Re: ****************** Ada vs C++ Help ********************

[miano@worldnet.att.net (John M. Miano)]

In article <3226aa7e.7480949@news.redstone.army.mil>, Sarner-BA-MD@PATRIOT-ccmail.SED.Redstone.Army.Mil wrote:
>Does anyone know of some good papers comparing Ada 95 to C++.  Would
>especially be interested in any studies involving the comparison of
>executables of similiar code modules in Ada95 and C++, ie size of .exe
>and execution speed and how the code itself was fairly constructed to
>be considered comparable to each other.   I am on a project that is
>being developed in C with tentative plans to go to Ada.  It is a
>government project that may fall under the "Ada mandate".  However,
>many of the contractors seem to have a definite *dislike* for Ada.  At
>650KSLOC, any Ada conversion will surely cost $$.  They would rather
>transition to C++.  Probably because they could then recompile it and
>call it that but also since C++ does support OO (the official party
>line).  

1. Code size and execution speed are a function of the compiler and not the 
language.

2. In almost any technical comparison from the theoretical point of view Ada 
is going beat C++.  C++ is simply a compromise.  C++ wins in most of the 
pragmatic areas: tools and available  programmers.

>Also, I have seen posts here saying C and C++ depend on pointer
>arithmetic.  Could someone please elaborate on this and why it is
>considered to be "dangerous" or otherwise bad?  I could sure use some
>Ada ammo!  

This is not the case.  The languages do not depend upon it but many 
programmers do.  They are allow to because if the [] operator is not redefined 
then *(POINTER + X)  is equivalent to POINTER [X].  For that matter in C
A[X} and X[A} are equivalent.

If you are really careful you can write programs that are just as clear and 
correct in C++ as you can in Ada.  The problem is that C++ allows one to do 
the stupidest things imaginable and stupid things are encouraged.  I have seen 
books and articles on C programming that claim

for (ii = 0 ; ii < 10 ; ++ ii)
  * a ++ = * b ++ ;

is preferable to 

for (ii = 0 ; ii < 10 ; ++ ii)
  A [ii] = B [II] ;

because it is "more efficient".    

John


World-Wide-Web: The CB Radio of the 90's