Official IOCCC guidelines

These guidelines are OFFICIAL.

Unless some significant problem turns up (such as a need to delay in when the IOCCC will close, or some major bug is discovered) they will remain OFFICIAL for this IOCCC.

See our FAQ on “rules, guidelines, tools feedback” as well as our FAQ on “about asking questions” about these guidelines. You might also find the FAQ in general useful, especially the FAQ on “how to enter the IOCCC”.

SUGGESTION: Watch both the IOCCC news and the IOCCC Mastodon feed for the latest news about any mentions changes.

HINT to mastodon users: You may wish to refresh the @IOCCC mastodon feed page and/or mastodon app from time to time to view IOCCC mastodon updates.

The IOCCC open for submissions

See “FAQ - How may I enter the IOCCC” for information on how to register for the IOCCC, how to create a proper bzip2 compressed tarball of your submission, and how to upload your bzip2 compressed tarball to the IOCCC submit server.

28th International Obfuscated C Code Contest Official Guidelines

Copyright © 2024-2025 Leonid A. Broukhis and Landon Curt Noll.

All Rights Reserved. Permission for personal, education or non-profit use is granted provided this this copyright and notice are included in its entirety and remains unaltered. All other uses must receive prior permission in writing by contacting the judges.

Jump to: top

These IOCCC guidelines are version 28.44 2025-03-04.

The markdown form of these guidelines is available for download.

IMPORTANT: Be SURE to read the IOCCC rules.

Jump to: top

← Lines that start with this symbol indicate a change from the previous IOCCC.

Most lines (we sometimes make mistakes) that were modified since the previous IOCCC start with a solid 4 pixel black left border (or, in the case of a code block or blockquote, just a vertical bar).

Jump to: top

This file contains guidelines intended to help people who wish to participate in the International Obfuscated C Code Contest (IOCCC).

These are NOT the IOCCC rules, though it does contain comments about them. The IOCCC guidelines should be viewed as hints and suggestions. Submission that violate the guidelines but remain within the rules are allowed. Even so, you are safer if you remain within the IOCCC guidelines.

Of course if a guideline is talking about a rule and you break that guideline, you stand a good chance of having your submission rejected for breaking the rule, not the guideline.

You SHOULD read the CURRENT IOCCC rules, prior to submitting code to the contest.

Jump to: top

This IOCCC runs from 2024-12-29 23:58:13.213455 UTC to 2025-06-05 04:03:02.010099 UTC. This IOCCC marks the 40th anniversary of the contest!

This contest will enter the pending state on or about 2024-12-29 23:58:13.213455 UTC. When this happens, you may begin registering to enter the IOCCC.

This contest will enter the open state on 2025-03-05 23:19:17.131107 UTC. When this happens, you may begin uploading your submissions to the IOCCC.

This contest will enter the judging state on 2025-06-05 04:03:02.010099 UTC. When this happens, you may no longer upload your submissions to the IOCCC.

IMPORTANT NOTE: Until the contest enters the open state, any or all of the above dates and times may change AT ANY TIME!

The reason for the times of day are so that key IOCCC events are calculated to be a functional UTC time. :-)

Until the contest status becomes open, the IOCCC rules, IOCCC guidelines and the tools in the mkiocccentry repo SHOULD be considered provisional BETA versions and may be adjusted AT ANY TIME before the contest status becomes open.

See the FAQ on “how to register and submit to the IOCCC” for instructions on registering and participating in the IOCCC, as the process may change from contest to contest.

To assist in the formation of the xz compressed tarball for submission, use the mkiocccentry(1) tool as found in the mkiocccentry repo.

See FAQs regarding:

• obtaining the latest mkiocccentry tools
• compiling the mkiocccentry tools
• installing the mkiocccentry tools
• using the mkiocccentry tools

Uploading a tarball not formed by mkiocccentry(1) puts you at a very big risk of violating Rule 17, especially as mkiocccentry(1) does a great number of things that are required, and it also runs many checks, and if any of those checks fail, you are at a very great risk of having your submission rejected for violating Rule 17.

We STRONGLY recommend you do install the mkiocccentry toolkit.

IMPORTANT NOTE: if you run an IOCCC related tool outside the mkiocccentry toolkit directory and you have not installed those tools, then you will have to specify the options (such as paths) for the tools that are required like chkentry(1), txzchk(1) and fnamchk(1).

See the FAQ on “obtaining and compiling the most recent mkiocccentry tools” and the FAQ on “how to enter the IOCCC” as that FAQ has important details on how to register as well as how to upload your submission to the IOCCC.

