Re: Future with Ada

["Marin David Condic" <dont.bother.mcondic.auntie.spam@[acm.org>]

Oh... I probably wouldn't do it out of concern for speed. I'd do it if it
seemed that constructing subprograms to squeeze something into some
arbitrary line or nesting limit was going to make the code less intuitively
obvious to the casual observer. Say you have an enumeration with 20 values
and a case statement handling it with 10 statements per enumeral. That's 200
lines + overhead, right? Would you create something like:

case (var) is
    when First_Enum =>
        Do_Ten_Lines_Of_Stuff ; --  I wouldn't generally object to this...
maybe
    when others =>
        Do_Another_Case_In_A_Procedure (var) ; -- I *would* object to this.
end case ;

procedure Do_Another_Case_In_A_Procedure (var : in var_type) is
begin
    case (var) is
        when Second_Enum =>
            Do_Ten_Lines_Of_Stuff ;
        when others =>
            Do_Yet_Another_Case_In_A_Procedure (var) ;
    end case ;
end Do_Another_Case_In_A_Procedure ;

And so on, and so on, and so on... Just to avoid a 200 line case statement
where and "end case Identifier;" might help you see what you were ending? I
guess I've written enough code where I've exceeded a screen's worth of lines
or had two or three levels of nesting and thought it would be useful to
identify what the "end" parts came from. If that's bad code by some
definition of "bad" then I guess I'll just have to live with the stigma.
(Bad Programmer! No Cookie! :-)

MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com
Web:      http://www.mcondic.com/


"Georg Bauhaus" <sb463ba@l1-hrz.uni-duisburg.de> wrote in message
news:a6831k$3ls$1@a1-hrz.uni-duisburg.de...
>
> yes, sometimes, highest speed dfa in inner loops, say?.
> my point was that this need not be reflected in "big" if, in a sense.
> With inlined subprograms, tag magic, ... you can do a lot to reduce
> the "textual length" of an if or case statement, without much or any
> loss in performance. At the expense of increased expressiveness,
> sometimes :-)
>