Symbols for Legacy Computing, IBM PC, MS-DOS, and HP 4260 Series compatibility

[PhMajerus]

I tried adding the Symbols for Legacy Computing and Symbols for Legacy Computing Supplement to Cozette.
These characters are important in terminals both for historical reasons and to give more flexibility to modern TUI apps, so much so that their original proposals were named terminals-prop and terminals-supplement. They include denser pseudopixels to complement the existing half-blocks and quadrants, extra box drawing and block elements, fill patterns, etc….
Most only have been added to Unicode very recently, some as late as September 2024, and therefore are not widely used yet, but having them in a few fonts is needed as the first step before any app can consider taking advantage of them.
They have been requested in #44, and this pull request contains all the characters needed to close that issue.

I couldn't get a build environment working, so this only provides 1065 new glyphs in the Cozette.sfd file, and 5 variants in the CozetteCrossedSeven.sfd file.
I'm also new to FontForge, so hopefully I didn't mess up anything else and the new code points work, but please check what was changed as it seems it touched some other parts of the file as well.
I couldn't get FontForge to acknowledge the new characters in their corresponding Unicode blocks in Edit > Font Info > Unicode Ranges. Symbols for Legacy Computing U+1FB00-U+1FBFF still shows 0/212 even though selecting that range highlights the new characters in the grid.
I probably missed something, but hopefully you can integrate them properly into your build system.

This pull request contains all of the Symbols for Legacy Computing (U+1FB00 - U+1FBFF), Symbols for Legacy Computing Supplement (U+1CC00 - U+1CEBF) blocks, characters missing for codepage 437 (required for all classic PC ANSI-art), and a few extras in other blocks that have been added either as part of legacy character sets but grouped into existing similar sets, or existing in other blocks and required as part of sets spread across Symbols for Legacy Computing and other blocks. It also includes chess pieces and playing cards (#153).

In more details, the characters included are as follows:

• Sextants (U+1FB00 - U+1FB3B, Unicode 13.0, from TRS-80, Teletext)
• Octants (U+1CD00 - U+1CDE5, Unicode 16.0, from Kaypro II).
• Smooth Sextants/Diagonals (U+1FB3C - U+1FB67, Unicode 13.0, from Teletext)
• Triangles (U+25E2 - U+25E5; U+1FB68 - U+1FB6F; U+1FB9A - U+1FB9B, Unicode 13.0)
• Shaded blocks (U+1FB8C - U+1FB92; U+1FB94 - U+1FB99; U+1FB9C - U+1FB9F, Unicode 13.0, from Sharp MZ, Sinclair ZX, Commodore,…)
• Box drawing diagonal lines#1 (U+1FBA0 - U+1FBAE, Unicode 13.0, from Amstrad CPC,…)
• Box drawing diagonal lines#2 (U+1FBD0 - U+1FBDF, Unicode 16.0, from Ohio Scientific, Sharp MZ,…)
• Semicircles and extra quadrants (U+1FBE0 - U+1FBEF, Unicode 16.0, required for octants)
• Extra box drawing lines (U+1FB70 - U+1FB7F, Unicode 13.0)
• Box drawing and tick lines (U+1CE0D - U+1CE19, Unicode 16.0, from Hewlett-Packard 2640 Series terminals)
• Extra block elements (U+1FB82 - U+1FB8B; U+1FBCE; U+1FBCF, Unicode 13.0)
• MouseText (U+1FB80; U+1FB81; U+1FBB0 - U+1FBBC, Unicode 13.0, from Apple II)
• Misc sprites, UI, etc... (U+1FBBD - U+1FBCD, Unicode 13.0, from TRS-80 and others)
• Separated Quadrants (U+1CC21 - U+1CC2F, Unicode 16.0, from Sharp MZ)
• Separated Sextants (U+1CE51 - U+1CE8F, Unicode 16.0, from Teletext)
• Segmented digits (U+1FBF0 - U+1FBF9, Unicode 13.0, from Atari ST, LCD/LED seven-segment displays,…)
• Outlined capital letters and digits (U+1CCD6 - U+1CCF9, Unicode 16.0, from Sharp MZ) - Also added crossed outlined seven (U+1CCF7) to CozetteCrossedSeven.sfd
• Large Type Pieces (U+1CE1A - U+1CE50, Unicode 16.0, from Hewlett-Packard 2640 Series terminals)
• Legacy systems arrows (U+1F8B0 - U+1F8BB, Unicode 13.0 and 16.0, from Atari 8-bit and others)
• Completed Control Pictures (U+2400 - U+2429, including Unicode 16.0 delete symbols from Apple II, TRS-80, and Amstrad CPC)
• Terminal graphic symbols (U+1CE00 - U+1CE0C, Unicode 16.0, from Ohio Scientific)
• 8×8-based patterns (U+1CE90 - U+1CEAF, Unicode 16.0, from Robotron Z9001, required for octants)
• More Sharp MZ symbols (stick figures, ruler, electronics, and more) (U+1CC00 - U+1CC20, Unicode 16.0, from Sharp MZ)
• Chess pieces + standing knight (U+2654 - U+265F; U+1CCBA - U+1CCD5, Unicode 16.0, from Sharp MZ)
• Smalltalk-72 symbols (U+1CEB0 - U+1CEB3, Unicode 16.0, from Xerox Alto and Macintosh Smalltalk environments)
• Sprites (U+2B24; U+1CC48 - U+1CCB9; U+2708; U+1F6E6 - U+1CDFF, Unicode 16.0, from Mattel Aquarius, Ohio Scientific, and Sharp MZ)
• MS-DOS CP437 (U+20A7; U+21A8; U+2310; U+25AC; U+25BA; U+25C4; U+25D8; U+25D9; U+263C, from IBM PC and MS-DOS)
• HP 4260 Series terminals (U+2628; U+2243; and extra U+2244, from Hewlett-Packard 2640 Series terminals)
• Playing cards (U+1F0A0 - U+1F0AE; U+1F0B1 - U+1F0BF; U+1F0C1 - U+1F0CF; U+1F0D1 - U+1F0DF, Unicode 7.0)