While the contest is open, you may modify your previously uploaded submission by rebuilding your submission with the mkiocccentry(1) tool and then re-uploading it to the same slot number on the submit server.

HINT: So that you do not have to repeatedly answer all the questions, the mkiocccentry(1) tool has the options -a answers, -A answers and -i answers, where -a will write to an answers file (if it does not already exist), -A WILL OVERWRITE THE FILE and -i will read the answers from the file. If you use -A, BE SURE you don’t overwrite another file by accident!

Be aware that even with the -i answers you will still be required to confirm most if not all y/n questions - you just don’t have to input the name of the submission, the abstract, the author details etc. If you really wish to circumvent this you can use the -Y option but we do not recommend this because if your update breaks a rule or there is some problem, you might not see it.

The overall size limit (see Rule 2a) on prog.c has been increased from 4096 to 4993 bytes. The Rule 2a size was changed from 4096 to 4993 to keep the “2b to 2a” size ratio to a value similar to the 2001-2012 and 2013-2020 IOCCC eras.

The Rule 2b size has increased from 2053 to 2503 bytes.

Your submission must satisfy BOTH the maximum size Rule 2a AND the IOCCC size Rule 2b.

To check your code against Rule 2, use the iocccentry(1) tool. For example:

    iocccsize prog.c

The IOCCC size tool algorithm can be summarized as follows:

The size tool counts most C reserved words (keyword, secondary, and selected preprocessor keywords) as 1. The size tool counts all other bytes as 1 excluding ASCII whitespace, and excluding any ‘;’, ‘{’ or ‘}’ followed by ASCII whitespace, and excluding any ‘;’, ‘{’ or ‘}’ byte immediately before the end of file.

ASCII whitespace includes ASCII tab, ASCII space, ASCII newline, ASCII formfeed, and ASCII carriage return.

When ‘;’, ‘{’ or ‘}’ are within a C string, they may still not be counted by the IOCCC size tool. This is a feature, not a bug!

In cases where the above summary and the algorithm implemented by the IOCCC size tool iocccsize(1) conflict, the algorithm implemented by the current version of iocccsize(1) is preferred by the IOCCC judges.

In other words, make sure iocccsize does not flag any issues with your prog.c.

There are at least 2 other reasons for selecting 2503 as the 2nd limit besides the fact that 2503 is a prime. These reasons may be searched for and discovered if you are “Curios!” about 2503. :-) Moreover, 2053 was the number of the kernel disk pack of one of the judge’s BESM-6, and 2503 is a decimal anagram of 2053.

Take note that this secondary limit imposed by the IOCCC size tool obviates some of the need to #define C reserved words in an effort to get around the size limits of Rule 2.

Yes Virginia, that is a hint!

Jump to: top

You are encouraged to examine the winners of previous contests.

Keep in mind that rules change from year to year, so some winning entries might not be valid submissions this year; what was unique and novel one year might be ‘old’ the next year.

A submission is usually examined in a number of ways. We typically apply a number of tests to a submission:

• look at the original source
• C pre-process the source ignoring #include lines
• C pre-process the source ignoring #define and #include lines
• run it through a C beautifier
• examine the algorithm
• compile it (with flags to enable all warnings)
• execute it

You should consider how your submission looks in each of the above tests. You should ask yourself if your submission remains obscure after it has been ‘cleaned up’ by the C pre-processor and a C beautifier.

Your submission need not pass all of the above tests. In certain cases, a test is not important. Submissions that compete for the ‘strangest/most creative source layout’ need not do as well as others in terms of their algorithm. On the other hand, given two such submissions, we are more inclined to pick the submission that does something interesting when it’s executed.

We try to avoid limiting creativity in our rules. As such, we leave the contest open for creative rule interpretation. As in real life programming, interpreting a requirements document or a customer request is important. For this reason, we often award ‘Best abuse of the rules’ or ‘Worst abuse of the rules’ or some variation to a submission that illustrates this point in an ironic way.

The IOCCC has a rich history of remarkable winning entries created by authors who skillfully employed various techniques to develop their code. While it is NOT required, you are allowed to use tools to develop and test your submission. These tools may include, but are not limited to code generators, code analysis tools, static code analysis tools, runtime analysis tools, machine learning tools, natural language models, code copilot tools, so-called AI services, large language models (LLMS), etc. If you do make use of such tools or services, then we ENCOURAGE you to describe how you used such tools in your remarks.md file.

