JC> In doing so, I discovered that fixed, aka:
JC> -misc-fixed-medium-r-semicondensed--13-120-75-75-c-60-iso8859-1
JC> has an ASCENT of 11 and DESCENT of 2 (totalling 13), whereas
JC> -b&h-Luxi Mono-medium-r-normal--[9.75 0 0 13]-0-75-75-m-0-iso8859-1
JC> while intended to match those metrics ends up with an ASCENT of 13 and
JC> DESCENT of 3.
Please, not that question.
In short, the font's point size, ascent and descent are three random
values that depend on the font, the font technology, and even the
individual foundry or font designer. No particular relationship
between the three values should be expected.
As you probably know, a molten lead fo(u)nt is a collection of blocks
of lead with glyphs embossed at the top. In order for the type to
align neatly in the galley, all the blocks in a single font have the
same dimension in the direction parellel to the paper but orthogonal
to the direction of writing; this dimension is known as the font's
/point size/, or /em size/, or /quad size/.
Obviously, it is physically impossible to typeset two lines of 10 pt
Garamond less than 10 pt away; on the other hand, nothing prevents you
from typesetting them 12 pt away by inserting 2 pt wide strips of
lead; you then speak of typesetting 10 pt Garamond with 2 pt /leading/,
or sometimes 10/12 pt Garamond.
Thus, with molten-lead fonts, for a given typeface, the choice of the
point size is merely a technical decision, and very similar fonts can
have different point sizes depending on the technical tradeoffs made
by the foundry.
While there have been attempts to rationalise this nowadays obsolete
notion (most notably by Knuth), it has carried over to digital fonts.
Nothing should be expected from the point size except that within a
single family produced by a single foundry the size of glyphs should
be roughly correlated with the point size.
What about the font's ascent and descent? As opposed to a glyph's
ascent and descent, which are well-defined notions, these have no
meaning at all. They are arbitrary values that are produced mostly at
random by the various font backends.
In the case of the TrueType fonts in the FT1 or FT2 backends, they are
merely scaled versions of values that are found in the font file --
randomly chosen by the font designer. For Type 1 fonts produced by
FT2, they are scaled versions of the font's bounding box (and thus
depend on the font's glyph coverage, but not the encoding). In the
case of the Type 1 backend, they are the max of the scaled values of
the bounding boxes of the glyphs present in the encoding (and thus
depend on the encoding). In the case of bitmap fonts, they are
just taken from the font file.
The only reasonable use of the font's point size is to allow the
(human) user to specify it. The only reasonable interpretation of the
font's ascent and descent is to use them as a user-tweakable default
for the positionment of baselines.
Juliusz
_______________________________________________
Fonts mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/fonts
