furuta@BRILLIG.UMD.EDU (Richard Furuta) (10/07/87)
Recently it came to my attention that the redesign of Scientific American was guided by Bigelow and Holmes. I asked Chuck Bigelow for more details about it and got the following message in reply. I found the message to be extremely interesting and believe that laser-lovers will enjoy seeing it as well. The redesign of Scientific American marks the first use of the Lucida family in a magazine setting. Readers of laser-lovers may remember earlier discussion of Lucida as available on laser printers (in particular those sold by Imagen). --Rick |Date: 03 Oct 87 1459 PDT |From: Chuck Bigelow <CAB@SAIL.Stanford.EDU> Scientific American typography Beginning in the September issue, Scientific American has a new typographic design and new typefaces. The October issue on "The Next Revolution in Computers" shows even greater use of the new types. The changes include new layouts for the table of contents and the departments -- including Science and the Citizen, and complete revision of the typefaces throughout the magazine -- titling, text, captions, and illustrations. For those who care to know more about such matters, here is a discussion of how the typography and types were planned and executed. The project took over a year, so many details are left out of this relatively short account, but there should be enough information to satisfy those interested in typographic matters. The last time Scientific American changed design was in March 1976 (which featured an article about "The Small Electronic Calculator"). At that time they changed from Didot titling faces to Garamond, and from Caledonia text to Times Roman. The reason for that change was a switch from metal to digital typesetting. The elegant Didot designs were unavailable in digital form (and remain so) and the Caledonia was available but looked bad. Over the past two years, Scientific American had been considering another re-design that would take better advantage of its digital typesetting technology and be more accessible to a wider range of readers, without losing the qualities of formality, clarity, orderliness, and even elegance that have characterized it over the past four decades. In August 1983, I co-authored an article on "Digital Typography" for Scientific American, for which most of the typographic illustrations were executed by Kris Holmes. When it came time for a re-design, we were chosen to develop it in cooperation with the magazine's editorial, art, and production directors. In doing the re-design in our studio, Kris Holmes and I were assisted by Carol Twombly, a graduate of Stanford's digital typography program. In some areas, notably the Table of Contents, we adopted an "iconic" or pictorial approach, using illustrations to complement the titles and abstracts of articles. This replaced the previous "bookish" Table of Contents. In the departments, we re-introduced "icons" or pictographs in the titling, harkening back to the Ben Shahn illustrations that enlivened those sections in the classic S.A. of the fifties. The new pictographs were developed by the S.A. art department and an outside artist. The overall layout of the articles remained much as before. Though we experimented with various changes to the articles, relatively few were ultimately adopted. The Science and the Citizen department was extensively re-worked, both from an editorial as well as a typographic perspective, and many of the design ideas here came from the magazine's own art department. The typefaces are completely new to the magazine (or to any magazine) and most of them we designed especially for it. The titling roman and italic used for the main article titles and subtitles were changed from Garamond to Galileo, a new kind of "modern". Galileo is in the style of Bodoni, Didot, Walbaum, and yes, Computer Modern, but is different from all of them. The stems are slightly flared, and the serifs slightly flared and bracketed, to have a more lively action than the standard "modern" or "didone" style faces. (For a discussion of the Modern style, see my note on "Modern Typefaces" in Donald Knuth's Computers & Typesetting, Vol. E.) Galileo does have pronounced difference between thick and thin, like the modern faces of 19th century typography still associated with scientific publishing. The Galileo italic is slightly more unusual. Instead of curling up in the normal cursive style seen in most italics, the serifs are horizontal though asymmetrical. Galileo italic shares this trait with Lucida italic. The horizontal serifs give a different texture than the common slanting serifs. The actual letter shapes are cursive, not slanted romans, though the `a' and `g' are "humanist" or "two-storied" rather than "chancery" or "one-storied". These forms were chosen to make the subtitling slightly more legible, since it is more difficult to confuse the humanist forms with other letters. The Galileo capitals are rather narrow to help fit long capitalized titles on one or two lines. Their narrowness