You may use git, or services such as GitHub to develop and maintain your submission. However, we DISLIKE submissions that require them in order to build/compile your submission.

Submissions will be judged in an environment that has no IDE. Any submission that fails to compile/build because it requires an IDE will be rejected.

We will use the make(1) tool, your Makefile, and other tools found in Single UNIX Specification environments and systems that conform to the Single UNIX Specification in the building and compiling of your submission.

We do realize that there are holes in the rules, and invite submitters to attempt to exploit them. We may award ‘Worst abuse of the rules’ or ‘Best abuse of the rules’ or some variation and then plug the hole next year.

When we do need to plug a hole in the IOCCC rules or IOCCC guidelines, we will attempt to use a very small plug, if not smaller. Or, maybe not. :-)

There may be fewer than 2^7+1 reasons why these IOCCC guidelines seem obfuscated.

IMPORTANT: Be sure that your submission works as documented in your remarks.md file. We sometimes make an effort to debug a submission that has a slight problem, particularly in or near the final round. On the other hand, we have seen some otherwise excellent submissions fall down because they didn’t work as documented.

If you submission has bugs and/or mis-features, you are MUCH BETTER off documenting such bugs and/or mis-features in your remarks.md file.

Consider an example of a prime number printing program that claims that 16 is a prime number. Noting such a bug in your remarks.md file could save your submission:

“this submission sometimes prints the 4th power of a prime by mistake”

Sometimes a strange bug or (mis-)feature can even help the submission! Of course, a correctly working submission might be better.

We tend to look down on a prime number printer that claims that 16 is a prime number. Clever people will note that 16 might be prime under certain conditions. ;-) Wise people, when submitting something clever will fully explain such cleverness in their submission’s remarks.md file.

People who are considering to just use some complex mathematical function or state machine to spell out something such as “hello, world!” really really, and we do mean REALLY, do need to be more creative.

Ultra-obfuscated programs are, in some cases, easier to deobfuscate than subtly-obfuscated programs. Consider using misleading or subtle tricks layered on top of or under an appropriate level of obfuscation. A clean looking program with misleading comments and variable names might be a VERY GOOD START to a great submission!

Jump to: top

mkiocccentry

Rule 17 (the mkiocccentry(1) rule) states that you REALLY SHOULD use the mkiocccentry(1) tool to package your submission tarball.

See the mkiocccentry repo for the mkiocccentry(1) tool and below for more details.

IMPORTANT NOTE: make CERTAIN you have the most recent version of the mkiocccentry toolkit! See the FAQ on “obtaining the mkiocccentry toolkit”.

mkiocccentry runs a number of checks, by the tool itself and by executing other tools, before packaging your xz compressed tarball, including running chkentry(1) on the submission directory.

If mkiocccentry encounters an error, the program will exit and the xz compressed tarball will not be formed. For instance, if chkentry fails to validate the submission directory, either because of a validation error in the .auth.json or .info.json JSON files or because of something else that mkiocccentry(1) creates or does, it is an error and possibly a bug that you should report as a bug at the mkiocccentry bug report page. PLEASE run the bug_report.sh script to help us out here! See the FAQ on “reporting mkiocccentry bugs”.

Once the tarball is packaged it will run txzchk(1), which will also run fnamchk(1), as part of its algorithm.

However, even if mkiocccentry or one of the tools it invokes reports an error, it does not necessarily mean it is a bug in the code. It might be an issue with your submission. Thus if you report an error as a bug it might not be something that will be fixed as there might not be anything wrong with the tools.

On the other hand, some conditions flagged by mkiocccentry(1) are warnings and it allows you to override these, if you wish. If you’re brave enough you can use the -W option to ignore all warnings but this is a big risk; the -y option will assume ‘yes’ to most questions but this is also a big risk. Using -Y will say ‘yes’ to even more prompts. Needless to say, we do NOT recommend these options.

In many places it will prompt you to verify what you input, allowing you to correct details as you go along. Be advised that there is a default answer and if you press enter it will proceed with that default, so make sure you have provided the correct answer.

If you wish to test that your submission passes the mkiocccentry(1) tests without having to type in in answers each time, you can use the -d or -s seed option to mkiocccentry for the tool to pseudo-randomly create answers for you. For example:

    mkiocccentry -d workdir topdir

IMPORTANT NOTE: if you run an IOCCC related tool outside the repo directory (specifying the absolute or relative path to the tool) and you have not installed the tools (and we STRONGLY recommend you do install them), then you will have to specify the options (such as paths) for the tools that are required like chkentry(1), txzchk(1) and fnamchk(1).

