Organizing Principles within the Occidental Glyphs

Contents

• Overview
• Division by Stroke Count
• Division by Size
  • Size Overview
  • Size Cartographic
  • Size Indexical
  • Size Normal
• Division by Tapering
  • Overview
  • Taper0
  • Taper1
  • Taper2
• Division by Typeface
  • Overview
  • Lineface
  • Roman
  • Italic
  • Script
  • Fraktur
  • Blackletter
  • Lombardic
• Application by Character Set
  • Character Sets vs. Glyphs
  • Latin Letters
  • Arabic Digits
  • Greek Letters
  • Cyrillic Letters
  • Symbols, Punctuation, Etc.

1 - Overview

The Hershey occidental glyphs are numbered is a sequence that runs, with gaps, from 1 to 3926 (or 3999 or 4000 to round it off). While their arrangement has considerable internal order, it is not related to the order of the ASCII character set. Subsequent additions have filled in some of the gaps, and (in the GNU Plotutils package [gp241]) added glyphs above 4000.

Dr. Hershey [Hershey 1967], [Wolcott & Hilsenrath 1976], Peter Holzmann [Holzmann 1986], and the GNU Plotutils [gp241] employ a number of not necessarily mutually consistent terms to describe the glyphs. I've adopted some of these here, but have omitted some, changed some, and have introduced new terms. I've done so to try to increase clarity, but I'm sure that I've just introduced confusion.

My basic premise is that the style of each range of Hershey glyphs can be defined by four terms: stroke count, size, tapering, and typeface. Given these divisions, each range may then be identified further by the character set is provides.

Note that the "ranges" identified here are not fonts. A "font" is a particular collection of glyphs, presumably in a consistent typeface, arranged to satisfy the needs of a particular character set (a definition such as this applies equally well to cast metal and digital fonts). They aren't typefaces either. These are just subdivisions of the Hershey Occidental glyph repertory by various features.

2 - Division by Stroke Count

The basic division of the Hershey glyphs is stroke count: 1, 2, or 3. This describes "the number of lines which are used in parallel to obtain a variation in line thickness." [Hershey 1967, p. 3], This division is even more basic than size, and can form the basis for an overall understanding of the glyph sequence.

In "Calligraphy for Computers," [Hershey 1967, p. 3], Dr. Hershey uses the terms "uniplex," "duplex," and "multiplex" for these stroke ranges. Wolcott and Hilsenrath [Wolcott & Hilsenrath 1976, p. 4] use the terms "uniplex," "duplex," and "triplex." I'll follow that usage here. The stroke count ranges in Dr. Hershey's original layout are continuous: uniplex from 1 to 1000, duplex from 1001 to 3000, and triplex from 3001 to 4000. However, the glyphs added by Wolcott at the NBS to the uniplex range (238-284 (with gaps) and 910) contain some duplex glyphs. The additional glyphs above 4000 contain all manner of things.

3 - Division by Size

3.1 - Size Overview

As a second descriptive term, the glyphs may be divided into three sets based on their capital height (or in the case of some symbols the capital height of the alphabetic characters best associated with them).

In "Calligraphy for Computers," [Hershey 1967, 3] Dr. Hershey identifies three glyph heights: 9 raster units (used for FORTRAN and Cartographic glyphs), 13 raster units (used for "indexical lines of print"), and 21 raster units ("available for principal lines of print.") [Wolcott & Hilsenrath, p. 4] simplify this terminology into: "Cartographic," "Indexical," and "Normal." These are the terms I'll use here.

3.2 - "Cartographic" Size

The "Cartographic" size, encompassing glyphs in the range 1 to 500 (all Uniplex, therefore), is the smallest. In the original Hershey distributions, contained only capital letters (in Roman and Greek), numerals, and a restricted set of symbols. Later distributions added an Italic typeface as well as lowercase and more symbols.

Although this range is termed "cartographic," it does not include a number of cartographic symbols included in other ranges.

In [Wolcott & Hilsenrath, p. 24-25], Wolcott (identified as the originator in [gp241]) adds symbols in the range 238-284 (with gaps) which are Normal, not Cartographic in size.

1.3. - "Indexical" Size

The "Indexical" size covers the range 1001 to 2000. All Indexical glyphs are Duplex, but the Duplex range extends beyond into the "Normal" size range.

3.4 - Normal Size

The "Normal" size appears in all three stroke ranges: Uniplex (501-1000), Duplex (2001-3000), and Triplex ( 3001-4000), as well as the "NBS" glyphs and glyphs above 4000. Only the Triplex stroke range is exclusively Normal size.

