Be sure to see the other challenge, Reverse ASCII character map!

The ASCII charset (American Standard Code for Information Interchange) is the most widely-used character encoding standard. ASCII codes represent text in computers, telecommunications equipment, and other devices.


Your challenge is to print a mapping of the ASCII character set as the user inputs them. GIF:


After the user enters every ASCII character, the output should look like this:



Each character has an assigned position on a 16x6 logical grid, starting with the space character in the top left position, and wrapping such that the digit 0 appears below it.

When printable ASCII input is received, print that ASCII character at its assigned screen location without deleting any of the characters currently onscreen.


  • Your program only needs to map out the printable ASCII characters, 0x20 to 0x7E.
  • Your program must not terminate and continue to map characters to the screen until all printable ASCII characters have been inputted. From here, your program can either terminate or run off into Neverland.
  • Your program can map characters any way you like, e.g. to a spreadsheet, table, console window, or a graphical window.
  • No matter how you display the mapping, it must be updated in realtime (as soon as it receives user input).
  • If your program does not read input silently, it must put the cursor out of the way, so the text won't get in the way of the map.


Here is the pseudocode algorithm I used to generate the GIF:

loop forever
    c = input
    y_coord = c / 16
    x_coord = c - y * 16
    if c is printable
        print c at (x_coord * 2 + 1, y_coord + 1)
    end if
end loop

There may be another way to achieve the required output. You can choose to use my algorithm or your own, but the output must be the same regardless.

Here's a useful ASCII table reference.


The answer with the least bytes in each language wins. Have fun!

  • \$\begingroup\$ Do we need to have spaces between the characters? \$\endgroup\$ Commented Jun 3, 2017 at 0:28
  • \$\begingroup\$ @musicman523 Yep, those are necessary. The output must look exactly as shown in the screenshots. \$\endgroup\$
    – MD XF
    Commented Jun 3, 2017 at 0:29
  • \$\begingroup\$ Are we allowed to assume the terminal's background color? \$\endgroup\$ Commented Jun 3, 2017 at 0:38
  • \$\begingroup\$ Is clearing the terminal, then redrawing the updated table for each char acceptable? \$\endgroup\$ Commented Jun 3, 2017 at 0:40
  • \$\begingroup\$ @DigitalTrauma - Redrawing each time is acceptable - I asked in the Sandbox post \$\endgroup\$ Commented Jun 3, 2017 at 0:41

19 Answers 19


JavaScript (ES6) + HTML, 114 + 16 = 130 bytes

Saved 16 bytes thanks to @Shaggy

a=Array(96).fill` `;onkeypress=k=>(a[k.key.charCodeAt()-32]=k.key,O.innerText=a.join` `.match(/.{1,32}/g).join`
<pre id=O></pre>

It's so unbelievably satisfying to just mash the keyboard...

  • 10
    \$\begingroup\$ "so unbelievably satisfying to just mash the keyboard" may or may not have been what I was going for. +1 \$\endgroup\$
    – MD XF
    Commented Jun 3, 2017 at 3:01
  • \$\begingroup\$ And yes, you can assume only printable ASCII is given. I'm pretty sure that's rule #1. \$\endgroup\$
    – MD XF
    Commented Jun 3, 2017 at 3:02
  • \$\begingroup\$ Can't you just use prompt() within a loop? It will save you from all the event handling and HTML. OP seems to allow it. See the Mathematica post's comments. \$\endgroup\$
    – Arjun
    Commented Jun 3, 2017 at 7:04
  • \$\begingroup\$ Only handles printable ASCII; subtract 7 bytes if we can assume only printable ASCII is given That does not seem to make sense. If it only handles printable ASCII, then how can assuming printable ASCII save any bytes? \$\endgroup\$
    – Arjun
    Commented Jun 3, 2017 at 7:07
  • \$\begingroup\$ You should be able to just use onkeypress by itself, allowing you to drop the body tag. Also, the pre tag can be shortened to just <pre id=O. Although, you'll need to include the closing > in order for it to work in a Snippet. \$\endgroup\$
    – Shaggy
    Commented Jun 3, 2017 at 9:43

QBasic 4.5, 81 85 bytes

Added 4 bytes to comply with the spacing-rule.

