Re: some questions re. Ada/GNAT from a C++/GCC user

[emery@grebyn.com (David Emery)]

On the topic of 'named things', one area where Ada could have improved,
in my opinion, is to allow an optional name at the end of a record
declaration. I've seen some code with a lot of relatively complex
record declarations, where this would have helped readability.  As a
coding convention, I usually provide the comment after "end record" when
the record declaration is "long" (greater than 10 lines, depending on 
context).  

Here's what I would have liked:

   type Very_Complex_record (Discr1, Discr2 : integer; Has_Blotz : boolean)
      is record
         f1 : t1;
         f2 : t2;
         case has_blotz is
            when true =>   
               case Discr1 is
                  when 3 | 6 | 9  =>
                     Mungification : Whatever;
                  when 42 =>
                     Life_The_Universe_and_Everything : Boolean := False;
                  when others =>
                     case Disrc2 is
                        when 69 =>    
                           null;    -- what did you expect??
                        when 100 => 
                           George_Burnes_Cigars : Stogies;
                        when others =>
                           Shoe_Size : natural;
                     end case;
               end case;
            when false =>
               Other_Stuff : Some_Other_Record (Sex => true);
         end case;
      end record Very_Complex_Record;  -- not legal Ada, at not least in
               -- this millennium...

               dave