[R6RS] Library syntax versioning

[William D Clinger]

Mike wrote:
> Two things are getting conflated here: The notion of the code you
> write and the notion of code that gets distributed to someone else.
> SRFI 83 is for the latter.

Repetition does not convince.

> I think you want to argue that, as this is
> what the R6RS will specify, people will use the same notation to write
> code.  That's OK, but in my mind, much less important.

We are of different minds.

> As to writing code, could you say why you don't want to write the
> following single line at the top of a file containing library module
> specifications?  (OK, two: there has to be a blank line after it.)
> 
> Content-Type: application/x-scheme; report-revision=6

I am not convinced that your proposal is any less verbose
than "scheme://r6rs".

Kent wrote:
> Perhaps we can plan on using the same tool for future extensions to the
> syntax:
> 
>   #!r7rs

I like that.

> We can even require #!r6rs to appear at the top of a library now, or at
> least allow it to be included.

Indeed.

I propose we add #!r6rs as a new external representation
that every R6RS-conforming implementation must support.
Its purpose is to flag code that is written in the lexical
syntax of R6RS, to ease the eventual transition from R6RS
to R7RS lexical syntax.

It hardly matters what #!r6rs represents, but I further
propose that it be an external representation for the
canonical unspecified value we voted into existence last
week; that value has no external representation as yet,
so no one can say #!r6rs is redundant.

Will