LOCATE i\16-1,(i MOD 16+1)*2

And the output will look like this (NOTE: Old screenshot, now every character is separated by a space):enter image description here

QBasic has the LOCATE command, which comes in handy here. A breakdown of this code:

DO                          Starts an infinite loop
LOCATE 7,1                  Moves the cursor out of the way
LINE INPUT A$:i=ASC(A$)     LINE INPUT gets user input; we need LINE INPUT instead of regular input
                            for support of <space> and <comma>. The ASC() function then takes the
                            ASCII value of the first character in the input, so it can deal with
                            inputs like 'Hello' - it will take ASC('H') and assign that to 'i'
LOCATE i\16-1               Here's the cool bit: LOCATE takes a row and a column to put the cursor on.
    ,(i MOD 16+1)*2         Row is determined by dividing the ASC value by 16, minus 1 (SPACE, ASC 32 
                            is placed on row 1), and for columns we take the modulo plus 1 (Again, SPACE 
                            mod 16 = 0, plus 1 = column 1). Multiplied by 2 gives us the spacing. 
                            We move the cursor to 1,2
?CHR$(i)                    PRINT a cast of the ASCII value to CHR at the specified location.
LOOP                        Ad infinitum
  • \$\begingroup\$ QBasic beats everyone! Wow! \$\endgroup\$
    – Arjun
    Commented Jun 3, 2017 at 9:12
  • 5
    \$\begingroup\$ @Arjun You kids and your Java... \$\endgroup\$
    – steenbergh
    Commented Jun 3, 2017 at 10:50

Java 8, 143 bytes

o->{for(;;){char c=System.console().readPassword()[0];if(c>31&c<127)System.out.println(String.format("\u001B[%d;%df%c",c/16+1,(c%16+1)*2,c));}}

Uses the ANSI control code CSI n ; m f to set the cursor position and Console.readPassword() to read the user input silently. Output of some characters:


  • 1
    \$\begingroup\$ The first time I see that Java has a chance in code golf, too ! Nice ! \$\endgroup\$
    – Luatic
    Commented Jun 3, 2017 at 17:27
  • 1
    \$\begingroup\$ +1, never seen readPassword() used like that. Oh, and you seem to be missing a semi-colon after the println. Also, isn't is possible to use System.out.printf somehow instead of System.out.println(String.format(? And you can change ()-> to o-> by using an unused empty parameter. \$\endgroup\$ Commented Jun 27, 2017 at 12:03
  • \$\begingroup\$ @KevinCruijssen Fixed, thanks! \$\endgroup\$ Commented Jul 3, 2017 at 12:24

BrainFuck, 355 Bytes


BrainFuck's options are pretty limited, so output is in the terminal and the screen is "cleared" with 20 newlines. Input should be the ASCII characters, separated by newlines.

Try it online!

Formatted and Documented

These are the debug notes I used to write the program. I used my interpreter which can optionally print the state of the tape at every '~' character for debugging.


    Calculate 16 * 6
    Resulting tape state:
    [0 0 0 0 0 0 16 96 0 0 0 0 ...]
    Note that, to obtain a 16-by-6 grid, the 16
    immediately to the right is decreased to 15
    (since we will decrease it by 1 each loop
    until we reach 0 and immediately reset)

    Our next goal is to make 96 sets of 3 cells each in the pattern [C D 0]
    The first cell will represent an entered character--when the corresponding
    input on the keyboard is pressed, it will change to the entered key.
    The first cell is initialized to 32 (' ').

    The second cell will represent the delimiter after that character.
    Every 16 cells, this should be 10 for '\n'. Otherwise, it should be 32 for ' '.

    The third cell is a buffer cell, used for traversal of the grid. In general,
    it should be only temporarily modified and then reset to 0.

