Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!cs.utexas.edu!uunet!mcvax!hafro!krafla!snorri
From: snorri@rhi.hi.is (Snorri Agnarsson)
Newsgroups: comp.lang.c++
Subject: Re: Eiffel vs. C++
Message-ID: <1021@krafla.rhi.hi.is>
Date: 9 Jun 89 11:19:28 GMT
References: <77300029@p.cs.uiuc.edu>
Organization: University of Iceland
Lines: 95

From article <77300029@p.cs.uiuc.edu>, by johnson@p.cs.uiuc.edu:
> 
>> Written  6:48 pm  Jun  4, 1989 by bertrand@eiffel.com 
> Responses written by Ralph Johnson
> 
>>Assertions
> I think that assertions are one of the most attractive things about
> Eiffel.  I said this to a more theoretical member of my department,
> and he made a snide remark, after which he pointed out that there are
> many assertions one might want to make about programs that the Eiffel
> assertion language will not let you make.  This is a bogus argument!

Well -- I agree with that argument.  However I agree with you that assertions
are good (at least they are better than nothing), but I think well
formulated preconditions and postconditions, written as documentation, are
necessary and sufficient.  Assertions can serve as scaffolding for testing.

>>Garbage collection
> 
> This is both a major advantage and disadvantage.  Would you advise
> us writing our operating system in Eiffel instead of C++?  What
> impact does your garbage collector have on real-time software?

In my opinion garbage collection is the single most useful feature
of a programming language.  Without garbage collection polymorphism
is severely handicapped and programmers tend to shy away from clean
designs because of the restrictions imposed by the lack of garbage
collection.  Consider the following C function:

     Usage:
       x:=plus(y,z);
     Precondition:
       y and z contain integers represented as zero terminated
       character strings of digits.
     Postcondition:
       x contains the sum of y and z, as a zero terminated
       character string of digits.

If C had garbage collection we could write code such as:

     x=plus(x,plus(y,z));
     ...
     y=x;

Instead we have to write as follows:

     t1=plus(y,z);
     t2=x;
     x=plus(x,t1);
     free(t1);
     free(t2);   /* assuming that the string pointed to by t2
                    is not pointed to by any other variable   */
     ...
     y=strcpy(x);  /* unless you want to risk multiple pointers
                      to the same string, in which case you
                      would need to be VERY CAREFUL when freeing
                      that string                                */

It is often said that garbage collection is too costly, but I suggest
that in the above example the solution with garbage collection is
probably more efficient, both in programmer effort AND in
the speed of the resulting program.  By the way, if garbage collection
is too expensive, how about 'malloc' and 'free'?  Do you think those
routines are less expensive?  Why do you think so?

We should also consider the effort needed to document memory management
issues for our (reusable) modules.  If garbage collection is not available
we get two bad effects:

  1) The effort of designing a module for a given problem,
     PLUS the effort of documenting all the memory management issues
     needed to use the module is too great.  Therefore you do not
     design or use modules.  Or perhaps you use them without
     adequate documentation, leaving a can of worms for future maintainers.

  2) You put a lot of effort into designing, coding and documenting
     modules, including all the memory management issues the user
     needs to know about to use the modules, so as not to get memory
     leaks or heap smashes.
     The design document for the modules become so huge and complicated
     that nobody uses the modules (or, perhaps, the modules are used,
     but incorrectly).

In my opinion garbage collection is the single most useful feature
of Prolog, LISP, SmallTalk, Eiffel, CLU, etc., etc.

How many would use the above languages if they did not have garbage
collection?  In my opinion, very few.

> Ralph Johnson
>>Bertrand Meyer
-- 
Snorri Agnarsson		|  Internet:	snorri@rhi.hi.is
Taeknigardur, Dunhaga 5		|  UUCP:	..!mcvax!hafro!rhi!snorri
IS-107 Reykjavik, ICELAND