Re: [css-text] Issue 18: character-based alignment in table columns

On Tue, Jan 26, 2016 at 2:55 PM, Alan Stearns <stearns@adobe.com> wrote:

> On 1/27/16, 6:36 AM, "Tab Atkins Jr." <jackalmage@gmail.com> wrote:
>
> >
> >This sounds like those tools just not considering the case of the
> >texts being in different fonts or faces.  I don't think there's a good
> >reason to copy that bug.  You clearly want the texts *aligned*, and
> >aligning centers seems most likely to get you that result.
>
> I agree. It’s likely this is an edge case that just never got considered
> or was never considered worth the time to address. In InDesign or TeX, when
> the problem crops up you can tell the user “If you want those things
> aligned, use the same font and size.” On the web that’s often not an
> option, so we can handle this edge case better.
>

I'm all in favor of making CSS better than InDesign. I'm not so sure about
the right thing to do. Let's look at the most common case of a bold row in
a decimal-aligned table. In Adobe Caslon Pro, the em-squares of both roman
and bold numerals are 500 units wide (1000 units/em). The em-square of the
roman period is 254 units wide, but 280 units wide for the bold period.

So where does that extra space go? In InDesign, numerals to the right of
the alignment character are pushed right a bit. This leaves numerals to the
left unaffected.

If we center the alignment character, then numerals to the left shift a
tiny bit right, and numerals to the right shift a tiny bit left.

I think the InDesign behavior might be more desirable, given that numbers
to the left are literally more important (and there are usually more of
them).

Note I'm assuming that we're not messing with the text layout itself, i.e.
not kerning around the alignment character to achieve some particular goal.
And the lengths we're talking about are rather small... .026em is not very
much. Of course if you're trying to align across significantly different
font sizes (etc.), you're gonna have a bad time.

Dave

Received on Wednesday, 27 January 2016 18:42:12 UTC