4 - Division by Tapering

4.1 - Overview

This division is more difficult than the earlier ones. It might be best to eliminate this division altogether, as it doesn't contribute as much as the others. Its only use is to distinguish the duplex parallel-lined glyphs in the range 2501-2800 from duplex tapered glyphs.

In "Calligraphy for Computers," [Hershey 1967, p. 3] Dr. Hershey identifies three new names for alphabets "because they are not identical with existing alphabets":

The word simplex has been selected to describe those alphabets which are composed of lines of uniform thickness and have no serifs or flourishes. The simplex style of character is known otherwise as gothic^*, sans serif, grotesk, light face, or block letter. The word complex may be applied to those alphabets which are composed of lines of variable thickness and do have serifs or flourishes. The complex style of character includes those which are known otherwise as standard, modern, boldface, or black letter...

^*Only in America is the term gothic applied to this style of character. [original footnote by Dr. Hershey]

In related but slightly different terminology, [Wolcott & Hilsenrath 1976, p. 4] define "simplex," "complex," and "gothic" on the basis of the quantity of tapering done in the glyphs. Note here that "gothic" means a blackletter face, where Dr. Hershey employed it in its idiomatic American sense to mean sans serif. Wolcott & Hilsenrath partially identify the ranges of glyphs corresponding to these types in their index (C-1).

I've changed these terms completely, for three reasons.

First, "gothic" has two meanings, as noted above. The "Fraktur" and "Blackletter" typefaces (of which there are examples in the Hershey glyphs in the 3301-3800 range) are often termed "Gothic." Wolcott & Hilsenrath use the term in this sense. Yet in a completely unrelated usage, certain sans serif typefaces are (at least in traditional US usage) termed "gothic" (Dr. Hershey acknowledges this sense). I feel that the word had too many disparate meanings to be used for this particularly difficult division of these glyphs.

Second, "Simplex" and "Complex" as stylistic identifiers are easy to conflate with the stroke counts "Uniplex," "Duplex," and "Triplex" ("Simplex" with "Uniplex," especially).

Third, it seems to me that terms such as "Simplex" and "Complex" best describe characteristics of typefaces (which is how Wolcott and Hilsenrath use these terms), not styles of typefaces (which is how Dr. Hershey uses them). Yet Wolcott & Hilsenrath use the term "Gothic," which indicates a style of typeface, to describe characteristics of a typeface.

I'll replace these terms with three newly coined ones: Taper0, Taper1, and Taper2. These aren't very good terms - in fact they're ugly - but they have the advantage of sticking out enough that it's unlikely they'll be conflated with other terms. They're also ugly enough that with any luck they won't be used at all.

4.2 - Taper0

"Taper0" (Wolcott & Hilsenrath's "simplex") glyphs don't taper at all. All Uniplex glyphs are Taper0. Some Duplex glyphs are as well. It wouldn't be impossible to have a Triplex Taper0 glyph, but there are none in the Hershey repertory.

4.3 - Taper1

Most Duplex glyphs are "Taper1" (Wolcott & Hilsenrath's "complex") except some parallel-line glyphs in the range 2501-2800 which are Taper0. Triplex glyphs in the range 3000-3300) are "Taper1."

4.4 - Taper2

Triplex glyphs in the range 3301-4000 are "Taper2" (Wolcott & Hilsenrath's "Gothic").

5 - Division by Typeface

5.1 - Overview

The final division of the glyphs based on their visual forms (as opposed to further divisions based on the characters, numerals, and symbols they represent) is that of "typeface." Once again, there is considerable room for discussion and difference.

5.2 - Lineface

The Uniplex faces are about as simple as writing can be: just single lines. They have no variation in width between letter components. Generally, they are "sans serif" in the sense that there are no fiddly bits sticking out to further define the glyph (though it's easy enough to imagine a serif'd lineface typeface).

In "Calligraphy for Computers," [Hershey 1967, p. 3] Dr. Hershey calls these "simplex," but his definition of the term could imply more than "uniplex simplex." [Wolcott & Hilsenrath] do not further identify the cartographic range typeface, but call the normal range uniplex typeface "simplex." [Holzmann 1986] terms these faces "plain" in the Cartographic size, while at the normal size they're called "Roman" and "Simplex." The GNU plotutils [gp241] differ further (though theirs is the most extensive arrangement into modern fonts).

It doesn't seem right to call these faces "Roman" for two reasons. First, they're not really derived from the Roman letterforms. These letters originate in scratching the sand with a stick and have their modern echoes in the ballpoint pen. They share in the great utility of these two forms of writing, but have little to do with carved Roman capitals.

I'll sidestep the issue by introducing an invented term, "Lineface," to designate this type of single-stroke typeface. It's almost as ugly a term as "TaperX," but it shares with it the distinctiveness of novelty.

