Can a segment of a curve

be scrutinized out of

context and still be

deemed good or bad?

Yes. And when every

curved segment of every

letter possesses a rhythmic

beauty unto itself—as

well as in context with the

whole—we can be sure

that our lettering is gonna

look right. It takes a good

eye and practice, and with Jim Parkinson and Jonathan Hoefler are two type designers who do a bet-

beziers, it also takes ter job than most at finessing curves. The examples above are entirely

lots of patience, without rudeness or lapses in artistic judgment. This, I believe, is the goal.

I read somewhere that, whereas we can use
tools to mechanically construct a letter, there
are sure to be places where we've no choice
but to draw a certain portion by eye. I'm
always amused to notice those places where
the logic of a designer's geometric structure
breaks down, as in the letter fragment above,
with nonmatching perfect circles on what
should be identical corners, and the blunted
circle in the counter that totally breaks style.

Above, although the two sides of this serif
cannot, of course, be identical, they ought to
appear to have once been a matched set
prior to italicization. Instead, these serifs are
of different parentage, and they're just
ugly curves, besides. For starters, it
would be an improvement if both bezier
points, at a-b, which transition from the
straight stem into the curving serif, had been
placed at the same height.

Above left, a piece chopped off a script letter from a major foundry.
Aside from ugly, lumpy curves, the lines of the stroke don't follow
through as they loop around. In the designer's defense, I will admit
there's nothing more difficult than rendering delicate scripts. And 1
certainly was reminded of this as I tried to improve these curves,
above right, which are now reasonably beyond reproach.

Above left, the outer curve of the hump is awkward but acceptable.
The inner curve, however, is obviously flawed. My corrections in the
version, upper right, consisted mainly of making the inner curve echo
the arc of the outer as if made by a single brushstroke.

Often a curve is good or bad relative only to its opposite side. Either
inner or outer could be good, but paired as opposite edges of one
stroke, they don't properly relate. We must choose, then, which curve
to fix so the two flow as one. At a, we see the original portion of a
letter whose curves I don't like. At b, the outer curve is isolated to
better scrutinize its very subtle bumps, flat spots, dips and corners.

At c, the counter is isolated, and my shorthand notational system
indicates flaws in the counter's curves. Basically, wherever we find a
bump, flat spot, dip or corner, we are sure to find bezier points
placed out of extrema, with handles poorly adjusted. At d, the inner
curve has been fixed, the outer left alone. True, if you saw these
curves in a whole-letter context, they might make more sense. Or not.

At e, both sides of the curve
have been designed to flow
smoothly and work together.
This was fairly difficult for me to
correct and that's the point: It
takes time to make good curves!

ШщЕШшиШШ? ífi0& Ш

ARS EXTREMA

Our Master's Voice:

"During the early sixties at the Renault corporation,
where I was an engineer at the time, I went to see my
supervisor to tell him that I had found a new
mathematical method for drawing curves that would
replace all previous rough calculations and other lathed
shapes and models. He saw my project, looked at me,
and said, 'Monsieur Bezier, if your thing worked, the
Americans would already be using it.'"

—Pierre Bézier

(by permission of Christian Lavigne,

from La Sculpture Numérique)