On Tue, Oct 27, 2009 at 11:22 PM, Brian Harvey <bh@cs.berkeley.edu> wrote:
> Arthur, you didn't reply to the most important paragraph of Ray's message:
>
>> It would not have been necessary to have more than one working
>> group in the first place if subsetting were a viable approach.
>> Rather, it would be appropriate for the working group to produce
>> a document specifying some constraints as "optional", as was done
>> in past RNRS reports. Indeed, this will be the net effect of the
>> current effort if the draft WG charters are ratified and executed.

Since WG2 Scheme is, by the charters, supposed to be compatible with
WG1, I don't understand why this statement is surprising or
controversial.  Separating WG1 from WG2 is just a way of organizing
the work.  The two groups will still have to work together closely to
achieve their independent goals.

> I think that the Steering Committee (amazingly, in my view) thinks that
> R6-haters merely hate the weight of the standards document, and if we
> could just leave out every other page, we'd have a nice WG1 Scheme.
> It's not true.  At least it's not true for /this/ R6-hater.  For me,
> R6 isn't just too much of a good thing; it's a huge step in a wrong
> direction.  The SC set up two WGs because they recognize that my "wrong
> direction" is other people's right direction.

Instead of spending energy trying to figure out what the Steering
Committee thinks other people think, let's concentrate on coming up
with a good set of goals for the two working groups -- for choosing
what goes into the core Scheme, WG1 Scheme, and what goes into the
more full-featured Scheme, WG2 Scheme.  That's the best way to move
past earlier disagreements and toward something that more people can
be happy about.

Do you have any specific recommendations for the two charters?

> For me, the paradigmatic issue is that the R6 people view the Lisp REPL as
> an unnecessary frill -- well, let me not put opinions in other people's
> brains and merely remark that they produced a document that makes proper
> REPL support impossible.  I am sitting here right now banging my head into
> the wall at the very thought of this!

Please don't hurt your head!

Can you suggest any specific changes to the charters that would
express your goals?

>                                        And, no, it doesn't help to have two
> modes.  It's vitally important that someone who has no idea that there's
> such a thing as a compiler or a macro should be able to understand the
> behavior of the language.  The only mode is REPL-compatible mode, even if
> under the hood there's a brilliant compiler.  No R6-compatible implementation
> can provide this behavior.

The draft charters don't require strict R6RS compatibility.  The WG1
charter says "...the language should be backwards compatible
with... an appropriate subset of the R6RS standard."  The WG2 charter
says "...working group 2 should consider all features provided by R6RS
Scheme, and all criticisms of those features."  This leaves
considerable room for change, as long as it is well justified, but
also encourages the working groups to preserve the "good parts" of
R6RS, whatever everyone agrees those are.

> In my opinion Ray is 100% right.  If the two languages were compatible,
> there wouldn't be all that blood on the floor.

As heated as the arguments have been, I prefer not to think of them in
such terms.  That's focusing on the past, but there's so much good
that we can do in the future if we put our minds to it.  Sure, there
will be disagreements, but let's work on them.

> PS The message from Ray to which you responded is /not/ the first commentary
> on the WG charters!  (It may be that Ray was first, but not in that message.)

But that's not what I said.  I said "Thank you very much for being the
first person to publicly comment on the *revised* [emphasis added]
charters (as far as I know)."

> Please review the archives of the mailing list.  There has been lots more
> detail about why the WG1 \subset WG2 approach can't work.

I've been following the discussion of the earlier drafts of the
charters on <r6rs-discuss> since the beginning.  I still believe that
we can find common ground, and I believe that arranging the effort
into two separate but cooperating working groups is the best way to
organize the work.  The groups will have different emphases, but their
goals will not be incompatible.

Let's think about what we *can* do, not about what we did or didn't do
right before!  To start, how specifically can we improve the charters?