But there are tens of hundreds of Chinese characters, and a 5-by-7 grid was too small to make them legible. Chinese required a grid of 16 by 16 or bigger—i.e., no less than 32 bytes of reminiscence (256 bits) per character. Were one to think about a font containing 70,000 low-resolution Chinese characters, the entire reminiscence requirement would exceed two megabytes. Even a font containing solely 8,000 of the most typical Chinese characters would require roughly 256 kilobytes simply to retailer the bitmaps. That was 4 instances the entire reminiscence capability of most off-the-shelf private computer systems within the early Eighties.

As severe as these reminiscence challenges have been, essentially the most taxing issues confronting low-res Chinese font manufacturing within the Nineteen Seventies and Eighties have been ones of aesthetics and design. Long earlier than anybody sat down with a program like Gridmaster, the lion’s share of labor came about off the pc, utilizing pen, paper, and correction fluid.

Designers spent years attempting to style bitmaps that fulfilled the low-memory necessities and preserved a modicum of calligraphic magnificence. Among those that created this character set, whether or not by hand-drawing drafts of bitmaps for particular Chinese characters or digitizing them utilizing Gridmaster, have been Lily Huan-Ming Ling (凌焕銘) and Ellen Di Giovanni.

Draft bitmap drawings of Chinese characters for the Sinotype III font.


The core drawback that designers confronted was translating between two radically other ways of writing Chinese: the hand-drawn character, produced with pen or brush, and the bitmap glyph, produced with an array of pixels organized on two axes. Designers needed to determine how (and whether or not) they have been going to attempt to re-create sure orthographic options of handwritten Chinese, akin to entrance strokes, stroke tapering, and exit strokes.

In the case of the Sinotype III font, the method of designing and digitizing low-resolution Chinese bitmaps was completely documented. One of essentially the most fascinating archival sources from this era is a binder filled with grids with hand-drawn hash marks throughout them—sketches that may later be digitized into bitmaps for a lot of hundreds of Chinese characters. Each of those characters was rigorously laid out and, typically, edited by Louis Rosenblum and GARF, utilizing correction fluid to erase any “bits” the editor disagreed with. Over high of the preliminary set of inexperienced hash marks, then, a second set of crimson hash marks indicated the “final” draft. Only then did the work of knowledge entry start.

An in depth-up of a draft bitmap drawing of bei (背, again, rear) exhibiting edits made utilizing correction fluid.


Given the sheer variety of bitmaps that the group wanted to design—no less than 3,000 (and ideally many extra) if the machine had any hopes of fulfilling shoppers’ wants—one would possibly assume that the designers regarded for methods to streamline their work. One means they may have completed this, for instance, would have been to duplicate Chinese radicals—the bottom elements of a personality—after they appeared in roughly the identical location, dimension, and orientation from one character to a different. When producing the numerous dozens of frequent Chinese characters containing the “woman radical” (女), for instance, the group at GARF might have (and, in concept, ought to have) created only one customary bitmap, after which replicated it inside each character during which that radical appeared.

No such mechanistic choices have been made, nevertheless, because the archival supplies present. On the opposite, Louis Rosenblum insisted that designers alter every of those elements—typically in almost imperceptible methods—to make sure they have been in concord with the general character during which they appeared.

In the bitmaps for juan (娟, sleek) and mian (娩, to ship), for instance—every of which accommodates the lady radical—that radical has been modified ever so barely. In the character juan, the center part of the lady radical occupies a horizontal span of six pixels, as in contrast with 5 pixels within the character mian. At the identical time, nevertheless, the bottom-right curve of the lady radical extends outward only one pixel additional within the character mian, and within the character juan that stroke doesn’t lengthen in any respect.

The bitmap characters for juan (娟, sleek) and mian (娩, to ship) from the Sinotype III font, recreated by the creator.


Across the complete font, this degree of precision was the rule slightly than the exception.

When we juxtapose the draft bitmap drawings towards their remaining varieties, we see that extra adjustments have been made. In the draft model of luo (罗, gather, internet), for instance, the bottom-left stroke extends downward at an ideal 45° angle earlier than tapering into the digitized model of an outstroke. In the ultimate model, nevertheless, the curve has been “flattened,” starting at 45° however then leveling out.

A comparability of two draft variations of the character luo (罗, gather, internet).


Despite the seemingly small house during which designers needed to work, they needed to make a staggering variety of decisions. And each one in all these choices affected each different resolution they made for a particular character, since including even one pixel typically modified the general horizontal and vertical steadiness.

The unforgiving dimension of the grid impinged upon the designers’ work in different, sudden methods. We see this most clearly within the devilish drawback of attaining symmetry. Symmetrical layouts—which abound in Chinese characters—have been particularly tough to signify in low-resolution frameworks as a result of, by the principles of arithmetic, creating symmetry requires odd-sized spatial zones. Bitmap grids with even dimensions (such because the 16-by-16 grid) made symmetry inconceivable. GARF managed to realize symmetry by, in lots of instances, utilizing solely a portion of the general grid: only a 15-by-15 area inside the general 16-by-16 grid. This lowered the quantity of usable house even additional.

Symmetry and asymmetry within the characters shan (山, mounting), zhong (中, center), ri (日, solar), and tian (田, subject).


The story turns into much more complicated after we start to match the bitmap fonts created by totally different firms or creators for various tasks. Consider the water radical (氵) because it appeared within the Sinotype III font (beneath and on the proper), versus one other early Chinese font created by H.C. Tien (on the left), a Chinese-American psychotherapist and entrepreneur who experimented with Chinese computing within the Nineteen Seventies and Eighties.

A comparability of the water radical (氵) because it appeared within the Sinotype III font (proper) versus an early Chinese font created by H.C. Tien (left).


As minor because the above examples may appear, every represented yet one more resolution (amongst hundreds) that the GARF design group needed to make, whether or not throughout the drafting or the digitization section.

Low decision didn’t keep “low” for lengthy, in fact. Computing advances gave rise to ever denser bitmaps, ever quicker processing speeds, and ever diminishing prices for reminiscence. In our present age of 4K decision, retina shows, and extra, it could be onerous to understand the artistry—each aesthetic and technical—that went into the creation of early Chinese bitmap fonts, as restricted as they have been. But it was problem-solving like this that in the end made computing, new media, and the web accessible to one-sixth of the worldwide inhabitants.

Source www.technologyreview.com