Re: ligature usage - WAS: How do we find out what assigned code points aren't normally used in text?

From: Jukka K. Korpela <>
Date: Tue, 13 Sep 2011 21:33:35 +0300

13/09/2011 00:42, Stephan Stiller wrote:

> On 9/12/2011 10:46 AM, Philippe Verdy wrote:
>> common-ligatures
>> Enables display of common ligatures (OpenType feature: |liga|). For
>> OpenType fonts, common ligatures are enabled by default.
>> This means that German documents will really need to use ZWNJ
>> (fortunately, this character should soon become standard on German
>> keyboards, and CSS3 would be a good motivation for including this key
>> mapping) for common ligatures like fi,fl, ff, ffi, ffl, ſt, or even tt...
> It would be nicer if the user were, "by default", offered a choice.

I’m afraid this discussion, though on-topic in my opinion, has become
rather specialized and technical (in terms of web techniques) for this
list. I share Philippe’s concern for the change: changing the way
browsers work in rendering texts is not a good thing when it changes the
_default_ behavior.

Even if a change, like using a ligature for “fi,” might be an
improvement in the average, that’s not enough. There are too many things
that may get broken that way—even if we don’t consider drastic (yet
realistic) issues like intentionally monospace text.

But I don’t think the Unicode Consortium, or the community supporting
Unicode at large, could make a useful move in this issue. It really
calls for common sense, rather than anything else, from browser vendors
and CSS specs authors to realize that the default rendering should be
left intact, as there are too many potential parameters to consider.

I see this primarily as an _author_ choice. The user should have the
last word, as he has if he really wants that, but for the most of it,
typographic issues like ligatures are not something that users can and
will deal with. Authors can be expected to do that, if they care, and it
should not be too much of a burden to write an author stylesheet that
suggest ligature behavior for all text, if that’s desirable and possible.

Received on Tue Sep 13 2011 - 13:36:38 CDT

This archive was generated by hypermail 2.2.0 : Tue Sep 13 2011 - 13:36:39 CDT