Re: Empty set

2013/9/13 Stephan Stiller <stephan.stiller_at_gmail.com>

> Hi Philippe,
>
> I don't think so : the ellipsis shoud still use the **same** extra inter
> letter spacing in justified lines as between letters within words to make
> the texxt more visually balanced.
>
> That providing a special font-dependent ellipsis glyph can/will prevent
> the ellipsis dots from being affected by typographic tracking was my point,
> right? :-) And the exact spacing of the ellipsis dots in relation to
> ordinary inter-word spacing and in relation to closed-up dots varies (and
> note that most ellipsis glyphs look more like "..." than ". . .", so this
> interestingly makes a stylistic choice for you already); as far as what you
> wrote is a prescription, it's not crazy, but I don't see why it should be
> absolute or which authority or principles it'd be coming from.
>

I dd not speak about inter-word spacing (this cont affect the rendering of
ellipsis itself) but about inter-letter spacing. If the ellipsis is a
single glyph mapped from a single character, the spacing between dots will
not vary with justification.

The single-character ellipsis is also not visually appealing in many
monospaced fonts (with aspect ratio about 0.5em to 0.7em), as they force
dots to become smaller than standard dots. That's why old typewriters did
not have an ellipsis character as they looked too much like an underscore
or dots could become invisible or incorrectly joined line an underscore on
baseline, far from the, ou would have beenso thin that they would have
pinned holes across the paper.

Dots need some surface, similar to the surface created by a circle
approximaly with the same radius as the the standard dot, and visually the
same surface as a square with its width equal to the width of legs of a
Latin small letter m

The dots being drawn with a minimum spacing distance between each other at
least like legs of this letter m, or the dots over "iii" (in a sans-serif
font). But if interletter justification applies, the same 3 letters 'iii"
will vary the distance of these upper dots and the ellipsis shoudl as well
follow the same visual distance (but still keeping their weight), to
preserve a balanced blackness in the fully justified line of text.
Interletter spacing is important for correct and readable rendering in
narrow columns (less than about 20em with fonts whose average letter width
is about 0.5em to 0.7em, i.e. on columns for less than 40 letters or
whitespaces). using only interword spacing gives poor result and should not
increase the width of normal word-separating spaces to more than about
1.5em to 2em (roughly the interline height) :

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).

Now if I spoke about dotted rulers (". . . ."), it is because they
are very different : they are not influenced by line jsutifications or by
adjusments of the interletter spacing, they keep the weight of their dots
(which may still be thiner than punctuation dots, for periods terminating
sentences or abbreviations, or for ellipsis), and these ruler dots are
vertically aligned using equal steps measured from the right margin of the
tabular column, to create a regular grid-like pattern. The dots of the
ellipsis on the opposite are positioned horizontally relative to the right
side of the characters before them.

Personnally I find the single-character ellipsis very ugly within justified
paragraphs with proportional fonts (as they don't preserve the visual
blackness of characters on the line), or within paragraph rendered
monospaced fonts (with or without justification, but if justification is
enabled and possible it is even worse !)

I will say the same thing if a font or renderer attempts to treat the
3-dots characters as as single ligature with fixed spacing between dots, or
in similar cases like when rendering "!!!" or "!?".I also think that the
non-breaking narrow space before some punctuations (notably in French for "
;" or " :" or " !" or " ?", or within angle quotation marks, shoudl also
apply the interletter spacing between the character components of these
sequences, even if they don't apply any variation to the inter-word
spacing). If justification narrows the standard inter-letter spacing (i.e.
reduce the default side bearings), it should as well narrow the spacing
between dots of the ellipsis. In other words,

----
Conclusion:
I'm convinced that the ellipsis is 3 characters and not just 1,even if
there should never exist any line wrap between them; they should behave
visually in a way similar to the unbreakable digraphs in words (e.g. "qu"),
 but not in a way similar their ligature, e.g. in "ts" and similar
unbreakable orthographies for affricates within the same morpheme).
If one wants to really avoid this expansion of ellipsis during
justification (cases that should be rarely desired), we could use
<period,ZWJ,period,ZWJ,period>, possibly surrounded with (unbreakable)
fixed-width narrow spaces, to hint their behavior as a single (unbreakable)
ligature. But may be this is the (rare) usage intended by the
single-character ELLIPSIS (but we know the caveats of this character if it
is mapped with a single glyph in monospaced fonts)