dvk@mi-cec.UUCP (Dan Klein) (02/27/84)
As an answer to George Rosenberg (and others who have sent me mail): My examples of the C "neatener" versus the Bliss "optimizer" were intended as an example to show how certain languages generate better code than others, and to enumerate some improvements I have found to be possible and easy (versus those that are hard) to make in Berkeley C. I did NOT want to compare Bliss to C, but chose them since C was good and Bliss was better (at generating efficient code). I did not want to say "if I was a compiler, I would..." but chose rather to say "there is a compiler that does...". I have my opinions about the relative merits of Bliss and C, and they are not relevant to the examples I posted. All that was being done was to show "this is what BSD C does, and here is how it can be made better". Comparing Bliss-10 and Bliss-36 would not be worthwhile, as most people don't have a PDP-10, but quite a few have VAXen. Similarly, comparing an optimizing C compiler with a non-optimizing one would be a pain, since I do not have access to anything except 4.1c BSD (as I suspect do a majority of VAX users). The final thing I was trying to accomplish (and I admit that I was being a bit high brow in doing this) was to give some people a little exposure and information on a subject that they might not otherwise have access to. Having received a large quantity of favorable mail (yes, I got flames too), I think I accomplished what I set out to do. -Dan Klein, Mellon Institute, Pittsburgh (412) 578-3382