home bbs files messages ]

Forums before death by AOL, social media and spammers... "We can't have nice things"

   alt.os.development      Operating system development chatter      4,255 messages   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]

   Message 4,070 of 4,255   
   Paul Edwards to Paul Edwards   
   Re: commercial competitor to windows   
   23 Dec 23 19:36:52   
   
   From: mutazilah@gmail.com   
      
   On 22/12/23 19:57, Paul Edwards wrote:   
      
   > It's one of these "forest for the trees" things.   
      
   Actually, I don't think it is a case of this at all.   
      
   > There is not a lot of code involved to "glue"   
   > a (limited target) Win32 executable so that it   
   > can run on OS/2, which is still commercially   
   > supported in the form of ArcaOS.   
      
   Actually, I believe it is working (e.g. on Linux   
   as the latest target) because I have managed to   
   combine two C libraries into one.   
      
   It is not obvious that it is possible to do that.   
      
   In addition, with msvcrt.dll in particular, for   
   fprintf(stdout ...) to work, it required mapping   
   of some standard names.   
      
   This was done by Alica years ago for producing our   
   equivalent DLL, and this code needed to be activated   
   in the Linux etc environment.   
      
   So yeah - it's a limited amount of code, but anyone   
   would be forgiven for thinking it was nominally   
   impossible to combine two different C libraries   
   into a single executable. I have a "runnum" variable   
   that knows it needs to switch parameter format for   
   numbers above 1 too.   
      
   So yeah - the incremental approach to getting this   
   working (including going via 16-bit systems, and   
   trying to get the abstraction right) - was part of   
   the process, not unnecessarily convoluted.   
      
   In hindsight there isn't a lot of code involved   
   and it's seemingly obvious.   
      
   But quite frankly, as per this correction - I'm only   
   vaguely aware of how it works at all. And at no point   
   during the construction was I 100% sure the various   
   steps would work. I had no idea whether there was   
   a show-stopper I hadn't thought of.   
      
   With the 16-bit in the path, one of the potential   
   show-stoppers was the ds pointing to the DGROUP,   
   but lo and behold, Microsoft C 6.0 and Watcom   
   both had an option for ds to not be assumed to   
   be pointing to DGROUP, requiring the ds to be   
   reloaded.   
      
   Someone once told me (I don't know if it is true),   
   that something like pennicilin was actually invented   
   in Africa. They had stumbled across a sequence of   
   steps - some of them unnecessary - to produce the   
   required antidote (by luck). So it's probably   
   something closer to that. There were a huge number   
   of "I wonder if this works?" along the way. Maybe   
   a lot of those questions had answers already known   
   to others, but the only answers I ever got from   
   others around at the time were "no, the format is   
   incompatible so it won't work" or something along   
   those lines. Never once did someone say "of course   
   it will work - how the hell do you think xyz works?".   
      
   So I never had any confidence that any of the   
   numerous steps would work.   
      
   Thus I don't think it is fair to say that I   
   couldn't see the forest for the trees. It's more   
   that I managed to grow one tree on top of another   
   tree and the entire forest didn't collapse as it   
   was always in danger of doing.   
      
   A bit like S/380. It's nominally impossible to   
   run a 31-bit program on a 24-bit OS. You could   
   say "oh - dos extenders do that", but it's not   
   the same, and before it was proven to work, there   
   wasn't one person saying "of course it will work -   
   it's just a simple dos extender", but there was   
   a gaggle of people saying it wasn't possible.   
      
   Basically it is nominally impossible to put 31   
   liters of water into a 24 liter container. But   
   is it? If the rules used the chemical equation   
   H2O, so that freezing wasn't prohibited, you   
   can indeed delicately balance 31 liters of ice   
   in a 24 liter container.   
      
   Or in my case - 2**31 liters in 2**24 liters -   
   much more useful.   
      
   BFN. Paul.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]


(c) 1994,  bbs@darkrealms.ca