The Mosaics design.png contains the details of how the smooth sextants/diagonals and triangles combine with straight slopes.
For the separated quadrants and sextants, I followed the same design as your existing Braille (which are pretty much separated octants).
The extra box drawing lines follow the same design principles as the existing ones, including the 1px overflow on the right of the cell.
The two checkerboard fill patterns U+1FB95 and U+1FB96 are reinterpretations of the original 4×4 checkerboards because it's impossible to make regular cherckerboards with a 6×13 resolution, so the more organic design I used provide a tiling pattern that looks more regular.
The sprites such as the running man, the stick figures, and the pointing hand are slightly adjusted versions of their original 8×8 characters from the Apple II and TRS-80, while the MouseText elements are based on Macintosh classic versions of the same GUI elements to better take advantage of the extra resolution.

You can test the pseudopixels using curl "https://raw.githubusercontent.com/PhMajerus/Documents/refs/heads/main/CheatSheets/Blocks%20tables.txt".
I also have some Sextants and Octants-based ANSI-art you can use to see the result.
Similarly, many of the new box drawing characters can be tested using curl "https://raw.githubusercontent.com/PhMajerus/Documents/refs/heads/main/CheatSheets/Lines%20tables.txt".

For Large Type Pieces, see curl "https://raw.githubusercontent.com/PhMajerus/Documents/refs/heads/main/HowTos/HowTo%20Large%20Type%20Pieces.txt" for explanations.

I included the 9 characters missing to support the original IBM PC / MS-DOS codepage 437, as that is required for legacy PC text apps. Those are also used by other codepages, and from my limited tests, it appears Cozette already contains all the other characters required for CP850 (MS-DOS Latin 1), CP1252 (Windows Latin 1), CP850 (MS-DOS Portuguese), CP863 (MS-DOS fr-Canada), and CP865 (MS-DOS Nordic).
So adding the missing CP437 characters also make Cozette support most MS-DOS codepages (except CP720 Arabic and CP932 Japanese).

I added 2 characters missing to support the standard HP 4260 charset ROMs U+2628 and U+2243 (and U+2244 for consistency, although it is not available in the original HP terminals). This is detailed in #146 and this PR closes that issue as well.

