[R6RS] Document endgame

[William D Clinger]

Mike wrote:
> - The "ledge": Will, could you give a rough summary on what you think
>   it would look like?

Not too different from what I see in the current
draft, whose organization looks better than I had
remembered.  In particular, I would suggest putting
the ledge before these things:

    declarations (currently section 5.2)
    conditionals (cond, case, and, or; currently 5.6.5)
    iteration (do, named let; currently 5.6.8)
    quasiquotation (currently 5.6.9)
    binding constructs for syntactic keywords (currently 5.7)
    length, append, reverse, list-tail, list-ref, map, for-each
        (currently in 5.14)
    substring, string-append, string->list, list->string,
        string-copy, string-fill! (currently in 5.17)
    vector->list, list->vector, vector-fill! (currently 5.18)
    apply, call-with-current-continuation, call/cc, values,
        call-with-values, dynamic-wind (currently 5.20)
    tail calls and tail contexts (currently 5.21)
    syntax-rules (currently section 7)
    records (currently section 8)
    et cetera

I would suggest that that some of the introductory
folderol on the numerical tower (currently section 6)
be moved before the ledge.  In particular, the current
draft of section 5.12 refers to exact and inexact numbers,
flonums, IEEE-754 single and double precision, infinities,
and NaNs, none of which are really motivated or explained
until section 6.

Will