copy of your code/data down to the end of your BASIC   
   program.  Once you SAVE, then you    
   can restore LOMEM back.  After you LOAD your program, you would need to copy   
   the code/data at the end of your program back up and adjust LOMEM and HIMEM   
   accordingly.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)