Make sure you have the latest version of of the mkiocccentry toolkit so that you do not violate Rule 17 by mistake.

Jump to: top

The synopsis of the mkiocccentry(1) tool is:

    mkiocccentry [options] workdir topdir

To help you with editing a submission, the mkiocccentry(1) tool has some options to write OR read from an answers file so you do not have to input the information about the author(s) and the submission more than once, unless of course you need to make some changes, in which case you can use the option that overwrites the file. If you do use the overwrite option, MAKE SURE you do not overwrite another file!

See the FAQ on the “mkiocccentry tool” for how to use this tool and the FAQ on the “finer details of mkiocccentry” for even more information.

The mkiocccentry(1) tool will execute a number of tools, some of which will execute one or more additional tools.

iocccsize

mkiocccentry(1) will use code from iocccsize(1) which detects a number of issues (such as Rule 2) that you may ignore, if you wish, as noted above.

In other words, you no longer need to run iocccsize manually. However, the checks described above are still made but through the mkiocccentry(1) tool itself.

Jump to: top

chkentry

mkiocccentry(1) will write two JSON files: .auth.json and .info.json. These files contain information about the author(s) and about the submission. After these files are formed, the chkentry(1) tool will be run on the submission directory; this tool not only validates the two JSON files (and runs specific checks on them after validating the JSON), but it also verifies that what is in the manifest in .info.json matches the contents of the directory, in addition to other checks. If chkentry(1) fails to validate anything at all, your submission WILL BE rejected!

IMPORTANT: although chkentry(1) validates the files in the directory, it only inspects the contents of the two JSON files; it performs other kinds of checks on the other files, and those checks are run on the two JSON files as well.

If you submit your own JSON files (other than .auth.json and .info.json) then they do NOT have to be valid JSON. However, if you do provide such invalid JSON files, PLEASE document and explain this in your remarks.md file.

If chkentry does not pass and you used mkiocccentry(1) it is very likely a bug and you should report it as a bug at the mkiocccentry issues page. PLEASE run the command:

    make bug_report

from the top level repo directory itself and attach the bug report file (it will tell you the name of the file). See also the FAQ on “reporting mkiocccentry bugs”.

Assuming that chkentry(1) successfully validates your submission directory, the tarball will be formed and then txzchk(1) will be executed on it. In this case, there should be no problems, as mkiocccentry(1) should NOT form a tarball if there are any issues.

If mkiocccentry(1) is used and chkentry(1) fails to validate your submission directory, then unless it is a system specific problem, it could be a bug in mkiocccentry(1), chkentry(1) or possibly jparse, although this is unlikely. Nonetheless, if you believe there is a bug, you may report it as explained above.

If you want to know what .auth.json is, see the FAQ on “.auth.json”. If you want to know what the .info.json file is, see the FAQ on “.info.json”. On the other hand, if you want to know a bit more details about chkentry(1), see the FAQ about “chkentry”.

chkentry uses the jparse library. See the jparse README.md in the mkiocccentry GitHub repo subdirectory jparse as well as the jparse library README.md file in the mkiocccentry GitHub repo subdirectory jparse, for more details.

The jparse parser, library and tools were co-developed by Cody Boone Ferguson and Landon Curt Noll in 2022-2025 and come from the jparse repo. However, the mkiocccentry tools use a clone of the jparse repo at a specific release. Thus the mkiocccentry will at times be behind the jparse repo!

You do NOT need to install jparse from the jparse repo! The mkiocccentry tools link in the static library from mkiocccentry’s clone.

The mkiocccentry toolkit also has a clone of both the dbg library and the dyn_array library; the dyn_array library uses the dbg library and the jparse library uses both libraries but unlike in the jparse repo, the libraries do not need to be installed separately, in order to use the tools in mkiocccentry.

In other words, mkiocccentry contains everything you need, and even if you do install the libraries from their respective repos, it/they will not be used when compiling the mkiocccentry tools. This is important to make sure that you’re using the correct versions, which is also verified by chkentry (in the JSON files created by mkiocccentry(1)). See Rule 17!

Please see the FAQ on “validating your submission directory for more details on chkentry(1).

You might also wish to see the FAQ on “.auth.json” and the FAQ on “.info.json” for much more information on these files.