I adjusted ≅ (U+2245) that was a pixel too high compared to other mathematical operators alignments. That is the only existing character that I touched in the PR, everything else is just new glyphs.

[PhMajerus]

Here are the characters in FontForge:

Some characters from other blocks are required to get complete sets, for example sextants and octants do not repeat patterns already found in quadrants. Also, the proposals for legacy computing symbols added some code points to existing blocks when it made sense. So here are all the blocks impacted by this pull request.

Control Pictures

These are used a lot in old terminals to show the C0 control characters on screen, and the Legacy Computing Supplement proposal added the U+2427 - U+2429 delete symbols used on Apple II, TRS-80, and Amstrad CPC.

Geometric Shapes

It is important for those triangles to be full-block and combining with the new triangles below.

Mathematical Operators

I added ≃ (U+2243) for compatibility with HP 4260 Series terminals and ≄ (U+2244) for consistency, and adjusted ≅ (U+2245) that was a pixel too high compared to other mathematical operators alignments.

Miscellaneous Symbols

I added ☨ (U+2628) for compatibility with HP 4260 terminals.
The chess pieces are used on some legacy computers, and while these are not strictly part of the Legacy Computing Symbols, larger 2×2 variants below are, so these are added for consistency. Also, being able to play chess in a terminal is cool.

Dingbats

The Airplane (U+2708) is required as part of the 4 airplanes directions with U+1F6E7, U+1CDF8, and U+1CDF9 below.

Miscellaneous Symbols and Arrows

The Black Large Circle (U+2B24) is required as the closed version of the pac-man-style "black large circle minus left|upper|right|lower quarter" below at U+1CC6B to U+1CC6E.

Transport and Map Symbols

The Up-Pointing Airplane (U+1F6E7) is required as part of the 4 airplanes directions with U+2708 above, and U+1CDF8, and U+1CDF9 below.

Supplemental Arrows-C

These arrows are inherited from legacy systems and have been added as part of the Legacy Computing and Legacy Computing Supplement proposals.

Playing Cards

CozetteCrossedSeven.sfd variants:

Symbols for Legacy Computing

This is the complete Symbols for Legacy Computing block, including both the Unicode 13.0 additions, and the Unicode 16.0 extras added at the same time as the Legacy Computing Supplement block.

Symbols for Legacy Computing Supplement

This is the complete Symbols for Legacy Computing Supplement block.

CozetteCrossedSeven.sfd variant:

MS-DOS CP437

These are U+20A7, U+21A8, U+2310, U+25AC, U+25BA, U+25C4, U+25D8, U+25D9, and U+263C from the original IBM PC character set.

---

The 2×2 Chess pieces are designed to be placed next to each other to build larger pieces. They follow a similar design as the ones from Miscellaneous Symbols, but slightly larger:

The U+1FB98 and U+1FB99 fill patterns are supposed to be 4×4 checkerboards. But fitting 4×4 into 6×13 gives horrible uneven rectangles, as seen on the first of the examples below. 2×4 works much better and keeps the original square checkerboard look, but still makes some moiré/aliasing artifacts because ¼ of the squares are rectangles (with an extra row of pixels, very noticeable to some of us), as seen on the second example. So I decided to instead build a more organic cells/scales pattern, which hides the lack of even repetitions, as seen on the third example.

This is similar to how many vectorial fonts do not use the original pixels patterns for ░ ▒ ▓ (U+2591 - U+2593) shading blocks to avoid moiré patterns.
I'm not sure if the cells pattern is better than a 2×4 checkerboard in real-world usage of those characters, but any hope of them joining with other patterns wouldn't work anyway if we don't respect the original 4×4 which looks horrible, so we'll see if the cells pattern works and maybe go back to 2×4 checkerboard at some point if it proves unsatisfactory. The cells pattern design has the benefit of making it clear those are fill patterns, and avoid confusion with U+1CD57 and U+1CD89 octants.

