mspace default attributes

Hi,
I was wondering if the default attribute values for mspace are
appropriate. At first sight, they seem to be harmless, all zero. But
let's take the following example. Let's take an accent, say a dot in a
LaTeX font. Well, since it's an accent, it's positioned rather high with
respect to the normal baseline. In fact, it has a negative descent,
meaning that the bottom most pixel of the glyph is ABOVE the baseline.
Let us say that the descent is -10 and the ascent is 11. Thus the dot
accent is fully contained in a box 2 pixels high. Right?
But what happens if we place such a dot just aside an mspace element
with a little width and zero for ascent and descent? The intention is
just to shift the dot a bit to the right. However, the bounding box for
such an mspace has ascent (=height) and descent (=depth) both set to
zero, and a some value strictly greater than zero as width. Now, what is
the combined bounding box of the mspace and the dot accent? Whichever
the widths of the two initial bounding boxes are, the resulting ascent
is obviously 11 (the greatest ascent among the two) and the descent is
(obviously?) 0 (again, the greatest one). Now, this box is much taller
than the box for the dot alone. The effect of this can be seen in my
widget (see the attached picture). In the first case I have the dot
alone, in the second case I have the dot with a small mspace aside.

Is my widget wrong? I mean, this is the most natural behaviour I see
now. But the result is far from being satisfactry. What if the default
values of mspace would be defined to be "undefined", so that they do not
compromise "strange" bounding boxes aside?

Regards,

	luca
<math>
  <mstyle mathsize="48pt">
    <mrow>
      <mover accent="true">
        <mi><mchar name="alpha"/></mi>
        <mo>&#x0307;</mo>
      </mover>
      <mo>,</mo>
      <mover accent="true">
        <mi><mchar name="alpha"/></mi>
        <mrow><mspace width=".3em"/><mo>&#x0307;</mo></mrow>
      </mover>
    </mrow>
  </mstyle>
</math>

Received on Wednesday, 9 May 2001 10:19:41 UTC