Quite awhile back, we had a discussion about removing "$" from the set
of allowed characters in operator names, and instead allowing it as a
non-first character in identifiers. (It'd have to be non-first to avoid
ambiguity with parameter symbols "$nnn".) See, eg,
http://archives.postgresql.org/pgsql-hackers/2001-08/msg00629.php
That discussion petered out without any definite consensus being
reached, but I think it's time to reconsider the idea. We're getting
flak about "x<$n" being parsed as "x <$ n" rather than "x < $n" (see
current thread in pgsql-sql). While this has always been a hazard for
SQL and plpgsql function writers, it is now also a hazard in direct
SQL, if you use PREPAREd queries. So I think the importance of avoiding
such problems has moved up a notch as of 7.3.
So, I'd like to put that proposal back on the table. Comments anyone?
regards, tom lane