L2/03-028

Title: Response to "UTC Agenda Item: Scope of Enclosing Marks"
Source: Asmus Freytag
Date: January 30, 2003

I find the proposed scope of Enclosing Marks very disturbing.

At 01:09 PM 1/30/03 +0000, Michael Everson wrote:
At 18:00 -0800 2003-01-29, Mark Davis wrote:

1. Given the sequence "1" + grapheme_joiner + "2" + enclosing_circle, should
the circle enclose the three previous characters or only the "2"?

All three.

I disagree. To set up the expectation that all reasonable renderers will  
support this mechanism is unrealistic. Since it's unrealistic, no user can  
rely on that text encoded in this way will be interchangeable. This is  
particularly true for the case of the digits ordinarily rendered with  
minimalistic rendering engines that are optimized for performance.


2. Given the sequence KA + VIRAMA + DDHA, should the circle enclose the
three previous characters or only the DDHA?

Definitely all three. You can't break up Indic orthographic syllables by  
ringing bits of them.

This is more realistic. Indic renderers need to do a number of complex  
things already and will know about KA + VIRAMA + DDHA. However, the task of  
scaling and positioning the circle to reasonable effect are not trivial.

Before we establish a definite mandate here, I would like to see a  
detailed description of what pieces of technology need to be invoked and  
what the implication(s) are of adding this to renderers and fonts.

Conclusion
Complex behavior, like circling more than one character should not be  
mandated by Unicode conformance. It should be left to higher level  
protocols and/or to best practice. Maintaining the second requirement would  
possibly result in some implementations taking the way out, by declaring  
that they simply don't support the combining circle etc. Maintaining the  
first requirement would have the result in their claiming the same thing  
for Grapheme Joiner in rendering. So we should not make either a  
requirement.

However, since circling the DDHA in isolation, or circling the 2 in  
isolation is not optimal, we should not *require* that the action of the  
combining circle be *limited* to that. Rather, those applications that can  
handle complex cases should be allowed to display something sensible in  
these cases. This makes the situation more similar to other cases, where we  
only mandate some minimum, but allow finer typography to do more  
sophisticated things.

A./