Re: Empty set

From: Philippe Verdy <>
Date: Fri, 13 Sep 2013 16:38:10 +0200

My opinoon anyway is that the single character ellipsis only exists in the
UCS for backward compatibility with legacy encodings, that were initially
designed at time where they were used on terminals with limited graphic
capabilities and using only monoospaced fonts (including for tetext,
broadcasted and display on old low resolution analog TV, displaying only
24/25 rows with about 40 character cells), or old computer monitors
displaying only monochromatic text in standard 25x80 cell grids, or old
printers producing voluminous listings for mainframes, or data input and
storage on disks or bands with limited capacity or bandwidth.

Only to save a couple of cells three dots in the same cell allowed
capacting text. It just remained readable, but not typographically good.

But you make a point with East-Asian (CJK) scripts rendered in monospaced
grids: using a full cell only to display a single dot is a waste of space
(cells are larger for these scripts with their many strokes) and changes
dramatically the speed of reading for runs.

But then the existing ellipsis is not a good candidate because it has the
incorrect metrics where it should use the sinographic metrics.

And that's why we have monospaced fullwidth or half-width variants encoded
for common Latin, Greek or Cyrillic alphabets, and some precomposed
Latin-based symbols like measurement units, or common abbreviations based
on Japanese Kanas. Hangul is different because its composition square
already allows composing some squares with base Jamo letters, but it will
also use some common Latin-based abbreviations or symbols.

They also have different versions for punctuation signs, that better fit
the East-Asian composition square. But the encoded ELLIPSIS does not fit
correctly there. These scripts are also not concerned the same way by
features like justification, "tracking" or by glyph scaling modifying the
aspect ratio. These scripts also are much more tolerant to line breaks, and
don't need complex hyphenation rules as you can break them after almost all
squares (most of these squares encoding a full syllable, and many words
being monosyllabic). The rare punctuation signs are rarely needed in
sequences so they could make an exception for the ellipsis compacted in the
same composiition square.

Minor problems may only occur with long numbers written with positional
digits, or with text mixes that are alternating runs of half-width
characters with runs written in full-width squares, or with
paragraph-ending punctuations that would be orphaned after a line break
(but I don't think they will alter the vertical alignment of squares (for
horizontal lines). But traditional vertical rendering or artistic painted
calligraphy would not care about these alignment constraints across
successive rows and aligment of final cells on the right or bottom margin.

2013/9/13 Stephan Stiller <>

> Once you've increased the width of these interword spaces to their
>>> maximum, all the characters (and these increased spaces) should be
>>> justified using interletter spacing, and this extra interletter spacing
>>> should be applied as well between the dots of the ellipsis (showing that
>>> they are effectively 3 separate characters and not just one with a fixed
>>> distance between dots).
>> You are right that tracking and glyph scaling exist,
> This is not glyph scaling
> Because it's called tracking.
> if the dots are circular, they remain circular.
> When *glyph scaling* is performed, it's by up to 2% (recommends
> Bringhurst). That'll make circles ovals (different meaning of "ellipsis").
> And noone's gonna notice.
> you may argue that the problem is similiar with the dots or diaresis or
> Hebrew and arabic points, but these dots are less essential for correct
> understanding of text on low resolution medias, or prints on basic paper
> quality and with low cost inks.
> Well, I'm not sure about Arabic i‘jām not being so important for visual
> letter discrimination :-)
> Btw, different aspect of the topic: math and Chinese are different; for
> those separate ellipsis glyphs make sense.
> Stephan
Received on Fri Sep 13 2013 - 09:41:36 CDT

This archive was generated by hypermail 2.2.0 : Fri Sep 13 2013 - 09:41:36 CDT