           The second cell of our 3-set should be 32, so the above line
           writes 32 to the 3rd cell from the beginning of the tape (0-indexed)
       <<<[ The second cell of our 3-set should be 10, and we must reset the line counter ] 

    [ At this point, the delimiting cell we need is two cells to the left. ]
    [ Debug Mode: In the previous loop, add a + in the string of 8 +'s to get visible spaces in the grid ($-signs) ]

[ Go back to the beginning of the tape and clear up the residual '15' ]



        Take input such that the state of the tape now looks like this:
        [0 0 0 0 0 c c-32 0 32 32 0 32 32 0 32 32 0 ...]
        Where 'c' was the entered character.
        We now set up 1's in the buffer zones of the first c-32
        3-sets and clear the character that is currently there.
        All that is left, then, is to copy c to that location.

    [ Set up the row of 1's. ]

    [ Clear the current character. ]

    [ Copy the new character. ]

    [ Clean up the 1's. ]

    [ Print the grid. ]

    [ Print a bunch of newlines ]

    [ Take a new input. ]

Mathematica, 108 bytes

a=" "~Table~16~Table~6;Dynamic@Grid@a

Try it online at https://sandbox.open.wolframcloud.com/

When you paste code and press Shift+Enter, a dialog box will pop up, you enter "a" for example for character a. The program runs forever.

Note: On Wolfram sandbox, the font is formatted differently from that in Mathematica in my computer. So the line/column spacing may looks weird.

  • \$\begingroup\$ Does this continually map out every character? I.e. do you have to run it more than once to see the desired output? \$\endgroup\$
    – MD XF
    Commented Jun 3, 2017 at 4:29
  • \$\begingroup\$ You run it once, and each time you press OK of an input box, another input box appear for you to enter input. \$\endgroup\$
    Commented Jun 3, 2017 at 4:30
  • \$\begingroup\$ Sounds valid then, thanks. Nice job! \$\endgroup\$
    – MD XF
    Commented Jun 3, 2017 at 4:56
  • \$\begingroup\$ I think i = ToString@Input[]] would be more convinient.Users should just type a and not "a" \$\endgroup\$
    – ZaMoC
    Commented Jun 3, 2017 at 6:09
  • \$\begingroup\$ or better i=InputString[] \$\endgroup\$
    – ZaMoC
    Commented Jun 3, 2017 at 6:19

Python 2, 115 bytes

s='\n'.join([' '*31]*6)
while 1:
 if 31<i<128:i-=32;i=i%16*2+i//16*32;s=s[:i]+c+s[i+1:];print s

Try it online!

Requires quotation marks (single or double) around the inputted characters (the TIO version does not).

  • 1
    \$\begingroup\$ You can change raw_input to input as it is community census that you can assume the input has quotes around it if needed. \$\endgroup\$ Commented Jun 3, 2017 at 1:09
  • 1
    \$\begingroup\$ Sounds good! When I was testing I was just entering keys, and it was unhappy that I was entering { without a matching }. \$\endgroup\$ Commented Jun 3, 2017 at 1:14

str, noncompeting, 18 bytes

Presenting my new semi-esoteric language.


Animated GIF

..;                   preamble
#C                    clear screen
   ...............    main program; each character is pushed to the stack before
   d                  duplicate
    b                 buffer the character
     y                convert to character code
      16#/            divmod by 16 (a / b, a % 6)
          ~2-~        subtract a / b by 2
              u       unbuffer the character
               #p     place that character in the given position
                 q    "quiet"; disable auto-printing
  • \$\begingroup\$ I don't see spaces between the characters... \$\endgroup\$
    – MD XF
    Commented Jun 7, 2017 at 16:47
  • 2
    \$\begingroup\$ @MDXF The spec says nothing about spaces in between characters. Not to mention there are plenty of answers that don't use spaces. \$\endgroup\$ Commented Jun 7, 2017 at 17:09

QBIC, 53 57 bytes

Added 4 bytes for the spacing.