Jump to: top

txzchk

txzchk(1) performs a wide number of sanity checks on the xz compressed tarball; if any issues are found (‘feathers are stuck in the tarball’ :-) ) AND if and ONLY IF you used mkiocccentry(1), then it is possibly a bug in one of the tools and you might want to report it as a bug at the mkiocccentry bug report page. PLEASE run the bug_report.sh script to help us out here! You may do this by running from the top level directory of the repo make bug_report. See the FAQ on “reporting mkiocccentry bugs”.

As part of its algorithm, txzchk(1) will run fnamchk(1) on the tarball to verify that the filename is valid. See the FAQ on “fnamchk” and the fnamchk section below for more details on this tool.

It is beyond the scope of this document to discuss the many tests that txzchk(1) performs; if you wish to know, we refer you to the source code or the man page. You might just find a fun option if you do either of these!

Of course, as txzchk does not extract the tarball, it is possible that if you manually package your submission tarball and/or find and exploit a missing check, you could still be violating Rule 17.

The maximum total number of files that may be submitted has changed to 39 files. However, of those files, 5 are mandatory (prog.c, Makefile, remarks.md and the two JSON files generated by mkiocccentry(1), .info.json and .auth.json).

Additionally, three of the files, if included, MUST be specific file(name)s AND in the top level directory, or they will be counted as an extra file, not an optional file.

In particular, any file that is not prog.c, Makefile, remarks.md, try.sh, prog.alt.c or try.alt.sh will be counted as an extra file, and if try.sh, prog.alt.c or try.alt.sh are not in the top level submission directory they will also be counted as an extra file. The .info.json and .auth.json files are not counted as extra files but are required.

In other words, the actual amount of EXTRA files is 31.

If you use an optional filename for something other than their intended use in order to get past the file limit, we will consider that an abuse of rules. For more details on the optional files, see the FAQ on the “try.sh script system” and the FAQ on “alt code”.

If you REALLY MUST include more files than the limit allows, then you may do so by including as an extra file, a tarball. This does NOT have to pass txzchk(1) tests; only the submission tarball must pass the txzchk(1) tests.

If you DO include a tarball, and the build process or the program extracts said tarball(s), the make clobber rule MUST remove the extracted files. Even so, if you include a tarball to get past the limit on the number of files, you MUST justify this in your remarks.md file.

IMPORTANT REMINDER: make SURE your tarball does NOT reveal who you are! The mkiocccentry(1) tool creates a v7 format tarball to prevent this. You can do the same like:

    tar --format=v7 -cJf foo.txz directory

See Rule 17 and in particular the part about the maximum number of files. If you do not follow these points, you are at a great risk of violating Rule 17!

NOTE: if you want to include a test-suite that requires a lot of files, please suggest this in your remarks.md and if your submission wins it can be done. In other words you should not use a tarball for a test-suite unless you have a very good reason for this (and if you do, make SURE you specify why in your remarks.md file).

See also the FAQ on “txzchk”.

Jump to: top

As an important part of its algorithm, txzchk(1) directly executes fnamchk(1). If the filename is invalid (or the filename does not match the directory name of the tarball) then it is an error and you risk violating Rule 17. Nevertheless, you can run the tool manually, should you wish to.

For more information on fnamchk and how to manually validate your submission tarball filename, see the FAQ on “fnamchk”.

It is extremely unlikely that fnamchk(1) reporting an invalid filename is a bug in fnamchk(1) and as such, ignoring such an issue risks violating Rule 17 which is a big risk. Of course, using mkiocccentry(1) would prevent this from happening as it would not create such a file anyway. If mkiocccentry(1) was used it would rather suggest a bug in one of the tools and you should report it as a bug at the mkiocccentry issues page. See the FAQ on “reporting mkiocccentry bugs”.

As you can see, the use of mkiocccentry(1) is HIGHLY RECOMMENDED, and at the risk of stating the obvious, you run A VERY BIG RISK of having your submission rejected if you package your own tarball, and there are ANY problems. For instance, if chkentry(1) found a problem in your .info.json file, the mkiocccentry(1) tool would not package it. But if you were to package it manually, you would be violating Rule 17. But even if everything checks out OK you should NOT assume that everything IS OK.

Although the tools have been tested extensively, it is nonetheless possible for bugs to exist in the code, as all programmers know. In this case, please ask for help or report what you think is a bug via the bug report issues page in the mkiocccentry repo.

Of course, it is also possible for mkiocccentry(1), or one or more of the tools it executes (or another tool executes), to fail, but NOT because of a bug. An example problem is if there is not enough memory available or if some other library or syscall fails. Nonetheless it might be worth reporting as a bug; it is a judgement call: if it’s a bug it’ll be addressed and if it’s not that’s OK too!

