Bug in TeX input mode?

Bug in TeX input mode?

[Deniz Dogan]

Hi

After some random discussion on #haskell about having Unicode
characters as operators, I found that \circ in the TeX input mode
inserts the following character:

--------

        character: ○ (9675, #o22713, #x25cb)
preferred charset: unicode (Unicode (ISO10646))
       code point: 0x25CB
           syntax: _ 	which means: symbol
         category: .:Base, c:Chinese, h:Korean, j:Japanese
         to input: type "\circ" with TeX
      buffer code: #xE2 #x97 #x8B
        file code: #xE2 #x97 #x8B (encoded by coding system utf-8-unix)
          display: by this font (glyph code)

Character code properties: customize what to show
  name: WHITE CIRCLE
  general-category: So (Symbol, Other)

--------

However, we feel like what should have been inserted is this:

--------

        character: ∘ (8728, #o21030, #x2218)
preferred charset: unicode (Unicode (ISO10646))
       code point: 0x2218
           syntax: . 	which means: punctuation
         category: .:Base
      buffer code: #xE2 #x88 #x98
        file code: #xE2 #x88 #x98 (encoded by coding system utf-8-unix)
          display: by this font (glyph code)

Character code properties: customize what to show
  name: RING OPERATOR
  general-category: Sm (Symbol, Math)

--------

I'm not sure whether this is a bug in TeX input mode or correct. Any ideas?

-- 
Deniz Dogan

Re: Bug in TeX input mode?

[Chong Yidong]

Deniz Dogan <deniz.a.m.dogan@gmail.com> writes:

> \circ in the TeX input mode inserts the following character:
>
>         character: ○ (9675, #o22713, #x25cb)
>   name: WHITE CIRCLE
>   general-category: So (Symbol, Other)
>
> However, we feel like what should have been inserted is this:
>
>         character: ∘ (8728, #o21030, #x2218)
>   name: RING OPERATOR
>   general-category: Sm (Symbol, Math)

Maybe.  But if you look at the TeX output of \circ, the size of the
circle looks a lot more like ○ than ∘ (at least, when using DejaVu Sans
Mono, I haven't checked with other fonts).  And it seems more reasonable
for "\circ" to mean WHITE CIRCLE than RING OPERATOR.

Re: Bug in TeX input mode?

[Teemu Likonen]

On 2009-06-17 17:26 (-0400), Chong Yidong wrote:

> Deniz Dogan <deniz.a.m.dogan@gmail.com> writes:
>
>> \circ in the TeX input mode inserts the following character:
>>
>>         character: ○ (9675, #o22713, #x25cb)
>>   name: WHITE CIRCLE
>>   general-category: So (Symbol, Other)
>>
>> However, we feel like what should have been inserted is this:
>>
>>         character: ∘ (8728, #o21030, #x2218)
>>   name: RING OPERATOR
>>   general-category: Sm (Symbol, Math)
>
> Maybe.  But if you look at the TeX output of \circ, the size of the
> circle looks a lot more like ○ than ∘ (at least, when using DejaVu Sans
> Mono, I haven't checked with other fonts).  And it seems more reasonable
> for "\circ" to mean WHITE CIRCLE than RING OPERATOR.

I don't know if this experiment has any value but I just compiled a
LaTeX document with $\circ$ to PDF file and then copied the character
from KPDF reader to Emacs. The character looks like this:

        character: ◦ (9702, #o22746, #x25e6)
  name: WHITE BULLET
  general-category: So (Symbol, Other)

Re: Bug in TeX input mode?

[Deniz Dogan]

2009/6/17 Chong Yidong <cyd@stupidchicken.com>:
> Deniz Dogan <deniz.a.m.dogan@gmail.com> writes:
>
>> \circ in the TeX input mode inserts the following character:
>>
>>         character: ○ (9675, #o22713, #x25cb)
>>   name: WHITE CIRCLE
>>   general-category: So (Symbol, Other)
>>
>> However, we feel like what should have been inserted is this:
>>
>>         character: ∘ (8728, #o21030, #x2218)
>>   name: RING OPERATOR
>>   general-category: Sm (Symbol, Math)
>
> Maybe.  But if you look at the TeX output of \circ, the size of the
> circle looks a lot more like ○ than ∘ (at least, when using DejaVu Sans
> Mono, I haven't checked with other fonts).  And it seems more reasonable
> for "\circ" to mean WHITE CIRCLE than RING OPERATOR.
>

That's weird, because I'm using the same font and I think it looks way
much more like ∘! :) The Haskellers I talked to said that \circ is the
standard way of writing the "function composition operator" in TeX and
IIRC, the function composition operator is sometimes called "ring
operator", due to its shape. Copy-pasting the composition operator
from Wikipedia reveals that they use the Unicode RING OPERATOR as
well, but of course, this needn't mean that TeX does...

-- 
Deniz Dogan

Re: Bug in TeX input mode?

[Kenichi Handa]

In article <87y6rqimc5.fsf@stupidchicken.com>, Chong Yidong <cyd@stupidchicken.com> writes:

> Deniz Dogan <deniz.a.m.dogan@gmail.com> writes:
> > \circ in the TeX input mode inserts the following character:
> >
> >         character: ○ (9675, #o22713, #x25cb)
> >   name: WHITE CIRCLE
> >   general-category: So (Symbol, Other)
> >
> > However, we feel like what should have been inserted is this:
> >
> >         character: ∘ (8728, #o21030, #x2218)
> >   name: RING OPERATOR
> >   general-category: Sm (Symbol, Math)

> Maybe.  But if you look at the TeX output of \circ, the size of the
> circle looks a lot more like ○ than ∘ (at least, when using DejaVu Sans
> Mono, I haven't checked with other fonts).  And it seems more reasonable
> for "\circ" to mean WHITE CIRCLE than RING OPERATOR.

But, Knuth categorizes \circ in "binary operator" in his
"The TeX book".

---
Kenichi Handa
handa@m17n.org

Re: Bug in TeX input mode?

[Stefan Monnier]

> I'm not sure whether this is a bug in TeX input mode or correct. Any ideas?

I agree it's a bug in the TeX input method, probably due to the fonts
available back when it was written.


        Stefan