{locate 7,1┘_?┘i=asc(A)┘locate i/16-1,(i%16+1)*2┘?chr$(i)

QBIC began development as a shorthand for QBasic, so I thought translating my QBasic answer would demonstrate this nicely. We've saved some 40% in the byte-count for a functionally identical program - and that's even when LOCATE, ASC, and CHR have no QBIC-functions yet. Fortunately, QBIC can pass code directly to QBasic to compensate for this. A side-by-side:

QBIC              QBASIC
------------      ------------
{                 DO
locate 7,1        LOCATE 7,1
                  note that the lower-case alphabet is left unaltered in QBIC.
_?                LINE INPUT A$  (LINE INPUT used instead of INPUT to handle comma's)
i=asc(A)          i=ASC(A$)
locate i/16-1     LOCATE i/16-1
   ,(i%16+1)*2       ,(i MOD 16+1)*2
?chr$(i)          ?CHR$(i)
                  LOOP   (implicitly added to QBIC at EOF)

Haskell, 133 bytes

g l=do c<-getChar;p"2J";mapM h(c:l);g(c:l)
h c|(y,x)<-divMod(fromEnum c)16=p$show y++';':show(2*x+1)++'H':[c]

Requires a terminal that understands ANSI escape sequences.

It's shorter to keep a list of all keys pressed so far and clearing the screen before printing all of them each round than turning off the echo in the terminal session. The latter needs import System.IO and hSetEcho stdin(2<1) which costs too many bytes.


C, 101 bytes


This was the program I used to make the graphics. Output is as shown in the GIF. ;)

  • \$\begingroup\$ 85 bytes \$\endgroup\$
    – ceilingcat
    Commented Oct 5, 2020 at 4:56
  • \$\begingroup\$ Building on @ceilingcat, 83 bytes \$\endgroup\$
    – gastropner
    Commented Nov 20, 2020 at 19:31

QBasic, 62 58 bytes

LOCATE a\16-1,1+2*(a MOD 16)

Tested with QB64. Should work fine on regular QBasic, too, although you may want to modify it to do a CLS on the first run.

Similar to steenbergh's answer, but uses INPUT$(1) to read characters one at a time. This approach is shorter and also displays no prompt. It also uses RUN for the infinite loop, since we don't have to store any state between iterations except the state of the screen.

  • \$\begingroup\$ Wow, nice. Didn't know about input$(). I like the TIPS-topic too btw. \$\endgroup\$
    – steenbergh
    Commented Nov 13, 2017 at 10:21

Ruby, 79 75 71 + 13 = 84 bytes

+13 bytes for -rio/console flag.

97.times{|n|print$/[(n+31).chr]||" ",["


loop {
  $/ += STDIN.getch
  97.times {|n|
    print $/[(n+31).chr] || " ", ["

Pascal, 112 chars

Uses crt;var c:char;Begin ClrScr;repeat c:=ReadKey;GotoXY(ord(c)and$F*2+1,ord(c)shr 4-1);write(c);until 1<0;End.

As my Mathematica solution takes many bytes in div, mod and ToCharacterCode[Input[]], I try making another answer with Pascal. But without ClrScr my compiler (FPC) left some compile information on the screen. ClrScr; takes 7 bytes.

The *2 used for proper spacing takes another 2 bytes.


LOGO, 90 bytes

rt 90
keyboardon[invoke[setxy 30*modulo ? 16 -30*int ?/16 label char ?]keyboardvalue]pu

Try it on FMSLogo.

After all, my Logo solution is the shortest, compared with my Mathematica and Pascal answer.

Add 3 bytes if the turtle is required to be hidden.


6502 machine code + Apple //e ROM, 31 bytes

Hex dump:

8000- 20 58 FC 20 0C FD 48 38
8008- E9 A0 48 29 0F 0A 85 24
8010- 68 4A 4A 4A 4A 20 5B FB
8018- 68 20 ED FD 4C 03 80

Commented assembly:

 1 HTAB     =     $24        ; HORIZONTAL POSITION OF CURSOR
 4 HOME     =     $FC58      ; CLEARS SCREEN
 6          ORG   $8000
 7          JSR   HOME
10          PHA              ; PRESERVE ACC
11          SEC              ; MAKE SURE CARRY IS SET TO SUBTRACT
12          SBC   #" "       ; SUBTRACT CHAR CODE OF SPACE
13          PHA              ; SAVE ACC
14          AND   #$0F       ; GET LOWER 4 BITS TO GET CURSOR X POSITION
16          STA   HTAB
17          PLA              ; GET OLD ACC
18          LSR              ; SHIFT HIGH NIBBLE
19          LSR              ; INTO LOW NIBBLE
20          LSR              ; TO GET CURSOR Y POSITION
21          LSR
22          JSR   SETVTAB
23          PLA              ; RESTORE ACC
24 *
25          JSR   COUT
26          JMP   GETINP

GIF demo

If the cursor invalidates it, here's a 36-byte version without a cursor:

8000- 20 58 FC AD 00 C0 10 FB
8008- 8D 10 C0 48 38 E9 A0 48
8010- 29 0F 0A 85 24 68 4A 4A
8018- 4A 4A 20 5B FB 68 20 ED
8020- FD 4C 03 80

SmileBASIC 3, 82 bytes

C$=INKEY$()IF""!=C$THEN V=ASC(C$)-32LOCATE V MOD 16*2,V DIV 16*2?C$;

In the SmileBASIC character set, ¥ is located where \ normally would be; hopefully this doesn't invalidate this answer completely.


Japt, 63 43 bytes

Seeing as I apparently have a reputation for making Japt do things it shouldn't be able to do (continuous input, in this case), golfing from my my phone, and golfing with a feed of beer inside me, it's high time I combined all 3 of those things!

;$oninput=$@OqE¬ËoOx`Í*s.value` 1Ãú ò16 m¸·

Test it (just start typing in the "Notes" field)

  • 1
    \$\begingroup\$ sooo.. How do we use this? \$\endgroup\$
    – Razetime
    Commented Nov 21, 2020 at 4:21
  • \$\begingroup\$ Just start typing in the input field, @Razetime. \$\endgroup\$
    – Shaggy
    Commented Nov 21, 2020 at 10:38

Applesoft BASIC, 134 bytes


This is a golfed version of the Apple ][ keyboard test, the program that inspired the challenge.

  • \$\begingroup\$ This is actually 134 bytes, since Applesoft BASIC is tokenized. \$\endgroup\$ Commented Aug 3, 2017 at 23:51
  • \$\begingroup\$ @insert_name_here Ah, yes. Thanks. \$\endgroup\$
    – MD XF
    Commented Aug 4, 2017 at 0:30

8086 machine code, MS-DOS .COM, 27 bytes

Runnable in DOSBox. The only assumption about register starting values is AH = 0, which is true for most DOS isotopes. This, and the fact that only BIOS interrupts are used, means it can be run from a boot sector by changing the first instruction to mov ax, 3 (and the org of course) at the terrible, terrible cost of one byte.

Expects only printable ASCII characters as input. Does not terminate.


00000000 : B0 03 CD 10 31 C0 CD 16 50 2C 20 D4 10 D0 E0 92 : ....1...P, .....
00000010 : B4 02 CD 10 58 B4 0E CD 10 EB E9                : ....X......


org 0x100
cpu 8086
    mov al, 3       ; Set video mode to 80x25 text
    int 0x10        ; Has the desired side-effect of clearing the screen
    xor ax, ax      ; Read keystroke into AX
    int 0x16        ; AH = scan code, AL = ASCII
    push ax         ; Save ASCII code for later
    sub al, 32      ; Make ASCII code into range 0..94
    aam 0x10        ; AH = AL / 16 (row), AL = AL % 16 (col)
    shl al, 1       ; Double column for spacing
    xchg ax, dx     ; Move cursor to proper position
    mov ah, 2
    int 0x10
    pop ax          ; Retrieve ASCII code
    mov ah, 0x0e    ; Output at cursor position
    int 0x10
    jmp main_loop

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.