The Herschey Fonts

Notes from Frank Grießhammer’s lec­ture on the Hershey fonts at the Cooper Union, 2016.

Video of the talk

Minotaur has a lom­bardic up­per­case made en­tirely out of straight lines.

The Herschey fonts are a col­lec­tion of vec­tor fonts de­vel­oped in 1967 at the Naval Weapons Laboratory (in Virginia). They’re pub­licly avail­able.

Archive.org (and Google Scholar) has Herschey’s pa­per Calligraphy for Computers. In it, he writes about type de­sign for a coarse grid as an artis­tic chal­lenge. He’s clearly an in­ter­est­ing per­son: artis­tic, but also ex­tremely tech­ni­cal.

The pa­per has 200 pages, most of which are draw­ings of let­ters, in­clud­ing the lom­bardic caps from ear­lier.

Who is Allen V. Herschey

Worked as a physi­cist at the Naval Weapons Laboratory. He was a the­o­re­tial physi­cis (fluid dy­nam­ics). He was also ap­par­ently a cal­lig­ra­pher.

How did he de­sign his let­ters? He drew out each let­ter on graph pa­per, us­ing only straight lines. He drew a com­plete Japanese type­face, too.

The guy wrote all these tech­ni­cal re­ports. Mathematical type­set­ting in the 60s was done on the Veritype. You also had a prod­uct called Type-It, which were ba­si­cally re­place­ment type that you in­stalled in your type­writer to get, say, Greek let­ters. Herschey wanted to im­prove this process.

The Naval Weapons Laboratory had a com­puter called the NORC (Naval Ordinance Reference Calculator, 1954) - the fastest com­puter in the world at the time.

It was at­tached to a line printer, which would print onto these long con­tin­u­ous sheets of pa­per: A General Dynamics S-C 4020.

Hey! This page con­tains em­bed­ded con­tent from Youtube, who might use cook­ies and other tech­nolo­gies to track you. To view this con­tent, click Allow Youtube con­tent.

General Dynamics: The Mark of Man (1963)

In this video they show a sten­cil method to pro­ject let­ter­forms (similar to pho­to­type­set­ting), but Herschey used the vec­tor draw­ing method in­stead (this just sent the beam through the pe­riod and used it to draw). Here’s some mar­ket­ing copy about the S-C 4020:

Once upon a time […] the com­puter-printer gen­er­ated only row af­ter row of num­bers and sym­bols, it was only used by the en­gi­neers. The took their arm­loads of pa­pers to the draft­ing de­part­ment, where an army of drafts­men and clerks con­verted the fig­ures into charts, graphs, and draw­ings. […] Then a bright en­gi­neer learned about a new kind of ma­chine called the S-C 4020 that takes the in­for­ma­tion froma com­puter and con­verts it into curves, graphs and pic­tures.

A page from the 4020 manual, showing an exploded drawing of the shaped beam tube

The CHARACTRON shaped beam tube of the S-C 4020

http://​www.chilton-com­put­ing.org.uk/​acl/​pdfs/​sc4020_in­f_­man­ual.pdf

The S-C 4020 was a tech­ni­cal mile­stone. It could do ar­bi­trary vec­tor draw­ings, and even pro­duce an­i­ma­tion. Most of this ma­te­r­ial seems to come from Chilton Computing.

Figure 9 from Proceedings of the 9th Annual Meeting showing a cartoon character and a cloud.

Vector draw­ings pro­duced by the S-C 4020.

http://​www.chilton-com­put­ing.org.uk/​acl/​tech­nol­ogy/​sc4020/​p011.htm

Herschey knew the ma­chine:

Did quite beau­ti­ful draw­ings of ships, too.

The type

Herschey’s de­signs are quite ex­haus­tive. he drew five op­ti­cal sizes: FORTRAN, Cartographic, Indexical, Principal, and Triplex (Going from cap­tion to dis­play). Multiple stroke styles (one, two, and three). There are also cal­li­graphic de­signs, a Schwabacher, the Lombardic cap­i­tals.

In to­tal he drew about 1500 Latin char­ac­ters, 800 Japanese. It’s quite the achieve­ment. Actually did real type-de­sign re­search for these, too.

The FORTRAN style comes from the Leroy Lettering Set (used for comics a lot).

While these type­faces are avail­able, they only come in an an­cient plain-text for­mat. Grießhammer wrote the nec­es­sary Python to turn this into prop­erly-en­coded OTFs (albeit with a stem width of 0) - looks like that’s this repo, though it does­n’t con­tain the con­verted font files. Here’s some­one else’s at­tempt from 1997.

This is spec­u­la­tion from me, Max: I’m as­sum­ing this ma­te­r­ial flowed into the MAD Family from Colophon (especially the three weights pro­duced by re­peat­ing the out­line), though they don’t seem to ac­knowl­edge it in their mar­ket­ing copy.