Encoding italic (was: A last missing link)

wjgo_10009@btinternet.com via Unicode unicode at unicode.org
Tue Jan 15 09:48:15 CST 2019


You are the gentleman who kindly made the Gentium typeface open source.

  Thank you for your generous gift to the world.

  > Use of variation selectors, a single character modifier, or 
combining characters also seem to be less useful options, as they act at 
the individual character level and are highly impractical. They also 
violate the key concept that italics are a way of marking a span of text 
as 'special' - not individual letters. Matched punctuation works the 
same way and is a good fit for italic.

Italics works differently from matched punctuation marks in that with 
italics there is a change to each glyph whereas with matched punctuation 
there is no change to the glyphs between the matched punctuation marks.

That difference leads to the significant difficult that there are thus 
two competing forces here.

One of those forces is what you have stated about the nature of italics. 
The other of those forces is that Unicode is not stateful.

Years ago I encoded some Private Use Area codes for such features as 
italics, with a start character and an end character to surround a span 
of text that would then be rendered in italics.  As a result of 
discussion and advice I learned that such characters are not acceptable 
for encoding into regular Unicode because the effect would be stateful. 
So yes, the method that I suggested and for which James Kass suggested 
an enhancement is peculiar when viewed against the theory of the way 
that italics are used, but neither the method nor the enhanced method is 
stateful and that is an important feature of them.

Now it would be possible for a software application program to have a 
feature for composing plain text where a span of text may be highlighted 
by a user of the software application program and every character 
(except perhaps spaces?) within that span of text has, at the click of a 
button, a VS14 character inserted after it.

I remember that when handsetting metal type the same space sorts were 
used with italics as with roman.

There could also be a button that could remove all VS14 characters, if 
any, from within a highlighted span of text.

So, for someone typesetting plain text and viewing plain text the effect 
could look to be in accordance with how you consider italics should be 
encoded, though for plain text interchange the encoding would still be 
by using a VS14 character after each character that one wishes to become 
displayed italicized.

William Overington
Tuesday 15 January 2019

More information about the Unicode mailing list