# Re: Too narrowly defined: DIVISION SIGN & COLON

From: Mark Davis ☕ <mark_at_macchiato.com>
Date: Tue, 10 Jul 2012 17:35:10 -0700

The main point is that asserting a general preference in an annotation for ∶ to
express a ratio, as Asmus had in his formulation, is simply wrong and
counterproductive. (We are not going to change the world's usage from : to ∶
by fiat; and and the glyphic difference is quite subtle, and missing in a
great many fonts. Compare that with the difference between hyphen-minus and
minus, which is much more pronounced, and much better carried across fonts.)

The most that we could say is that in certain mathematical contexts ∶ is
preferred to : for expressing ratios, not that it is generally preferred.

By the way, here's your list with visible characters instead of the U+'s.

- HYPHEN-MINUS
− MINUS SIGN

/ SOLIDUS (Unicode 1.0 called it "SLASH")
∕ DIVISION SLASH

\ REVERSE SOLIDUS (Unicode 1.0 called it "BACKSLASH")
∖ SET MINUS

* ASTERISK *// you had U+003A = : instead of *.*
∗ ASTERISK OPERATOR

◦ WHITE BULLET
∘ RING OPERATOR

• BULLET
∙ BULLET OPERATOR

| VERTICAL BAR
∣ DIVIDES

‖ DOUBLE VERTICAL BAR
∥ PARALLEL TO

: COLON
∶ RATIO

~ TILDE
∼ TILDE OPERATOR

· MIDDLE DOT
⋅ DOT OPERATOR

------------------------------
Mark <https://plus.google.com/114199149796022210033>
*
*
*— Il meglio è l’inimico del bene —*
**

On Tue, Jul 10, 2012 at 5:05 PM, Ken Whistler <kenw_at_sybase.com> wrote:

> On 7/10/2012 4:22 PM, Mark Davis ☕ wrote:
>
> I would disagree about the preference for ratio; I think it is a
> historical accident in Unicode.
>
>
> Not really.
>
> The following pairs dating from Unicode 1.0 were deliberate:
>
> U+002D HYPHEN-MINUS
> U+2212 MINUS SIGN
>
> U+002F SOLIDUS (Unicode 1.0 called it "SLASH")
> U+2215 DIVISION SLASH
>
> U+005C REVERSE SOLIDUS (Unicode 1.0 called it "BACKSLASH")
> U+2216 SET MINUS
>
> U+003A ASTERISK
> U+2217 ASTERISK OPERATOR
>
> U+25E6 WHITE BULLET
> U+2218 RING OPERATOR
>
> U+2022 BULLET
> U+2219 BULLET OPERATOR
>
> U+007C VERTICAL BAR
> U+2223 DIVIDES
>
> U+2016 DOUBLE VERTICAL BAR
> U+2225 PARALLEL TO
>
> U+003A COLON
> U+2236 RATIO
>
> U+007E TILDE
> U+223C TILDE OPERATOR
>
> U+00B7 MIDDLE DOT
> U+22C5 DOT OPERATOR
>
> If anything, the "accident" is that the use of "!" for factorial was not
> distinguished with a separate symbol character. I don't recall the
> argument in detail -- it was discussed. But I suspect that it came down
> to most of the math operators being in principle distinguishable because
> they are rendered on the math centerline, rather than the baseline,
> whereas nobody could think of a good reason for a layout distinction
> for the factorial -- so it fell instead into the bucket already occupied
> by "." as full stop versus decimal point (versus record separator
> versus...)
>
> Now subsequent history has since led to more systematic distinctions,
> both in use and in glyph design, for some of the pairs listed above.
> For example, the two tildes generally look different. The SET MINUS was
> discovered to actually be distinct from a backslash, with a different
> angle and length. And so on. So that has whittled down the list of
> characters that people, after the fact, come to think of as accidental
> duplicates.
>
> But trying to rationalize these decisions by examining only the latest
> charts, while ignoring the history of how these distinctions came about
> in the first place is not a productive direction, IMO.
>
> Incidentally, one of the reasons the set of symbols in the U+2200
> Mathematical Operators block got a somewhat different treatment than
> generic punctuation or other symbols or combining marks, when it comes
> to unification versus non-unification decisions back in the original
> draft charts in 1989 and 1990 had something to do with the intuition
> back then that having unambiguous encodings for the math operators
> would be important for machine processing of mathematical data
> (as in algebra systems). It isn't so clear now, in retrospect, whether
> some of the disunifications were a good idea or not. But those
> decisions are what we have inherited in the standard now, for better
> or worse.
>
> --Ken
>
>
>
>
>
Received on Tue Jul 10 2012 - 19:36:35 CDT

This archive was generated by hypermail 2.2.0 : Tue Jul 10 2012 - 19:36:36 CDT