Jorgen Schaefer scripsit:

> That Scheme does not allow read syntax modification is, in my
> opinion, a good thing.

Distinguo.  Scheme does not *forbid* read syntax modification; it merely
provides no standard way to specify it.

> I'm not sure I've ever seen a really useful
> read table modification that would have needed a special
> character. Indeed, the special characters usually only lead to
> badly-readable code.

Readability is in the eye of the beholder.  Quite apart from the [Yy]ahoos
who condemn Scheme as unreadable *tout court*.

> I would advocate against reserving too many characters.

You will already have about 4000 symbols (900 of them mathematical in nature)
to use in identifiers, to say nothing of the vast supply of letters and
their analogues for spelling out names in virtually any natural language.

> The currently reserved ones suffice

Of course they suffice for someone who doesn't believe in using them.

> - and that only includes the curly
> braces after the syntax modification of this SRFI/R6RS.

Section 7.1.1 of R5RS specifically says:

	The following five characters are reserved for future extensions
	to the language: [ ] { } |

> Your list also includes quite a few characters which I
> definitively would like to allow in identifiers, if we allow
> Unicode characters at all[1] (These include the reversed question
> mark, among others).

I wouldn't be averse to removing a few characters ad hoc on the grounds
that they have close ASCII relatives that we already allow in identifiers.

--
John Cowan  xxxxxx@reutershealth.com  www.ccil.org/~cowan
Female celebrity stalker, on a hot morning in Cairo:
"Imagine, Colonel Lawrence, ninety-two already!"
El Auruns's reply:  "Many happy returns of the day!"