"unordered" operator

[uunet!att!attunix!dfp]

>> Tom,
>> 
>> Thanks for pointing out the problem in the NCEG draft document regarding
>> the ?? relational operator.  Do you have a suggestion for an "available"
>> character that could represent "unordered" in relational operators?
>> 
>> -Jim
>> 
>Yes.  Use <>, as in x <> y, pronounced "x box y" if you want a short
>pronunciation for <>.  NCEG can either introduce the new lexical element
><> as a single token, which will mean x < > y won't be acceptable,
>or can introduce it as the two tokens < followed by >.  Either way
>introduces no problem with respect to the ANSI standard.
>
>The only objectors might be Pascal programmers.  They are used to <>
>meaning "not equal".  Perhaps then we should choose x >< y.  In fact
>I like that!  It looks weird enough it easily conveys "not comparable to",
>but what should be the monosyllabic description?
>Perhaps "ouch".  Perhaps "never" (even though 2 syllables)
>as in "x never y" -- i.e., x and y are never comparable.
>
>Sounds like >< is the best choice, with <> a close second.
>I'd avoid using a new character.

At the NCEG meeting, I suggested using

	<>=	(instead of <=>)

for less than, greater than, or equal, and

	!<>=

for not ..., or unordered.  The proposal already uses

	<>

for less than or greater than, and

	!<>

for not ..., or equal to or unordered.

In all of the above, the <> sequence could be written as >< if
people like that better.  I find the <> ordering much easier,
possibly because it is used in a pretty similar manner in other
languages, and because that's the order of the characters on my
keyboards.

Dave