From: Asmus Freytag <asmusf_at_ix.netcom.com>

Date: Mon, 09 Jan 2012 12:39:35 -0800

Date: Mon, 09 Jan 2012 12:39:35 -0800

Philip,

In your text, you write:

Geometric shapes are normally centred on the math axis, and may be

presumed to be independent of the baseline. The height of the math axis

above the baseline, however, and consideration of caps height, x-height,

etc, can be expected to produce differences in the way different fonts

implement geometric shapes.

This may be true for geometrical shapes used in a mathematical context,

but might not be universally the case.

I've just finished looking at a set of symbols where some shapes are

deliberately aligned to the baseline (and because they are not tall,

this makes an obvious difference). As an example, a small circle

centered on the baseline is not the same symbol as a small circle

centered on the math axis.

As a general principle in ecoding symbols, not only geometric symbols,

it's not enough to match up the "ink" (or worse, merely the basic

strokes and there connectivity or intersection). On the contrary,

details of positioning, weight and proportion must also be considered as

important - unless it can be positively demonstrated that they are the

kinds of artifacts of font implementation.

In other words, unification "at arms length" is as inappropriate for

symbols as it was for ideographs, except that for symbols, one needs

careful analysis of each context to understand what the unification

criteria are in each case. This is made more difficult where samples are

"polluted" by bad implementations or people having used "fallback"

representations for lack of other alternatives.

In the case of the Web and Wingdings, there's the added problem that the

goal is to create a compatibility encoding for the set of shapes (more

or less) as found in the fonts. In that context, unification has to be

approached with considerable caution.

I haven't had the time for more than a quick look at your document, so I

may have more and more detailed comments later.

A./

On 1/9/2012 11:59 AM, philip chastney wrote:

*> A little work on the size tables given on page 3 of ISO/IEC
*

*> JTC1/SC2/WG2 N 4115, Proposal to add Wingdings and Webdings
*

*> Symbols, shows that the two tables used in N4115 (for diamonds
*

*> and squares) are near enough identical to each other, and to the
*

*> range of sizes illustrated in Table 2.5 in UTR 25 (Unicode and
*

*> Mathematics).
*

*>
*

*> This suggests that the classification “slightly small”, introduced in
*

*> N4115, is redundant, as are the 2 “slightly small” shapes, 1F7D7
*

*> and 1F786.
*

*>
*

*> the details are given in
*

*>
*

*> http://www.chastney.com/~philip/shapes/slightly_small.pdf
*

*>
*

*> regards . . . /phil chastney
*

Received on Mon Jan 09 2012 - 14:42:28 CST

*
This archive was generated by hypermail 2.2.0
: Mon Jan 09 2012 - 14:42:29 CST
*