is usually not noticed in the normal titling, but when all-capital titling is used, as in Science and the Citizen, they seem somewhat condensed. Therfore, when the opposite goal is intended, to make a few words stretch over an entire line, Galileo small capitals are used. These are wider and very slightly heavier, in proportion, than the capitals. The text was changed from Times Roman to Lucida Bright, a sharpened version of the Lucida available for laser printing. (See "The Design of Lucida: an Integrated family of Types for Electronic Literacy", by C. Bigelow and K. Holmes, in Text Processing and Document Manipulation, J.C. van Vliet, ed., Cambridge U. Press, 1986.) Although Lucida was first designed for laser printing, most of its features made it easily adaptable to high-resolution digital typesetting. The major changes were: thinning of hairlines and serifs, lengthening of serifs, tightening of letterspacing, and slight emboldening of capitals. The term "Bright" for the new design was suggested by the (former) art director of Publish! magazine. In a sense it is brighter, because the spectrum of spatial frequencies of text set in the design will have more high-frequency components, of higher energy. Or at least that is conjectured. We haven't done Fourier analysis of it. Yet. Lucida has such a large x-height that 8.5 point Lucida looks larger than 9 point Times Roman. The character count per line remained approximately the same. Thus, we achieved a sense of larger type without giving up much economy, and even added a little more leading to let a little light into the text columns. The text is thus set in 8.5/10.2 point type on a 13 pica measure. The Lucida capitals are rather small and light, and thus can be used in acronyms withoutt causing odd, dark tangles on the page, but we also designed true small capitals, about the same weight as the lower-case and very slightly taller. We also adjusted the use of initial drop-capitals in the text, for the first time they correctly align with the base-lines. The captions are composed in Lucida Bright Demi-bold, as are the sub-heads in the "long" article in most issues, except the special October issue. The Demibold was fine-tuned to exactly the density that the art direction liked best; stems are approximately the square-root of two times the normal weight stems. The cover now includes text set in Demibold Italic. For the labels in illustrations Lucida Sans replaced Helvetica. The Lucida Sans capitals are slightly narrower than Helvetica caps, have more differentiation of forms for easier recognizeability, and more inter-letter spacing, for an easier, less crowded rhythm. When italic or bold face are need (see p. 92 in the October issue), Lucida Sans Italic and Demibold are used. The Lucida Sans Italic is a true cursive, with different forms than the roman. If it were uprighted, it wouldn't look like the roman. In this way, Lucida is different from most sans-serifs. The italic is actually based on the chancery cursive style of the Renaissance, and so could be called a "chancery sans". This kind of italic can be more easily discriminated from the roman, and hence aids readability. It also has a more dynamic rhythm. A subliminal advantage of using Lucida and Lucida Sans together is greater harmony of letterforms -- text, captions, and labels in illustrations share certain underlying letter forms, and thus there is greater graphic harmony between them different typographic units. The special versions of the types were designed and drawn by traditional methods, and then digitized with Ikarus software (from URW) on a VAX system. The basic characters were digitized as spline outlines in the Ikarus format (a hybrid of Hermite cubic and arc & vector, see Peter Karow's book, Digital Formats for Typefaces) and these were scan-converted to high-resolution raster data (from 200 x 240 to 800 x 960 pixels per em square -- yes, a rectangular pixel aspect ratio on the Videocomp -- depending on the range of sizes each font was intended to set). The raster data was converted to run-length code used by the Videocomp, a venerable digital CRT typesetter, based on a device originally designed by Dr. Hell in Germany in the late 1960's, and still used to typeset many magazines, including Newsweek, Time, Scientific American, and others. Scientific American is typeset at approximately 1200 - 1800 dots per inch, but even at these high resolutions, the raster font data was improved by scan-conversion preceded by "conforming" the outlines to the raster grid by using routines to coerce the extreme points on the contours to have "nice" integer or half-integer grid values. The designs were proofed on an Imagen 8/300 laser printer for testing of weight, spacing, shapes, etc., but actual proofing on the typesetter and a test print run was neeeded to test all features. Even so, typography and type design is often a welter of details, and the designs may change in small details as we review their usage over the next few months. --Chuck Bigelow