LucidaBright kerning of uc letters

Lucida fonts and lucimatx macros topics

Moderators: PTIForAdmin, WaS

Post Reply
TomK
Posts: 16
Joined: Thu Feb 09, 2006 4:34 am
Location: Ljubljana, Slovenia

LucidaBright kerning of uc letters

Post by TomK »

Capitalized words typeset in LucidaBright font using LaTeX's lucimatx
and amsmath packages exhibit visually irregular intra-letter spacing.
E.g., in the word {\large MATHEMATICS} letters are very tightly spaced
(much more so than in CM fonts) except for noticeable gaps between A and
T; the effect is as if the word was written on a typewriter with a
sticky carriage. This elicits the question if the kerning data
were properly implemented.

In response, the author of lucimatx package offered the following
dubious explanation: "Kerning of upper-case letters is always crucial.
There is no font in the world, which would yield satisfactory results
with _any_arbitrary_word, regardless of how elaborated its kerning data
are. The Lucida text fonts are somehow special: They were designed to
go without any kerning data at all."

On the other hand, in his "Travels in TeX Land: Using the Lucids Fonts"
David Walden claims, on p. 2, that Y&Y created "the necessary TeX font
metrics (.tfm) for all Lucida fonts, including the fussy details TeX
needs for math typesetting."

Since these statements are contradictory, I wonder who is correct.
PTIForAdmin
Posts: 84
Joined: Thu Oct 06, 2005 10:08 pm
Location: San Francisco, CA
Contact:

Lucida UC kerning

Post by PTIForAdmin »

A comment from the font designer:
Concerning kerning of uppercase letters,
it is true that the Lucida Bright capitals are fit
rather tightly. This goes back to the origins of the
design for magazine typography (Scientific American).
For ALL CAPITAL setting, my recommendation would be
to add a little space between the letters, which used
to be standard good typographic practice, and then
tighten up only the combinations that need it, like,
say, AV LA TA and so on.
WaS
Posts: 27
Joined: Tue Feb 07, 2006 7:13 pm
Location: Erlangen, Germany
Contact:

Re: LucidaBright kerning of uc letters

Post by WaS »

TomK wrote:Capitalized words typeset in LucidaBright font using LaTeX's lucimatx
and amsmath packages exhibit visually irregular intra-letter spacing.
E.g., in the word {\large MATHEMATICS} [...]
This elicits the question if the kerning data were properly implemented.

In response, the author of lucimatx package offered the following
dubious explanation: "Kerning of upper-case letters is always crucial.
There is no font in the world, which would yield satisfactory results
with _any_arbitrary_word, regardless of how elaborated its kerning data
are. The Lucida text fonts are somehow special: They were designed to
go without any kerning data at all."
Using the default kerning data (which happens to be none in Lucida),
the word MATHEMATICS looks indeed ugly when typeset in Lucida Bright,
as compared with Computer Modern. However, the result from using CM
isn't the optimum, either.
Typesetting capitals needs almost always some manual intervention, which
depends on the particular circumstances, so I'd say that Lucida does in the
average
not perform worse than most other font familiies.
TomK wrote: On the other hand, in his "Travels in TeX Land: Using the Lucids Fonts"
David Walden claims, on p. 2, that Y&Y created "the necessary TeX font
metrics (.tfm) for all Lucida fonts, including the fussy details TeX
needs for math typesetting."
Since these statements are contradictory, I wonder who is correct.
D. Walden's quotation does not contradict mine. The tfm files for the Lucida fonts
do include all the information required by TeX, in particular for mathematical
typesetting. Kerning data, however, are not TeX-specific. The deficiency seen
above would affect non-TeX applications as well. Yes, you may blame
B&H for not having provided a few obvious kerning pairs between capitals, but
even if they'd exist, they would -- in the general case -- not eliminate the need for
further manual corrections, anyway. And creating these data subsequently would
require too much effort, as compared with the benefits.

I hope I was able to make things a bit clearer.

Walter
TomK
Posts: 16
Joined: Thu Feb 09, 2006 4:34 am
Location: Ljubljana, Slovenia

Post by TomK »

In a private email, WaS suggested to me to have a look at TeX/LaTeX 'soul' package (found at CTAN's tex/macros/latex/contrib/soul); the document soul.pdf is an elaborate description of the kerning problem, and its 'soul' solution.
Post Reply