From: Karlsson Kent - keka

To: 'kenw@sybase.com' ; 'mark.davis@us.ibm.com'

Cc: 'Asmus Freytag' ; mark@macchiato.com

Sent: Thursday, May 27, 1999 10:35 AM

Subject: Tildes and micro sign decompositions

Please forward to the UTC for the June meeting.
I hope this e-mail format is sufficient.  If not, please tell
me which form to fill in.

                Kind regards
                /Kent Karlsson

=======================================

NOT TILDE (2241) should be tied to the tilde operator in the
decomposition, not the spacing tilde accent:

     2241;NOT TILDE;Sm;0;ON;007E 0338;;;;Y;;;;; --->
     2241;NOT TILDE;Sm;0;ON;223C 0338;;;;Y;;;;;

The NOT TILDE is like the other negated math characters. 
The 007E TILDE, however, is not normally seen as a math
character, even though it is used in several programming
lanugages as a (usually unary) operator.  But so is
HYPHEN-MINUS, which is not classified as a math character.

TILDE is often printed/displayed *above* the math operator
vertical centre line, whereas TILDE OPERATOR and NOT
TILDE are set *on* the math operator vertical centre line,
just like =, <, etc.

------------------------------

The spacing tilde accent, TILDE (007E), is not decomposed like
the other spacing versions of the accents, as it should. 

    007E;TILDE;Sm;0;ON;;;;;N;;;;;  --->
    007E;TILDE;Sk;0;ON;<compat> 0002 0303;;;;N;;;;;

    02DC;SMALL TILDE;Sk;0;ON;<compat> 0020 0303;;;;N;SPACING TILDE;;;; --->
    02DC;SMALL TILDE;Sk;0;ON;<small> 007E;;;;N;SPACING TILDE;;;;

This is probably closest to the original intent and use of the TILDE in ASCII.
It is also closer to the *current* view of tilde in many keyboard mappings,
where a "dead key" tilde can be followed either by a letter (producing
õ, ñ, etc.) or by space. The latter produces ~, 007E TILDE.  Indeed, that
is the only way of producing TILDE on many keyboards (hex code input
disregarded).

(SMALL TILDE can be made canonically equivalent to TILDE.)

---------------------------

Alternatively (to the above), continue to regard TILDE as a
math operator, but then make TILDE OPERATOR canonically
equivalent to it:

   223C;TILDE OPERATOR;Sm;0;ON;;;;;Y;;;;;  --->
   223C;TILDE OPERATOR;Sm;0;ON;007E;;;;Y;;;;;

This, however, departs from the common view of 007E TILDE.

------------------------

The micro sign should be *canonically* equivalent to Greek mu:
 
    00B5;MICRO SIGN;Ll;0;L;<compat> 03BC;;;;N;;;;; --->
    00B5;MICRO SIGN;Ll;0;L;03BC;;;;N;;;;;

Just because it was thought important enough (for use in units) to include
this character in Latin-1 does not make this mu any different from
the mu encoded in the Greek block.  The situation is very similar
to that for the Kelvin sign, the Ångström sign, and the Ohm sign,
which are all canonically equivalent to the corresponding letter.

This has mistakenly lead to some slight appearance differences
between these equivalenced characters, even though they are or
should be canonically equivalent.  This should be regarded as font
mistakes, and not grounds for not making them canonically equivalent.

-----------------------------------------------------------------