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

RE: RE: Jinx's EVAL proposal

>From: oz@sis.yorku.ca
>Status: R
>In a message dated Mon, 18 May 1992 10:26:29, Bill Rozas outlined
>a "compromise" proposal for eval. It seems to me that this is basically
>the first installment of a two-part effort to get first-class environments
>into a R^nRS, sort of like laying down the tracks before the locomotive
>rumbles by. I am *not* suggesting that there is anything wrong with this
>approach, but I would prefer to see this proposal be considered within
>its appearent context, rather than as if it has been made independent
>of the particulars of first-class environments.

This would indeed be a disturbing situation were it in fact the case;
however, it seems to me that you are very far from the truth. My
office is but a few doors down from Bill Rozas and Chris Hanson, who
are the movers and shakers of MIT Scheme these days. I have had long
discussions with both Jinx (Bill) and Chris about first class
environments. I believe I am accurate in saying that there is a good
healthy amount of doubt about the efficacy of first class
environments. To paraphrase Chris, first class environments seem to be
too powerful, too much bang for the buck. They complicate a great many
optimizations you would like to be able to make and they give you too
little in return. This is a sign that perhaps the proposed benefit of
first class environments could be better gotten through some other
less powerful mechanism or combination of mechanisms. More careful
research is needed on this issue before any concrete proposals are
made. Indeed, first class environments may well be dropped in the next
``offical'' major release or MIT Scheme.

In short, the people whom you accuse of trying to slip in a hidden
agenda for first class environments are, in my experience, the two
least likely to do so. They recognize that first class environments
were, from the outset, an experiment. They played with it, learned
from it, and are re-assessing their position on it. They are being
very open and honest in their endeavors in the finest academic and
scientific tradition. Had you solicited their opinions on the topic,
as I have done, you might see things differently.

Be that as it may, I put to you that it is their desire to explore new
language mechanisms which motivated them to so passionately insist
that any standard contain wording that allow them to experiment with
novel language concepts without being relegated to ``non-compliance''.
This is also why they occassionaly bemoan tha fact that standardizing
Scheme may well have been the first step toward killing it, and why
they sometimes feel that the RnRS-authors list has reached a point of
impasse on most interesting and/or important issues: they want to
experiment with novel, experimental, unproven, research features and
understand them deeply before rushing into proposals to mandate that
all other implementors unquestioningly adopt their whims.  They do not
want to be discouraged by paranoid observers who are uncertain about
their research directions. Fear of the unknown is the greatest killer
of new ideas. Nor do they want to be badgered into accepting the
research of others without personal experience and careful
consideration. Patience is a virtue.

I believe that Jinx's EVAL proposal was exactly what it purported to
be: a proposal for bringing EVAL into the newest language report--- no
more and no less. I believe its merits and shortcomings should be
debated openly on that basis and that basis alone. I believe it is
counterproductive to imply, even politely, that some alterior motive
may be at work here.  The proposal either stands on its own or it
doesn't. If you suspected that there was some other motive at work,
you could easily have asked Jinx, either privately or on the RnRS list,
if this was indeed the case. An example of just such a less destructive,
non-accusative query might have looked something like:

    I am uneasy with your EVAL proposal because it seems to me to be
  the first step along the slippery slope towards a proposal for first
  class environments in Scheme, to which I am opposed. Could you clarify
  your intentions for me, please? Do your motives go beyond the obvious
  desire just to bring EVAL into the new language report? Do you or any
  of your MIT Scheme colleagues, in fact, have intentions of issuing a 
  proposal for first class environments in the near future? Etc.

I believe that such a query could more effectively have conveyed your
concerns without subtely implying that Jinx's proposal should be rejected
solely on the basis of vague accusations of political skulduggery.

Naturally, the content and tone of your original message make me wonder
if perhaps you too may be (perhaps unconciously) a bit disingenuous.
``Methinks thou dost protest too much.'' To wit:

    I am confused about what precisely you are suggesting in your
  objection to Bill Rozas' EVAL proposal. Do you, in principle, object
  to having EVAL per se be specified in the language report? If so,
  on what basis? Do you have some well-reasoned aversion to EVAL or
  is your objection based solely on the presumption that Bill is
  paving the way towards a secondary proposal which you deeply oppose,
  namely, first class environments? Please clarify. 

    It would be unfortunate indeed if the discussion at the next report
  meeting were somehow disrupted or de-railed by innuendo. I am trying to
  avert this by encouraging you to more precisely voice your technical
  objections to Bill's proposal. We are all mature adults here and we
  should exhibit respect for each other's professionalism. Specifically,
  we should not be afraid to openly express our technical concerns or to
  openly request clarification on points we do not fully understand. For
  example: What, specifically, in Bill's proposal leads you to feel that,
  if passed, would compell the committee in the direction of also adopting
  first class environments?

    It is my opinion that this is not Bill's intent, but perhaps I am
  mistaken. Admittedly, Bill can be a bit crafty at times... I believe it
  is an aspect of his acute (and uniquely Spanish) sense of humor.

Anyway, I sincerely hope this clears the smoke a little. I just wish I
could make it to the RnRS meeting. I hope some very exciting results will
come out of it and I look forward to reading the minutes!

Which brings to mind... is there any chance someone might be psyched to
video tape the sessions? I would dearly love to be able to experience
it, if only virtually.

Maybe next time around we could try to arrange a tele-conference format.
Our lab here at MIT just installed a ma$$ively high-tech conference room
which can support tele-conferencing and tele-broadcasting. It might be
neat to experiment with a tele-workshop.


ziggy@lcs.mit.edu    Michael R. Blair   MIT Laboratory for Computer Science
(617) 253-6833 [O]         -.           545 Technology Square --- Room 437
(617) 577-1167 [H]          /\.         Cambridge, MA   02139