---

Here is the full list of the 1065 new characters included in this pull request:

₧ ↨ ≃ ≄ ⌐ ␁ ␂ ␃ ␄ ␅ ␆ ␇ ␐ ␑ ␒ ␓ ␔ ␕ ␖ ␗ ␘ ␙ ␚ ␛ ␡ ␢ ␣ ␥ ␦ ␧ ␨ ␩ ▬ ► ◄ ◘ ◙ ◢ ◣ ◤ ◥ ☨ ☼ ♔ ♕ ♖ ♗ ♘ ♙ ♚ ♛ ♜ ♝ ♞ ♟ ✈ ⬤ 𜰀 𜰁 𜰂 𜰃 𜰄 𜰅 𜰆 𜰇 𜰈 𜰉 𜰊 𜰋 𜰌 𜰍 𜰎 𜰏 𜰐 𜰑 𜰒 𜰓 𜰔 𜰕 𜰖 𜰗 𜰘 𜰙 𜰚 𜰛 𜰜 𜰝 𜰞 𜰟 𜰠 𜰡 𜰢 𜰣 𜰤 𜰥 𜰦 𜰧 𜰨 𜰩 𜰪 𜰫 𜰬 𜰭 𜰮 𜰯 𜰰 𜰱 𜰲 𜰳 𜰴 𜰵 𜰶 𜰷 𜰸 𜰹 𜰺 𜰻 𜰼 𜰽 𜰾 𜰿 𜱀 𜱁 𜱂 𜱃 𜱄 𜱅 𜱆 𜱇 𜱈 𜱉 𜱊 𜱋 𜱌 𜱍 𜱎 𜱏 𜱐 𜱑 𜱒 𜱓 𜱔 𜱕 𜱖 𜱗 𜱘 𜱙 𜱚 𜱛 𜱜 𜱝 𜱞 𜱟 𜱠 𜱡 𜱢 𜱣 𜱤 𜱥 𜱦 𜱧 𜱨 𜱩 𜱪 𜱫 𜱬 𜱭 𜱮 𜱯 𜱰 𜱱 𜱲 𜱳 𜱴 𜱵 𜱶 𜱷 𜱸 𜱹 𜱺 𜱻 𜱼 𜱽 𜱾 𜱿 𜲀 𜲁 𜲂 𜲃 𜲄 𜲅 𜲆 𜲇 𜲈 𜲉 𜲊 𜲋 𜲌 𜲍 𜲎 𜲏 𜲐 𜲑 𜲒 𜲓 𜲔 𜲕 𜲖 𜲗 𜲘 𜲙 𜲚 𜲛 𜲜 𜲝 𜲞 𜲟 𜲠 𜲡 𜲢 𜲣 𜲤 𜲥 𜲦 𜲧 𜲨 𜲩 𜲪 𜲫 𜲬 𜲭 𜲮 𜲯 𜲰 𜲱 𜲲 𜲳 𜲴 𜲵 𜲶 𜲷 𜲸 𜲹 𜲺 𜲻 𜲼 𜲽 𜲾 𜲿 𜳀 𜳁 𜳂 𜳃 𜳄 𜳅 𜳆 𜳇 𜳈 𜳉 𜳊 𜳋 𜳌 𜳍 𜳎 𜳏 𜳐 𜳑 𜳒 𜳓 𜳔 𜳕 𜳖 𜳗 𜳘 𜳙 𜳚 𜳛 𜳜 𜳝 𜳞 𜳟 𜳠 𜳡 𜳢 𜳣 𜳤 𜳥 𜳦 𜳧 𜳨 𜳩 𜳪 𜳫 𜳬 𜳭 𜳮 𜳯 𜳰 𜳱 𜳲 𜳳 𜳴 𜳵 𜳶 𜳷 𜳸 𜳹 𜴀 𜴁 𜴂 𜴃 𜴄 𜴅 𜴆 𜴇 𜴈 𜴉 𜴊 𜴋 𜴌 𜴍 𜴎 𜴏 𜴐 𜴑 𜴒 