Regarding Mac and what was available to the public, you needed Adobe Type Manager (ATM) for outline fonts on the screen until the arrival of TrueType in the early 1990s (System 7.1, I think?). Certainly not a contender for a “first!”.
BTW, does anyone know anything about font rendering on the Star? It used Interpress for printing, but what was used for screen rendering?
I’m pretty sure it used bitmap fonts. That was definitely the case for the earlier Alto-based systems at Xerox.
How’s the LINC emulator coming along? Running Spacewar yet?
Oh, well, I’ve become distracted… :-/
It seems, the game is similar to Spacewar for the PDP-4.
If there is interest, I’ve a transcript of the source code available.
1 Like
Yes! Very much so.
(Add to make 20 characters.)
Here you are (including source context, 11MB):
https://masswerk.at/spacewar/sources/LINC-Spacewar.zip
I see, I actually wrote an assembler for verification (see the listing.txt for list output and octal object code at the very bottom of this) – I hope this does work correctly, never tested.
PS: I also found this list of “instructions used” (I guess, this is an extensive list?):
ADA 1100 CLR 0011 ROL 0240 STA 1040
ADD 2000 COM 0017 SAM 0100 STC 4000
ADM 1140 DIS 0140 SCR 0340 XSK 0200
APO 0451 JMP 6000 SET 0040
AZE 0450 LDA 1000 SNS 0440
BCL 1540 NOP 0016 SRO 1500
Thanks! I’m inspired to make a LINC emulator, maybe as an addition to SIMH. But not yet; no promises.
Here are two cool LINC videos made recently:
2 Likes
There’s however the challenge of somehow translating that haptic interface to a terminal environment: besides the keyboard and sense switches, there are also 8 knobs with a digital scan range of -177…+177, each.
You might be interested in these articles (which I hadn’t seen back when we had this discussion). Liesbeth Flobbe is one of the few people who seem to really understand the issues involved in handwriting fonts:
2 Likes
Yes, every character needs many variants in typical cursive writing. None of that is impossible with OpenType, but it’s a lot of work. If you want to see some serious character variations, the formal scripts of Arabic and Farsi given OpenType’s features a real workout. Neatroff handles them well: https://litcave.rudi.ir/neatstart.pdf
Handwritten text, especially if a slightly flexible nib is used, is made up of nuanced lines: they vary in thickness with pressure (or how much pigment is left, if a brush is used). Getting those to join neatly is complicated.
The letter “t” in that Cogncur font is an affront to legibility. Then again, I can barely read any cursive. It seems my schooling fell exactly between cursive fads. When I had to re-learn handwriting again in my mid 40s, Getty-Dubay italics are what I chose.
Me too. Mine runs your copy of Spacewar now. But I don’t know what to for the input devices.
It also passes most of these diagnostics: Directory of images/GJohns/classic-test.linc
2 Likes
In a roundabout way, I recently did a conversion (of a subset) of these fonts as TrueType. IMHO they turned out excellent.
See: dps8m / font-gct-multics · GitLab
The full-width README might display a bit better for most users: README.md · master · dps8m / font-gct-multics · GitLab
This project wasn’t to convert the Hershey fonts, but the Multics GCT (Graphics Character Tables) to TrueType, and it just so happens the 12 Multics Graphics fonts are derived from the Hershey set.
I also developed a GCT linter and the opposite conversion tool, one which converted SVG stroked font data to Multics GCT for use with the Multics Graphics System.
2 Likes
For those that don’t want to click through to the whole page (at README.md · master · dps8m / font-gct-multics · GitLab) here you can see how the “Gothic English” font turned out in TrueType (selected to show here since it’s the most intricate) of the Multics fonts:
If there is demand I could probably do a conversion of the full Hershey font data using the same techniques. Although the process is >99% automatic, it is still somewhat time consuming if you strive for perfection.
4 Likes
An interesting thing about the LINC is that the DSC (DiSplay Character) instruction draws a 2x6 dot pattern on the screen using the bits of the Y register (what LINC calls the memory location addressed by the instruction) at the position indicated by the accumulator (vertical) and register 1 (horizontal). You call this twice (each instruction takes t+112µs to run) to draw a full character. Each dot is spaced 4 coordinate values horizontally and vertically.
1 Like
An interesting aspect of these old character sets is the general absence of the exclamation mark, while there’s still room for characters of a more decorative nature (e.g., a bullet). I don’t think that anyone would seriously consider that we may do without it, nowadays. Which may hint at a change in communication styles.
I wouldn’t be shocked to find that the vertical bar third from the end on the penultimate row stood in for exclamation mark, or that the “decorative bullet” is actually an asterisk or perhaps a multiplication dot. Scientific computation was central enough in those days that providing relevant symbols (I see a rho, mu, and theta there, for sure, and what I assume is a superscript star for set notation or perhaps a degree sign?) seems to have trumped simple typography.
Then again, maybe you didn’t need an exclamation point because everything was shouting all the time anyway, without minuscules. 
Good guess, it’s not a bullet point but a hash mark #
It’s not rho, mu, or theta though. It lower case p, u, and the strange “origin character”. I render it latter as the squared minus ⊟
1 Like