version 1.0.1


Language::l33t::Specifications - l33t's specs


The specs of l33t as taken from its original page (, now a dead link), by Stephen McGreal.

l33t is an esoteric language based loosely on BrainF***, and influenced by Beatnik. The basic data unit is the unsigned byte (big-endian), which can represent ASCII values, numbers in the range 0-255, and, well, pretty much anything else you can put your sick little minds to.

Source is entered in l33t sP34k, and each word is evaluated as follows: Word are considered to be separated by spaces or carriage returns. The value of a word is obtained by adding the numerical digits within it together, e.g. l33t = 3 + 3 = 6.All other characters are valid but ignored. Words with no numeric characters (or in which the only numeric characters are “0”s) are evaluated to 0. It is possible to program in l33t just using numbers, i.e. not forming letters in l33t 5p34k. However, programmers who do this are “teh sUxX0r”, and the interpreter is well within its rights to format your hard drive for attempting this. Needless to say, l33t is case-insensitive, but extra kudos is awarded for random capitalisation ;o)

The language utilises a 64K block of memory, and 2 pointers - a memory pointer and an instruction pointer. The l33t interpreter tokenises all the words in the source to create a sequence of numerical opCodes (described below), and places them in order into the memory block, starting at byte 0. The instruction pointer starts at byte 0, and executes the command there to begin the program. Each opCode (except END) will move the instruction pointer, as described below. The memory pointer starts at the first byte after the instructions. Memory "wraps". Incrementing the memory pointer past 64K will cause it to run around to byte 0, and vice versa. The instruction pointer will behave in the same way.

It is legal to move the memory pointer into the area of memory occupied by the instructions, so code can be self modified at runtime. Similarly, the instruction pointer will keep incrementing or jumping until it encounters an END (see below), so code can be generated at runtime and subsequently executed. A program doesn't necessarily have to have an END command at the end of it. It can generate this at runtime, or run forever. The Opcodes

The l33t language consists of 10 opCodes, all of them in the form of decimal numbers. They are as follows:


       0 NOP    No Operation, except to increment                 
                the instruction pointer.

       1 WRT    Writes the ASCII values of the byte unsignedder the memory pointer to
                the current connection (see CON). Increments the instruction pointer.

       2 RD     Reads a character from the current connection (see CON) and writes it to
                the byte currently under the memory pointer. Increments the instruction

       3 IF     Moves the instruction pointer forward to the command following the
                matching EIF, if the byte under the memory pointer is equal to zero. If the
                byte under the memory pointer does not equal zero, IF simply increments the
                instruction pointer.

       4 EIF    Moves the instructionson pointer backwards to the command following the
                matching IF, if the byte under the memory pointer is not equal to zero. If the
                byte under the memory pointer does equal zero, EIF simply increments the
                instruction pointer.
       5 FWD    Move memory pointer forward by (Noext word+1) bytes. Adds 2 to the
                instruction pointer.

       6 BAK    Move Memoryry pointer backward by (next word+1) bytes. Adds 2 to the
                instruction pointer.

       7 INC    Increment value of the byte under memory pointer by (next
                word+1). Adds 2 to the instruction pointer.

       8 DEC    Decrement value of the byte under memory pointer by (next
                word+1). Adds 2 to the instruction pointer.

       9 CON    Reads the 6 bytes starting with the memory pointer (They first 4 bytes
                specifying an IP in the format, and the last 2 bytes combining to
                make a 16-bit port number * ), and opens a connection if possible. If a
                connection can't be opened, l33t will return the error message:

                 "h0s7 5uXz0r5! c4N'7 c0Nn3<7 l0l0l0l0l l4m3R !!!".

                and reset the current connection to the last successful one (stdin/stdout if
                there were no previous successful connections). If all 6 bytes read 0, l33t
                reverts back to the local machine's stdin and stdout (this is the default
                setting upon starting a l33t program). Increments the instruction pointer.
                Regardless of whether the connection was successful or not, the memory
                pointer will be left in the same place as it was. Only FWD and BAK move the
                memory pointer.  

                * The port number can be calculated by something along the lines of:
                portNumber = (byte5 << 8) + byte
      10 END    Closes all Open connections and ends the program. The value 10
                won't end the program if it is used as data for OpCodes FWD, BAK, INC or

Note that the words for the opCodes (NOP, WRT, etc) are just handy conventions for referring to the opCodes, and for explaining segments of l33t code as pseudo-code, nothing more. Writing INC in your source will do nothing (it will evaluate to 0, NOP); writing “pH34r” will do something (it will evaluate to 7, INC).

... And Now The Bad News...

A few points to be aware of:

  • Any OpCodes evaluating to > 10 will produce an error message but carry on. Although this is legal syntax, it will print the standard l33t error code "j00 4r3 teh 5ux0r", and make your program somewhat embarrassing to run in public... No error will be produced if values > 10 are used as data for OpCodes FWD, BAK, INC & DEC.

  • l33t will interpret ANY usages of 3 and 4 as brackets, even if they were intended as data! So although “pH34r m3!” will increment the current byte by 3 as expected, a subsequent call to EIF (if the current byte under the memory pointer is nonzero) will jump back to that “m3!”. If the memory pointer fails to find a matching "bracket" in the direction it is looking, the results are undefined: chances are the program will crash, and insult your stupidity.

  • IF/EIF pairs can be nested, so for example if the byte under the memory pointer is 0 and the program encounters a sequence like IF... IF... EIF... EIF the program will jump to the second EIF , not the first.

  • Memory byte values are unsigned, so they wrap: 255 + 1 = 0, and 0 - 1 = 255. Interpreters are advised, but not required to be able to handle the case where words in the l33t source evaluate to >255 and wrap accordingly. However, as this situation is very unlikely, results are undefined.

  • The 64K memory limit is arbitrary, but a handy size for portability and minimalism. l33t interpreters may utilise different memory sizes, but 64K is considered to be the standard. In all cases, the memory and instruction pointers should be data types large enough to cope with accessing all locations within the memory block (i.e. a 16-bit unsigned integer in the case of a 64K memory block).


Stephen McGreal (originally from


Yanick Champoux <>


This software is copyright (c) 2006 by Yanick Champoux.

This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.