Two variants on "Lineface" need to be distinguished in order to maintain uniqueness of identifications: Lineface Italic (which is really just more of a slanted than an Italic face) and Lineface Script. (A certain logic would then insist that regular Lineface is Lineface Roman, but that obviates the entire purpose of this name; let's not go there.)

5.3 - Roman

As used here, "Roman" designates a style of typeface, not a character set. (The character set employed by the Romans and much of the subsequent Western and Nonwestern world will be called "Latin.") In particular, some of the Greek and all of the Cyrillic glyphs are in a Roman typeface.

5.4 - Italic

"Italic" has, more or less, it's normal meaning. Certain of the typefaces I identify as Italic really have more of the feel of a mere slanting of a Roman face, especially in the uppercase.

"Italic" is entirely different from the "Italian" typeface identified by [Wolcott & Hilsenrath] (and termed here "Lombardic," see below).

5.5 - Script

"Script" also has its conventional meaning.

In "Calligraphy for Computers" [Hershey 1967, p. 15], Dr. Hershey indicates that he adapted his Script glyphs from "a Headliner Typemaster of the Varityper Corporation."

5.6 - Fraktur

The typeface identified by [Wolcott & Hilsenrath] as "German" is a form of Gothic or Blackletter typeface. The GNU plotutils [gp241] call it "GothicGerman."

In "Calligraphy for Computers" [Hershey 1967, p. 15], Dr. Hershey calls it "German Gothic" and indicates that it was adapted from a Fraktur in Jan Tschichold's Treasury of Alphabets and Lettering [Tschichold 1966]. An examination of this work suggests that of the several versions of Fraktur illustrated the Fraktur intended is the "Walbaum Fraktur" on page 179. Tschichold identifies this as a "late" Fraktur (dating to the period 1803-1828) by J. G. Justus Erich Walbaum (1768-1839). His description of it is charming.

I'll call it "Fraktur."

5.7 - Blackletter

In "Calligraphy for Computers," [Hershey 1967, p. 15], Dr. Hershey identifies the typeface for the glyphs which now appear in the 3301 range as "English Gothic." He say that the source of this typeface was "a Le Roy lettering set for Old English." The GNU plotutils [gp241] reverse this and term it "GothicEnglish."

As indicated elsewhere "gothic" is an overloaded term that I wish to avoid. To an American printer, "gothic" might mean sans serif. To Wolcott & Hilsenrath, "gothic" represents not a typeface but the quantity of tapering I'm calling "Taper2" here.

I probably should simply call it "Old English," then. But to me "Old English" is a language, not a typeface (a language not written in "Blackletter" for that matter, and which depends heavily upon characters not present in this typeface).

Instead, I'll simply term it "Blackletter," as it is an example of a blackletter typeface.

5.8 - Lombardic

In "Calligraphy for Computers," [Hershey 1967, p. 15], Dr. Hershey identifies the typeface for the glyphs which now appear in the 3801 range as "Italian Gothic" "because of its Lombardic origins." He identifies it as "an adaptation of a font of the American Type Founders Company," and cites page 785 of the famous 1923 Specimin Book and Catalog of the that company [ATF 1923]. The GNU Plotutils editor identifies this as "Missal Initials," and an examination of [ATF 1923] confirms that it is so.

[Wolcott & Hilsenrath 1976] term this face "Italian." While this is not incorrect, it is easily confusable (in name, not in appearance!) with "Italic." The GNU plotutils call it "GothicItalian."

It is similar to the face identified as "Lombardic Gothic Versal" in Frederic W. Goudy's classic The Alphabet and Elements of Lettering [Goudy 1952]. I'll call it Lombardic, after its origins.

6 - Application by Character Set

6.1 - Character Sets vs. Glyphs

A character set is the underlying communications entity separate from the form of the glyphs used to represent it.

6.2 - Latin Letters

The ordinary Western alphabet is generally termed "Latin" (e.g., the various ISO® "Latin" character set standards).

6.3 - Arabic Digits

The decimal digits: 0, 1, 2, 3, 4, 5, 6, 7, 8 ,9

6.4 - Greek Letters

The Hershey glyphs include the standard Greek characters (but not archaic characters such as digamma). No accents are provided, but some variant characters such as the final sigma are.

6.5 - Cyrillic Letters

I know nearly nothing about Cyrillic.

6.6 - Symbols, Punctuation, Etc.

The Hershey glyphs encode a wide range of symbols, but these do not necessarily correspond to the symbols and punctuation used in ASCII or any other character set. Certain common symbols are omitted. It is perhaps a sign of the recent sharp decline of our society as we grasp more and more for less and less that only three decades ago Dr. Hershey did not find it necessary to encode the copyright and registered trademark symbols in his original repertory.