TMG-L Archives

Archiver > TMG > 2000-05 > 0957222502


From: Christopher Brooks <>
Subject: [TMG-L:] TMG-L: RE: Typography (WAS Font Handling Problem)
Date: Mon, 01 May 2000 19:08:22 -0400
In-Reply-To: <200005011635.LAA17755@just.doit.com>


Jeff Owens wrote:

>Quite a long explanation of what seems to be your personal observations
>and preferences.
>
>I just did a most unscientific survey. Here in my messy office I picked
>up at random, (everything's at random) six catalogs and magazines, and
>six corporate annual reports. All twelve of my highly scientific survey
>were in ARIAL typeface.
>
>What authority do you quote as your definitve pronouncement that Times
>New Roman is the world's most readable typface for English?

I'm not the one whose advice is being questioned, but based on 20+ years of
experience in publishing, typesetting, and book design, I'll offer my own
advice, FWIW.

Typography 101 teaches that serif faces are always preferred for body text.
All sorts of experiments and research over the years have shown that both
comprehension and reading speed increase with serif faces, while fatigue
and eyestrain decrease -- the reason being that the eye and brain can more
easily distinguish the letterform shapes with serifs on them. Pick up any
basic entry-level magazine on design and you'll see this advice ("use serif
for text, sans serif for headlines"), along with other truisms like, "Don't
use more than three different faces in a document."

Of course, rules are meant, in a sense, to be broken, and some exciting
(and even readable) work is effective precisely because it breaks the
rules. For instance, we might visualize a poster printed entirely in black,
with the type "reversed" or "knocked out" -- ie., noninked, thus appearing
as white (or the color of the paper). Research suggests this is harder to
read, but in the case of a poster, you're after a dramatic impact first,
and then comprehension. And in the case of website design, there sometimes
don't seem to be any criteria at all ... <g>

But for standard documents with lengthy body text, yes, a serif face, and
minimal shock value, is recommended ... :-)

As for Arial and Times Roman, they're both considered cliches in the
publishing industry. That is to say, both have been so overused for so many
years that they're considered shopworn. Arial is actually Microsoft's
copyright-evading name for Helvetica, which is the near-universal European
typeface used for public sign lettering. It can be seen on Germany's
Autobahn, and most streetcorners and subway stations in western Europe.
Attractive, fresher sans serif alternatives to Arial/Helvetica are Gill
Sans, any of the Accidenz family, or Franklin. Futura's almost as overused
as Arial/Helvetica, and I avoid it accordingly.

If all of the documents Jeff examined appear to employ Arial as the body
text face, I might suggest that in the case of non-typographically-tutored
support staff who are assigned the typing of a manual, Arial is more often
than not the first font in the alphabet found in the Windows typeface
selection box. Just as in the case of political elections, there's a
certain advantage to being "first on the ballot" in alphabetical order.

The original Times Roman was developed exclusively for The London Times,
and since then has become quite universal. It's even more shopworn than
Arial. I use Times Roman when preparing genealogical files to send to
correspondents, because I know they'll have it on their machine, but there
are much fresher alternatives available in spades for your printed work --
the more moderate of the various Garamond faces, Palatino (also called Book
Antiqua), Caslon, Perpetua, Baskerville and Bembo come to mind.

Poster Bodoni, mentioned in an earlier message, is a display face which you
wouldn't want to use for lengthy text passages. (It's also the IBM
corporate typeface). Book faces of Bodoni can be used for body text, though
it has a rather "corporate" feel to it and is probably best suited to
things like corporate sales letters and brochures.

I was surprised that the discussion hasn't touched on probably the most
important factor in readability. It's not the typeface, or the point size
-- it's line length.

Let's say that to economize on paper, I set narrow margins. The result is a
very wide line of type which the reader's eye must scan from left to right
in its entirety before returning across the page to the left margin to
continue onto the next line down. The wider the traverse, the more fatigue
in reading the document. The shorter the traverse, the easier and faster it
will be to read and comprehend.

The book publishing industry's rule of thumb is that line length, expressed
as the number of characters *and* blank spaces, should be five to six times
the point size of the body face. Thus, if you're using 12-point Quasimodo,
you'd set your margins so as to not exceed 72 characters per line. 60
characters per line, at 12 points -- a ratio of 5:1 -- is quite luxurious,
and your readers would bless you for it. If you use smaller sizes -- my
personal work generally uses 9.5 to 11 point text -- your line length
should diminish accordingly. With 9.5 point type, I like to use a
two-column format for letter-sized pages, each column with a line length of
50 characters.

You can prove or disprove this theorem for yourself. Pick up any
well-designed hardbound book from a major (not reprint or bargain basement
publisher), and count the number of characters in a line of body text. I
checked a volume of "The Great Migration Begins," the definitive NEHGS
series on Puritan emigrants to New England. At a price of something like
$135 for this set, we would logically expect high design quality, and we
get it. Line length is 72 characters.

Finally, there's no mystery to point sizes either. 72 points = one inch, so
12-point type will be 12/72, or one-sixth, of an inch tall.

Hope this sheds light rather than makes heat. :-)


Christopher Brooks
Franconia, NH

-----------------------------------------------------------------------------
TMG-L The Internet Mailing List for The Master Genealogist
To unsubscribe: Send an e-mail to with 'UNSUBSCRIBE TMG-L'
or... if you get the digest version: 'UNSUBSCRIBE TMG-L-DIGEST'
-----------------------------------------------------------------------------

This thread: