Re: [CSS21][css3-text] letter-spacing, word-spacing and justification

On 07/06/2013 05:07 PM, Andrew Cunningham wrote:
> There is a sleeper here. Assuming that letter-spacing works as identified,
> there definitely needs to be a way of disabling inter letter spacing.
>
> For complex scripts that are not currently not speced out in OT specs,
> are supported by default rendering. Which usually means using mark,
> rlig, clig, and liga features.

'mark' should be alright, since diacritics and such should be part of the
same grapheme cluster as the other characters they combine with. Right?

Don't see why liga would be used in place of rlig in cases where the
complex script would require it. That seems like a bad choice. The
spec doesn't say to break required ligatures, only optional ones.

clig could indeed be a problem, if it's needed across grapheme clusters.

Are there situations where such ligatures occur across grapheme clusters?
The spec forbids spacing within a grapheme cluster.

> At least one rendering engine used by at least one browser will disable
> optional features during letter spacing, breaking rendering.

Indeed, the spec requests that optional ligatures be broken when
letter-spacing is not zero. I think this is appropriate for explicit
tracking: you can't get even spacing in a word like "find" otherwise,
if "fi" ligatures are enabled.

> As it is, in places where letter-spacing is non-zero, it is sometimes
> necessary to reenable specific font features for certain fonts.

~fantasai

Received on Friday, 12 July 2013 00:29:00 UTC