[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

agenda for the June 25 R5RS meeting: Proposed Objection

   Date: Mon, 27 Jul 92 13:57:07 -0400
   From: David Kranz <kranz@lcs.mit.edu>
   Sender: kranz@ORBIT.LCS.MIT.EDU

      I believe that the adding of a OOP system of ths form to Scheme is a
      bad idea.  To me it is yet one more step toward the CommonLisp-ization
      of Scheme.  It is one more added feature that does nota appear to be
      implementable in terms of a simple set of core Scheme primitves.  Ode
      for the days when one reasonably intelligent individual could
      implement a reasonably efficient and complete scheme in man weeks or
      months, not man years.
      Morry Katz

   I think you are putting the cart before the horse.  I would hope that the
   purpose of scheme is to provide what we think is a good thing for people 
   who *write programs*.  This may or may not be compatible with a small
   implementation effort.  It is reasonable to argue that a proposed feature
   could be implemented efficiently from other primitives or that it is bad
   in some way from a programming point of view, but to reject it *solely*
   because it makes more work for implementors is a bit absurd.  The swipe
   at CommonLisp is sort of like trying to discredit someone's argument by
   saying that they are a Communist (or perhaps, today, a Liberal).  Note
   that I am not favoring this particular proposal.

You miss my entire point.  Maybe it was unclear.  I see no way to
implement this proposal in terms of the conceptually and
implementationally simple core of the current scheme language.  I
furthermore see no compelling argument for this new functionality that
convinces me that it is worth increasing the size of the core to
facilitate it.  My comment about Commonlisp was not intended as a
swipe, but as an observation.  As the core of a language grows, it
becomes much more difficult to implement.  This adversely effects
those of use attempting to do language work using scheme.  I could
never do my reasearch in Commonlisp.  It would take me an extra 2
years of implementation effort (if I didn't just give up first). 

In order to convince me that an addition to scheme is desirable at
this point, one has to satisfy at least on of the following:
1)  It can be implemented in terms of the existing core.  Most such
    proposals belong in a library, not in the language, IMHO.
2)  The functionality offered
    a) cannot be achieved in existing scheme,
    b) has been well researched, implemented, tested, etc., and has
       been demonstrated to be feasible to implement, useful, etc., and
    c) can be added in a semantically clean fashion.

Furthermore, additions should, if possible, not conflict with
programming styles, extensions, etc. that are in common use by
significant fractions of the Scheme community.  (e.g., an OOP proposal
that couldn't work with first-class environments or broke the types of
analysis tools that individual at Indiana have developed over the
years would be suspect.)
Morry Katz