[texhax] vertical placement of symbols

[Uwe Lueck]

Zbigniew Nitecki <zbigniew.nitecki at tufts.edu> schrieb am 05.05.2009 15:15:41:
> What you suggested seems exactly what I was looking for, but it is a 
> disaster. 

ok, sorry, it must be \vcenter{\hbox{$\scriptstyle\circ$}} (now tested), or \mathbin{\vcenter{\hbox{$\scriptstyle\circ$}}} after Krishnan, or a compromise suggestion of mine: \mathinner{\vcenter{\hbox{$\scriptstyle\circ$}}.

The inside-\v... error is a matter where I need help myself. Whenever I try to understand this, I think to learn that the width of a \v... is the maximum width of what is inside, but actually the width so often is the outer \hsize. It seems that the latter happens at entering hmode. The \hbox avoids entering hmode. If you replace \hbox by \mbox, you have the same problem as before. 

I wonder about \mathbin. I would agree when the subject is a group of transformations, but it would be misleading with 

    f o g(x) = ... 

The latter could be improved to 

    (f o g)(x), 

but you may have an audience that knows mathematics as a tool for its own science only and won't understand the notation easily. 

Once I knew this matter (of style in this case), I have forgotten it. But I think sometimes it is against conventions to set binary operations as \mathbin. Somewhat obvious: 1/2, 1/3 etc. About the same with a backsla&#347;h: I think when I saw \setminus the first time, I thought it was wrong, I thought there usually is no space around the set difference backslash. Or it is more usual to typeset it as \mathinner, i.e., with \, space instead of \,\, (= \:). 

? 

-- Do you avoid "reply to all" deliberately? 

-- Uwe.


> Someone else suggested using \mathbin, which doesn't affect vertical 
> spacing, but does adjust horizontal spacing, although as the attached 
> shows, it seems to make relatively little difference.
> I enclose source and output.
> 
> On May 4, 2009, at 11:00, Uwe Lueck wrote:
> 
> Zbigniew Nitecki <zbigniew.nitecki at tufts.edu> schrieb am 04.05.2009 
> 16:35:15:
> Yes, but \circ alone gives a circle that is really too big.
> 
> If so, $f \comp g$ with \newcommand{\comp}{\vcenter{$\scriptstyle\
> circ$}}. 
> -- Uwe. 
> 
> On May 4, 2009, at 10:29, Uwe Lueck wrote:
> 
> Zbigniew Nitecki <zbigniew.nitecki at tufts.edu> schrieb am 04.05.2009 
> 15:14:08:
> To create the symbol for composition of functions, I have the 
> following code
> 
> \newcommand{\compose}{{\scriptstyle \circ}}% composition symbol
> \newcommand{\comp}[2]{\ensuremath{#1\compose #2}}% composition by 
> function name
> 
> This works fine, except that the circle is not vertically centered---
> it is too low. How do I adjust it?
> 
> Isn't it fine when you drop \compose and replace it by \circ? 
> 
> I even would not be willing to type `\comp{f}{g}' in place of `$f \
> circ g$' or, in equations, `f \circ g'. Or \let\comp\circ to see "
> what you mean" rather than "what you get". 
> 
> I guess not the vertical placement is bad, it is choosing \scriptstyle 
> 
> that chooses a font size too small between normal size symbols. 
> 
> 
>