𜴓 𜴔 𜴕 𜴖 𜴗 𜴘 𜴙 𜴚 𜴛 𜴜 𜴝 𜴞 𜴟 𜴠 𜴡 𜴢 𜴣 𜴤 𜴥 𜴦 𜴧 𜴨 𜴩 𜴪 𜴫 𜴬 𜴭 𜴮 𜴯 𜴰 𜴱 𜴲 𜴳 𜴴 𜴵 𜴶 𜴷 𜴸 𜴹 𜴺 𜴻 𜴼 𜴽 𜴾 𜴿 𜵀 𜵁 𜵂 𜵃 𜵄 𜵅 𜵆 𜵇 𜵈 𜵉 𜵊 𜵋 𜵌 𜵍 𜵎 𜵏 𜵐 𜵑 𜵒 𜵓 𜵔 𜵕 𜵖 𜵗 𜵘 𜵙 𜵚 𜵛 𜵜 𜵝 𜵞 𜵟 𜵠 𜵡 𜵢 𜵣 𜵤 𜵥 𜵦 𜵧 𜵨 𜵩 𜵪 𜵫 𜵬 𜵭 𜵮 𜵯 𜵰 𜵱 𜵲 𜵳 𜵴 𜵵 𜵶 𜵷 𜵸 𜵹 𜵺 𜵻 𜵼 𜵽 𜵾 𜵿 𜶀 𜶁 𜶂 𜶃 𜶄 𜶅 𜶆 𜶇 𜶈 𜶉 𜶊 𜶋 𜶌 𜶍 𜶎 𜶏 𜶐 𜶑 𜶒 𜶓 𜶔 𜶕 𜶖 𜶗 𜶘 𜶙 𜶚 𜶛 𜶜 𜶝 𜶞 𜶟 𜶠 𜶡 𜶢 𜶣 𜶤 𜶥 𜶦 𜶧 𜶨 𜶩 𜶪 𜶫 𜶬 𜶭 𜶮 𜶯 𜶰 𜶱 𜶲 𜶳 𜶴 𜶵 𜶶 𜶷 𜶸 𜶹 𜶺 𜶻 𜶼 𜶽 𜶾 𜶿 𜷀 𜷁 𜷂 𜷃 𜷄 𜷅 𜷆 𜷇 𜷈 𜷉 𜷊 𜷋 𜷌 𜷍 𜷎 𜷏 𜷐 𜷑 𜷒 𜷓 𜷔 𜷕 𜷖 𜷗 𜷘 𜷙 𜷚 𜷛 𜷜 𜷝 𜷞 𜷟 𜷠 𜷡 𜷢 𜷣 𜷤 𜷥 𜷦 𜷧 𜷨 𜷩 𜷪 𜷫 𜷬 𜷭 𜷮 𜷯 𜷰 𜷱 𜷲 𜷳 𜷴 𜷵 𜷶 𜷷 𜷸 𜷹 𜷺 𜷻 𜷼 𜷽 𜷾 𜷿 𜸀 𜸁 𜸂 𜸃 𜸄 𜸅 𜸆 𜸇 𜸈 𜸉 𜸊 𜸋 𜸌 𜸍 𜸎 𜸏 𜸐 𜸑 𜸒 𜸓 𜸔 𜸕 𜸖 𜸗 𜸘 𜸙 𜸚 𜸛 𜸜 𜸝 𜸞 𜸟 𜸠 𜸡 𜸢 𜸣 𜸤 𜸥 𜸦 𜸧 𜸨 𜸩 𜸪 𜸫 𜸬 𜸭 𜸮 𜸯 𜸰 𜸱 𜸲 𜸳 𜸴 𜸵 𜸶 𜸷 𜸸 𜸹 𜸺 𜸻 𜸼 𜸽 𜸾 𜸿 𜹀 𜹁 𜹂 𜹃 𜹄 𜹅 𜹆 𜹇 𜹈 𜹉 𜹊 𜹋 𜹌 𜹍 𜹎 𜹏 𜹐 𜹑 𜹒 𜹓 𜹔 𜹕 𜹖 𜹗 𜹘 𜹙 𜹚 𜹛 𜹜 𜹝 𜹞 𜹟 𜹠 𜹡 𜹢 𜹣 𜹤 𜹥 𜹦 𜹧 𜹨 𜹩 𜹪 𜹫 𜹬 𜹭 𜹮 𜹯 𜹰 𜹱 𜹲 𜹳 𜹴 𜹵 𜹶 𜹷 𜹸 𜹹 𜹺 𜹻 𜹼 𜹽 𜹾 𜹿 𜺀 𜺁 𜺂 𜺃 