rendering brackets

[Michael J. Baars]

On Thu, 2020-03-05 at 07:51 +0000, David Carlisle wrote:
> On Thu, 5 Mar 2020 at 06:14, Michael J. Baars
> <mjbaars1977.tex-live at cyberfiber.eu> wrote:
> > On Tue, 2020-03-03 at 20:58 +0000, David Carlisle wrote:
> > > On Tue, 3 Mar 2020 at 20:31, Michael J. Baars
> > > <mjbaars1977.tex-live at cyberfiber.eu> wrote:
> > > ....
> > > > 1/a: 13.20952pt+6.85951pt
> > > > a/1: 11.07062pt+6.85951pt
> > > > 
> > > > 
> > > > so the first form is over 2pt taller and so gets a larger
> > > > bracket
> > > > if you auto-size the brackets with \left\right.
> > > > 
> > > > 
> > > > Yes, which is really strange if you ask me. One would expect
> > > > them
> > > > to be exactly the same size.
> > > 
> > > why? 1 is clearly taller than a.
> > 
> > Yes, you are completely right. '1' is not the same height as 'a',
> > but
> > the sum of the two should be constant, therefore '1/a' and 'a/1'
> > should
> > be the same height.
> 
> as I said before, the sum never comes in to the layout at all, if
> they
> were closely stacked so that the sum affected the total height the
> the
> fraction bar on 1/a woul dbe lower than the one on a/1 and that is
> not
> what happens. the fraction bar (if at all possible) is set on the
> math
> axis, the same height as - or the bar in +

Let's look at the other styles, as you can see from the attachment,
both '1' and 'a' are first placed in boxes of the same height. Both the
height of the individual characters and the sum of the heights of the
characters is then of no importance. All the different styles produce
fractions of the same height and are rendered with the same brackets,
except the one rendered in displaystyle.

> > > > I do not know the exact details on height and depth as brought
> > > > forth by tex-live, but I can't remember mentioning 3D letters.
> > > 
> > > 3D ??
> > 
> > 3D always has the z-axis, or depth, perpendicular to the face of
> > the
> > monitor. I don't see what 'depth' as to do with a 2D box.
> 
> please see any tutorial about how tex boxes work, a box consists of
> three numbers width height (above the baseline) and depth (below the
> baseline)  a fraction, like g and y has non zero depth/
> > > all boxes in tex have height (above the baseline) depth (below
> > > the
> > > baseline) and width.
> > > 
> > > > I don't see any connection between the depth and the height of
> > > > the
> > > > denominator.
> 
> the depth  of the fraction is controlled by the depth of the
> denominator so is teh same for 1 or a but would be greater for y
> 
> > > As normal in tex, the placement is based on aligning characters
> > > by
> > > the
> > > baselines so in the denominator the fact that 1 is taller than a
> > > has
> > > no effect as the default position of the characters in the
> > > denominator
> > >  has the baselines in the same level.  the same is true of the
> > > distance between the fraction bar and the baselines in the
> > > numerator,
> > > so the baselines of the 1 and a are in the same place so the top
> > > of
> > > the 1 is higher than the top of the a.....
> > > 
> > > > That 1 and a are different in height should not be a problem,
> > > > since
> > > > the sum of the heights is constant,
> > > 
> > > The sum of the heights isn't involved as they are set such that
> > > the
> > > fraction bar is on the math axis you are not simply stacking 1
> > > over a
> > > and a over 1 and getting a stack of the same height,
> > > 
> > > 
> > > > but what you are saying, is that using a monospace font should
> > > > solve the problem? I'll look into that tomorrow morning.
> > > 
> > > No I never mentioned monospace. Monospace fonts have characters
> > > equal
> > > width but typically 1 will be taller than an a still. The widths
> > > of
> > > the characters are not invoved here.
> > > 
> > > 
> > > David
-------------- next part --------------
A non-text attachment was scrubbed...
Name: brackets2.tex
Type: text/x-tex
Size: 946 bytes
Desc: not available
URL: <https://tug.org/pipermail/tex-live/attachments/20200307/6f26efa5/attachment.bin>