Negative/Negation Sign

[Asmus Freytag]

On 10/29/2022 3:42 PM, Sławomir Osipiuk wrote:
> On Saturday, 29 October 2022, 17:56:20 (-04:00), Asmus Freytag via 
> Unicode wrote:
>
>     I can see no indication that the TI engineers had some other
>     symbol in mind, that is had they had the choice of a
>     Unicode-encoded outline font, they would have chosen something
>     with an appearance very distinct from SUPERSCRIPT MINUS. Unless
>     and until someone can come up with a very cogent argument that
>     they were really trying to model something that is visually
>     distinct from a superscript minus sign, there is no reason to
>     reject that mapping.
>
>
> The argument is simple enough: a minus sign as part of the exponent 
> should be visually distinct from a negation sign in the base.

And they are. One comes before the base, the other one after the base. 
And since negation is unary, it's never preceded by anything other than 
an operator, delimiter or a space.

>
> The TI engineers were trying to visually separate subtraction and 
> negation. To the extent we can try to deduce their reasoning, they 
> would not have wanted to immediately confuse negation with negative 
> exponents, which is what the superscript does.

>
> Someone with the appropriate calculator can confirm: What does 3⁻¹−-3 
> ("three to the power of negative one, then subtract negative three") 
> look like? If the negative symbol in the exponent and the one 
> preceding the three are the same, I'll admit the superscript is fine 
> in this case.

You don't need the calculator - you can look up the 7x5 bitmaps for the 
fonts. The result looks like

3⁻¹−⁻3

which is clear and unambiguous. The second ⁻ cannot ever be part of an 
exponent.

BTW, the same engineers have a provided a precomposed symbol for ⁻¹. The 
Wikipedia suggests mapping that to <207B 00B9>, while the other source 
maps that to  <203E 00B9> which uses the clearly inappropriate overline.

>
> My view is that the modifier letter minus (U+02D7) is the best option 
> to respect the intended semantics, while the plain hyphen-minus 
> (U+002D) would be my second choice.
>
> As for Wikipedia, it's ridiculous to say that one person's opinion on 
> an extremely esoteric detail, left uncontested (or more likely 
> unnoticed and unquestioned) is enough to form some kind of de facto 
> standard. But, if we are going by that logic, I suggest you check the 
> page again. ;-)

No more ridiculous than your personal choice, immediately contested here. :)

The fact is that a centerline glyph, no matter whether shorter than the 
minus sign or not, does not match the conventions used by TI. You are, 
of course, free to suggest that your notation is superior; it's just not 
a better *mapping* for what is available on those calculators. However, 
nothing prevents you from using it in your own documents.

Neither makes an argument for encoding a new characters - which is what 
had started this thread.

A./

PS: incidentally, the TI font has both a "DASH" and  "HYPHEN" (names 
given in one of the listings for the characters). The former has the 
width of what Unicode encodes as 2212 (same as the while the latter is 
shorter. The location in the original set near "+" and "=" makes clear 
that "DASH" is meant for the minus sign and the mapping in both source 
and Wikipedia therefore has 2021 for it, while the "HYPHEN" is mapped to 
2010.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://corp.unicode.org/pipermail/unicode/attachments/20221029/c08e926a/attachment.htm>