W3C home > Mailing lists > Public > w3c-math-erb@w3.org > September 1996

Re: notation for extending precedence

From: Bruce Smith <bruce@wolfram.com>
Date: Thu, 19 Sep 1996 20:50:37 -0700
Message-Id: <v02130509ae6736ced879@[199.182.131.98]>
To: raman@Adobe.COM
Cc: w3c-math@w3.org
At 8:51 AM 9/18/96, T. V. Raman wrote:
>AsTeR used about 12-15 precedence levels (I think)
>and did not  handle associativity but I could mail out the file of lisp code
>containing the precedence definitions.
>Would this be useful?
>--the file contains all the operators  with the precedence defined (all this
>is from memory --I'll have to go back and look)

This would be interesting to see, but I'm not sure if my original
question was clear. I didn't mean to ask for a suggestion of
precedences themselves, but rather, for a suggested file format for a
large table of operators and their precedences, which doesn't use
numerical precedence levels (like your suggested format
for defining new operators doesn't), but which also doesn't require
each operator to mention the name of the preceding operator in the
list (like your suggested format would require, if used for this).

That is, I see how your suggested format could be used to define
each built-in operator as having precedence less than the previous
one in the file, but I would prefer a file format in which it was
easy to insert (by text-editing operations on that file) a new operator
between two of the existing ones, or to reorder some of the operators.

This could be done by requiring them to be listed in order
of non-decreasing precedence, with some indication of when the
precedence level changes (since many of them have the same level),
but the question is, how well can this coexist with the ability to
define new operators incrementally -- that is, can one format for
operator definitions solve both problems well?
Received on Thursday, 19 September 1996 23:50:07 UTC

This archive was generated by hypermail 2.4.0 : Saturday, 15 April 2023 17:19:57 UTC