𜺄 𜺅 𜺆 𜺇 𜺈 𜺉 𜺊 𜺋 𜺌 𜺍 𜺎 𜺏 𜺐 𜺑 𜺒 𜺓 𜺔 𜺕 𜺖 𜺗 𜺘 𜺙 𜺚 𜺛 𜺜 𜺝 𜺞 𜺟 𜺠 𜺡 𜺢 𜺣 𜺤 𜺥 𜺦 𜺧 𜺨 𜺩 𜺪 𜺫 𜺬 𜺭 𜺮 𜺯 𜺰 𜺱 𜺲 𜺳 🂠 🂡 🂢 🂣 🂤 🂥 🂦 🂧 🂨 🂩 🂪 🂫 🂬 🂭 🂮 🂱 🂲 🂳 🂴 🂵 🂶 🂷 🂸 🂹 🂺 🂻 🂼 🂽 🂾 🂿 🃁 🃂 🃃 🃄 🃅 🃆 🃇 🃈 🃉 🃊 🃋 🃌 🃍 🃎 🃏 🃑 🃒 🃓 🃔 🃕 🃖 🃗 🃘 🃙 🃚 🃛 🃜 🃝 🃞 🃟 🛧 🢰 🢱 🢲 🢳 🢴 🢵 🢶 🢷 🢸 🢹 🢺 🢻 🬀 🬁 🬂 🬃 🬄 🬅 🬆 🬇 🬈 🬉 🬊 🬋 🬌 🬍 🬎 🬏 🬐 🬑 🬒 🬓 🬔 🬕 🬖 🬗 🬘 🬙 🬚 🬛 🬜 🬝 🬞 🬟 🬠 🬡 🬢 🬣 🬤 🬥 🬦 🬧 🬨 🬩 🬪 🬫 🬬 🬭 🬮 🬯 🬰 🬱 🬲 🬳 🬴 🬵 🬶 🬷 🬸 🬹 🬺 🬻 🬼 🬽 🬾 🬿 🭀 🭁 🭂 🭃 🭄 🭅 🭆 🭇 🭈 🭉 🭊 🭋 🭌 🭍 🭎 🭏 🭐 🭑 🭒 🭓 🭔 🭕 🭖 🭗 🭘 🭙 🭚 🭛 🭜 🭝 🭞 🭟 🭠 🭡 🭢 🭣 🭤 🭥 🭦 🭧 🭨 🭩 🭪 🭫 🭬 🭭 🭮 🭯 🭰 🭱 🭲 🭳 🭴 🭵 🭶 🭷 🭸 🭹 🭺 🭻 🭼 🭽 🭾 🭿 🮀 🮁 🮂 🮃 🮄 🮅 🮆 🮇 🮈 🮉 🮊 🮋 🮌 🮍 🮎 🮏 🮐 🮑 🮒 🮔 🮕 🮖 🮗 🮘 🮙 🮚 🮛 🮜 🮝 🮞 🮟 🮠 🮡 🮢 🮣 🮤 🮥 🮦 🮧 🮨 🮩 🮪 🮫 🮬 🮭 🮮 🮯 🮰 🮱 🮲 🮳 🮴 🮵 🮶 🮷 🮸 🮹 🮺 🮻 🮼 🮽 🮾 🮿 🯀 🯁 🯂 🯃 🯄 🯅 🯆 🯇 🯈 🯉 🯊 🯋 🯌 🯍 🯎 🯏 🯐 🯑 🯒 🯓 🯔 🯕 🯖 🯗 🯘 🯙 🯚 🯛 🯜 🯝 🯞 🯟 🯠 🯡 🯢 🯣 🯤 🯥 🯦 🯧 🯨 🯩 🯪 🯫 🯬 🯭 🯮 🯯 🯰 🯱 🯲 🯳 🯴 🯵 🯶 🯷 🯸 🯹