Per Bothner wrote, "Semantically, ... is arguably syntactic sugar for
...".  (Details of ... are in the mail archive.)  In other words,
these are supposed to be semantically equivalent.  I showed that they
are not.  Now you (Per) write

> Perhaps I should re-phrase my point.

The problem I see is with the point, not with its phrasing.  And I'm
afraid this restatement doesn't help me much either:

>					Consider:
> (G) (set! top-level-variable v)
> (L) (set! lexical-variable v)
> (F) (set-member! table selector v).
>
> I claim (F) is closer semantically to (G) than (G) is to (L).

I honestly don't understand this notion of semantic "close"ness.
Equality I know; proximity, no.  As such, therefore, I have no way of
verifying your claim.  It doesn't help that you have only asserted it.

If this is unrelated to your justification for overloading set!, then
let's move on.  But if it is, formalizing this would be a great help.

'shriram