From: engelbert_buxbaum@hotmail.com   
      
   RadSurfer wrote:   
      
   > Does anyone know of any URL's that show Benchmark tests of Borland   
   > Turbo   
   > Pascal, and Borland Pascal Compilers? This would prove most useful.   
   > Thanks!   
      
   With benchmarks you can check very different things:   
      
   *Ease of programming*: In a University programming class, students were   
   given the free choice to do their final assignments in either C,   
   Turbo-Pascal or Modula-3, with Modula-3 being suggested. Most   
   programming beginners followed that advice, and 3/4 of them completed   
   their assignments on time. In TP, that number was 2/3. Those students   
   who choose to do their work in C usually described themself as   
   experienced programmmers, however, none of them was able to produce   
   their assinment on time. When given extension time, 1/3 of the students   
   finally produced a solution, but the code quality was much lower than   
   that of that obtained with the other languages (taken from the Modula-3   
   homepage).   
      
   *Compile time*: Structured languages like Pascal allow the construction   
   of efficient one-pass compilers, simply because the compiler knows a lot   
   about the code in advance. C-compilers are usually multi-pass, in the   
   first pass the compiler has to analyse the code, rather than translating   
   it. That was one of the main reasons why Borland used a Pascal-dialect   
   for Delphi (described by M. Starke, one of the original Delphi   
   programmers, in "Borland Delphi", Munich (tewi) 1995, Foreword). Rapid   
   compile times are of course essential during programm development with   
   the many code-compile-test cycles.   
      
   *Run time of the finished maschine code*: That comparison is difficult,   
   because no single-pass C compilers and few multi-pass Pascal compilers   
   are available. When the output of a highly optimising multi-pass C   
   compiler is compared with a single-pass Pascal compiler (that is of   
   course a comparison between apples and pears), programs from Pascal-code   
   often run marginally slower than those produced from C. However, that   
   small difference is usually irrelevant with the high-powered computers   
   today, where the computer waits for user input most of the time. I have   
   however repeatedly stated that a multi-pass option in Turbo-Pascal or   
   Delphi for the final shipment-ready product would be a nice thing, if   
   only to show those C-freaks where the hammer hangs.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)