There are a variety of Unicode graphics drawing character sets, similar to the Large Character Microvector Set ROM from the HP 2640 series, which let you draw each character from a “3×3 matrix of smaller characters”.
I was thinking it might be fun to abuse some of these character sets for asciinema plotting and the like.
These are used to build up large math symbols over several character cells. I don’t know where they come from but I suspect they’re in Unicode for round-tripping compatibility.
(loop for i from #x239b to #x23b3 do (insert i))
⎛⎜⎝⎞⎟⎠⎡⎢⎣⎤⎥⎦⎧⎨⎩⎪⎫⎬⎭⎮⎯
You can use them like this:
⎛ ⎞ ⎡ ⎤ ⎧ ⎫ ⎰ ⎱
⎜⎛ ⎞⎟ ⎢ ⎥ ⎪ ⎪ ⎧ ⎱ ⎰
⎜⎜()⎟⎟ ⎣ ⎦ ⎨ ⎬ ⎮ x² dx
⎜⎝ ⎠⎟ ⎪ ⎪ ⎭ ⎲
⎝ ⎠ ⎩ ⎭ ⎳
This is the set from U+2500 to U+257F from Videotex mosaic characters:
(loop for i from #x2500 to #x257f do (insert i))
─━│┃┄┅┆┇┈┉┊┋┌┍┎┏┐┑┒┓└┕┖┗┘┙┚┛├┝┞┟
┠┡┢┣┤┥┦┧┨┩┪┫┬┭┮┯┰┱┲┳┴┵┶┷┸┹┺┻┼┽┾┿
╀╁╂╃╄╅╆╇╈╉╊╋╌╍╎╏═║╒╓╔╕╖╗╘╙╚╛╜╝╞╟
╠╡╢╣╤╥╦╧╨╩╪╫╬╭╮╯╰╱╲╳╴╵╶╷╸╹╺╻╼╽╾╿
This contains light, heavy, and double line weights in horizontal and vertical orientations, though not all combinations of these, plus double, triple, and quadruple-dashed variants of light and heavy straight lines, plus light diagonals, plus rounded corners. With these you can do arbitrary combinations of vertical and horizontal lines in half-character-cell increments. I think all the combinations of light and heavy lines are provided; that would be 3⁴ = 81 characters, which are the ones from U+2500 to U+254B inclusive except U+2504 to U+250B inclusive, plus the ones from U+2574 to U+257F: (+ (- (- #x254C #x2500) (- #x250c #x2504)) (- #x2580 #x2574)) gives 80, so either I miscounted or one is missing. All three combinations of diagonal lines are provided, but most of the light–double combinations and all of the heavy–double combinations are missing.
So with these Videotex characters you can do things like this:
╭──────┒ ╲╳╱ ╱╲ ╔══════╗
│ ┃ ╱╲ ╟──────╢
┕━━━━━━┛ ╱╲ ║ ║
╚══════╝
There’s a 2018 “Graphics for Legacy Computing” proposal to add 64 sextant characters to Unicode starting at U+1FB00 compatible with the TRS-80 “pseudopixel” or “semigraphics” set, or teletext systems including Minitel. But since 1991 Unicode has contained “quadrant” characters, like the Sinclair ZX-80 and ZX-81 or the Commodore line, with four pseudopixels per character cell, from U+2596 to U+259F:
(loop for i from #x2596 to #x259f do (insert i))
▖▗▘▙▚▛▜▝▞▟
This is visibly missing the all-empty and all-full configurations (■, which follows them, is not the all-full configuration), but normally a space can be used with or without inverse video.
For monochrome or 2-color graphics, these characters plus inverse video permit doubling the character grid resolution, with full color freedom.
This is the set from U+2580 to U+2590, in “Blocks” (now “Block Elements”):
(loop for i from #x2580 to #x2590 do (insert i))
▀▁▂▃▄▅▆▇█▉▊▋▌▍▎▏▐
These characters allow you to divide character cells vertically or horizontally (but not both) into two colors with a resolution of ⅛ cell. They’re commonly used, for example, for plotting sparklines. They are clearly designed for use with inverse video (^[[7m in ANSI).
This is particularly useful for bar plots, as provided by UnicodePlots.jl, where the blocks divided left to right can provide 640 pixels of horizontal precision for your bars on an 80-character screen. This includes stacked bars in which different colors divide a bar horizontally and may include inline legends. However, experimentation seems to show that my terminal here renders them incorrectly, with some unfilled space above them in the character cell:
▊▋▌▍▎▏▐
▊▋▌▍▎▏▐
▊▋▌▍▎▏▐
Similar are the “horizontal scan line” characters, of which there are only four starting at U+23BA:
⎺⎻⎼⎽
These are explained in Frank da Cruz’s proposal L2/00-159 as being for round-trip compatibility with some old terminals:
E0D6 Scan 1 DSG 06/15, H19 07/10, WG3 05/00, TVI 09/00, IBM SV300400
E0D7 Scan 3 DSG 07/00, WYA 01/01, WG3 05/00, IBM SV300200
E0D8 Scan 5 DSG 07/01, WYA 02/02, IBM SV300300, IBM SM920000
E0D9 Scan 7 DSG 07/02, WYA 01/03, WG3 05/01, IBM SV300100
E0DA Scan 9 DSG 07/03, H19 07/11, WG3 05/01, TVI 09/01, IBM SV300600
They’re intended to join up with U+23B8 “⎸” and U+23B9 “⎹” to make boxes in a similar way to the Videotex box-drawing characters above, and there are supposed to be five of them, but this does not work in my current font:
⎹⎽⎸⎹⎺⎸⎹⎻⎸⎹⎼⎸
These can be used for sparklines in a similar way to the vertically-divided eighth blocks, but with half the resolution. Sometimes U+2500 is considered a part of the set, but at least in the font I’m using at the moment, it doesn’t fit:
⎺⎻─⎼⎽
The proposed “legacy computing” characters would augment these with, among other things, 8-position horizontal and vertical lines.
As sort of noted above, the characters ⎹ U+23B9 and ⎺ U+23BA link
up:
⎹⎺⎹⎺⎹⎺⎺⎺⎺⎺⎺⎺
⎹⎺⎹⎺⎹⎺⎺⎺⎺⎺⎺⎺⎺
In my current font, successive rows of ⎹ don’t quite link up the way
they’re supposed to, but successive columns of ⎺ do. From the
“eighth blocks” area we have another couple of characters (or actually
various pairs of characters) that can potentially be applied in the
same way:
▁▏▁▁▏▁▁▏▁▁▏
▁▏▁▏▁▏▁▏▁▏
These also fail to link up from one line to the next in my current font.
The aforementioned “Graphics for Legacy Computing” proposal includes more such characters, including four corners intended to link up with the above “eighth blocks”:
(loop for i from #x1fb7c to #x1fb7f do (insert i))
🭼🭽🭾🭿
In theory, with three printable characters like this plus a space, you could lay out a grid of thin lines with the resolution of the character grid, with the lines beginning and ending at, say, the upper left-hand corner of each character cell. This would be potentially more parsimonious than the box-drawing characters we did get, which can end and join at the center of each character cell, but be interrupted at half-character-cell intervals, a relatively useless ability. But for this ability we need 15 graphics characters (for a single line width) rather than 3.
The three “shade” characters from U+2591 to U+2593 can be used to dither between a foreground color and a background color; really you only need two of them if you have inverse video or full liberty in color choice:
░▒▓
This doesn’t increase the resolution of your display any, though.
The Braille block from U+2800 to U+28FF offers a full selection of 256 binary patterns of 8 pixels:
(loop for i from #x2800 to #x28ff do (insert i))
⠀⠁⠂⠃⠄⠅⠆⠇⠈⠉⠊⠋⠌⠍⠎⠏⠐⠑⠒⠓⠔⠕⠖⠗⠘⠙⠚⠛⠜⠝⠞⠟
⠠⠡⠢⠣⠤⠥⠦⠧⠨⠩⠪⠫⠬⠭⠮⠯⠰⠱⠲⠳⠴⠵⠶⠷⠸⠹⠺⠻⠼⠽⠾⠿
⡀⡁⡂⡃⡄⡅⡆⡇⡈⡉⡊⡋⡌⡍⡎⡏⡐⡑⡒⡓⡔⡕⡖⡗⡘⡙⡚⡛⡜⡝⡞⡟
⡠⡡⡢⡣⡤⡥⡦⡧⡨⡩⡪⡫⡬⡭⡮⡯⡰⡱⡲⡳⡴⡵⡶⡷⡸⡹⡺⡻⡼⡽⡾⡿
⢀⢁⢂⢃⢄⢅⢆⢇⢈⢉⢊⢋⢌⢍⢎⢏⢐⢑⢒⢓⢔⢕⢖⢗⢘⢙⢚⢛⢜⢝⢞⢟
⢠⢡⢢⢣⢤⢥⢦⢧⢨⢩⢪⢫⢬⢭⢮⢯⢰⢱⢲⢳⢴⢵⢶⢷⢸⢹⢺⢻⢼⢽⢾⢿
⣀⣁⣂⣃⣄⣅⣆⣇⣈⣉⣊⣋⣌⣍⣎⣏⣐⣑⣒⣓⣔⣕⣖⣗⣘⣙⣚⣛⣜⣝⣞⣟
⣠⣡⣢⣣⣤⣥⣦⣧⣨⣩⣪⣫⣬⣭⣮⣯⣰⣱⣲⣳⣴⣵⣶⣷⣸⣹⣺⣻⣼⣽⣾⣿
At the expense of a little dottiness, background bleedthrough, and spacing jitter, this can be used to get 8× character cell resolution for things like plotting points and lines on a character display; 15360 pixels in a standard 80×24 terminal window. This is better resolution than even the proposed sextant characters, and the pseudopixels are usually squarer. The bit positions within the character cell, with x increasing right and y increasing down, are (0, 0), (0, 1), (0, 2), (1, 0), (1, 1), (1, 2), (0, 3), and (1, 3), in that order.
For example, you can plot this circle:
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⣀⡤⠤⠴⠒⠒⠒⠒⠒⠒⠢⠤⠤⣀⣀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣀⡤⠖⠋⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠉⠓⠦⣄⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⠴⠋⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⠲⣄⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣠⠞⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠑⢄⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⠞⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠑⢄⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⡰⠃⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⢳⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⡼⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠱⡄⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⡼⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠹⡄⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢰⠃⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢣⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⡏⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⡆⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢰⠃⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢣⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢸⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢸⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢸⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢸⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠸⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣸⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⡇⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⡇⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢸⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢸⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢧⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢠⠇⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⢧⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢠⠏⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⢳⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣠⠏⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢄⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⡴⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠳⣄⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⡴⠋⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠑⢦⣀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣠⠔⠋⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠓⠦⣄⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣀⡤⠖⠋⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠙⠒⠢⠤⢤⣀⣀⣀⣀⣀⣀⣠⠤⠤⠖⠚⠉⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
(Some terminals display this suboptimally with the non-active Braille dots also drawn, as empty circles.)
I did that with this simple Python program:
from __future__ import division, print_function
import sys
bitpos = [(0, 0), (0, 1), (0, 2), (1, 0), (1, 1), (1, 2), (0, 3), (1, 3)]
try:
unichr
except NameError:
unichr = chr
class Canvas:
def __init__(self, text_cols, text_rows):
assert text_rows > 0
self.pixels = [[0 for x in range(text_cols * 2)]
for y in range(text_rows * 4)]
def pixelsize(self):
return len(self.pixels[0]), len(self.pixels)
def pset(self, x, y):
try:
self.pixels[y][x] = 1
except IndexError:
pass # clip silently
def render(self):
p = self.pixels
ww, hh = self.pixelsize()
return '\n'.join(''.join(
unichr(0x2800 + sum(1 << i if p[y+by][x+bx] else 0
for i, (bx, by) in enumerate(bitpos)))
for x in range(0, ww, 2)
) for y in range(0, hh, 4))
def write(self, fileobj):
fileobj.write(self.render() + '\n')
def draw_circle(canvas):
ww, hh = canvas.pixelsize()
mindim = min(ww, hh)
cx, cy = ww/2, hh/2
r = mindim/2 - 1
x, y = r, 0
for i in range(1000):
canvas.pset(int(round(x + cx)), int(round(y + cy)))
x -= y * .01
y += x * .01
if __name__ == '__main__':
import cgitb; cgitb.enable(format='text')
canvas = Canvas(80, 24)
draw_circle(canvas)
canvas.write(sys.stdout)
These probably are not useful as mosaic characters like the ones in the “Graphics for Legacy Computing” item above; faced with the choice between making them mate properly for mosaicing and giving them 45° angles, font designers have typically chosen the latter:
◢◣
◥◤