We recommend AND encourage you to use the example Makefile, as the starting point for your submission’s required Makefile:

• view example Makefile
• download example Makefile

Feel free to modify the Makefile to suit your obfuscation needs.

Please add a space between the = and the value of variables, in the Makefile, making sure that the = comes immediately after the name. See the example Makefile for examples.

The rest of this section and its subsections will assume that you are using some variant of the example Makefile, again renamed as Makefile.

We suggest that you compile your submission with a commonly available -std=gnu17 (ISO C 2017 with GNU extensions) C compiler, via clang(1) and/or gcc(1).

Unless you clearly state otherwise in your remarks.md file, AND put in your submission’s Makefile, we will compile using -std=gnu17 -O3!

It is OK if you need to require your submission to NOT be compiled using the default -std=gnu17 -O3 settings. Simply explain why your submission should NOT be compiled using -std=gnu17 -O3 in your remarks.md file, AND adjust your Makefile accordingly.

One reason that you might have to change the flags, is that the optimiser is known to break some programs, but there are certainly other possible valid reasons. Again, just update the Makefile and explain it in your remarks.md. See the optimiser section for details for changing optimiser flags.

For more fun when it comes to optimisers breaking code, see 1986/marshall/compilers.html.

IMPORTANT NOTE: The use of -std=gnu17 does NOT imply the use of the gcc compiler! We often start by compiling using the clang C compiler instead.

PLEASE NOTE: in macOS, the compiler gcc found at /usr/bin/gcc is in truth the clang compiler, as /usr/bin/gcc --version will show!

You may change the standard under which your submission is compiled by modifying the CSTD Makefile variable. For example, to use c99 instead:

    CSTD= -std=c99

You may change the level of optimization and compiler debug level that your submission is compiled with, by modifying the OPT Makefile variable. For example, to compile without optimization, and to include debug symbols:

    OPT= -O0 -g3

The default warning flags are set via the CWARN variable, as shown in the example Makefile:

    # Common C compiler warning flags
    #
    CWARN= -Wall -Wextra ${CSILENCE} ${CUNKNOWN}

For details on CSILENCE and CUNKNOWN, see the section on disabling warnings.

For compilers, such as clang, that have the -Weverything option, while you may wish to try it, you should read our FAQ on “clang -Weverything”. We do NOT recommend that you put the use of -Weverything into your submission’s Makefile for the reasons cited there. This goes even if your version does not trigger a warning as some other version might!

On the other hand, if ${CC} has “clang” in the name, the example Makefile will automatically enable -Weverything, so you might have to use -Wno-foo options anyway, as detailed below. See the FAQ on “-Weverything” for more details.

If “clang” is NOT in ${CC}, the CWARN variable will not be further modified.

There is no real penalty for compiler warnings. Sometimes compiler warnings cannot be helped: especially in the case of obfuscated C. :-) So if you cannot easily get rid of a compiler warning, try not to fret too much.

We LIKE code that has a minimum of warnings, especially under the more strict -Wall -Wextra -pedantic mode:

    CWARN= -Wall -Wextra -pedantic

The two previous guidelines might be thought by some as being somewhat contradictory. Isn’t life, and isn’t trying to satisfy “contradictory customer requirements” all too often like that? :-) Try to minimize warnings if you can.

If you manage to produce very few warnings, or perhaps no warnings at all under the -Wall -Wextra -pedantic mode, then by all means brag about it in your remarks.md file AND BE SURE TO TELL US the OS, OS version, compiler and compiler version in which you observed this occurring (in case our OS and compiler produces a different result: so your submission won’t be penalized for not meeting your claims).

On the other hand, some warnings cannot be disabled and are enabled by compilers without any warning option specified. These are sometimes inevitable in obfuscated code and even in some non-obfuscated code, and you should not worry about this, though it might be worth pointing out.

For instance, some compilers like to warn about use of pointers as arrays, which seems to be dubious, as it obviously can’t (always) be avoided, being a big part of C, so you should not worry about this either; this is the warning -Wunsafe-buffer-usage and the way to disable it is -Wno-unsafe-buffer-usage. See also the FAQ on “forced warnings” and the FAQ on “-Weverything”.

Disabling warnings

If you do have to disable warnings due to -Weverything automatically being included, you might wish to state this fact in your remarks.md file. And even without -Weverything there can be warnings, as noted above.

If your submission issues lots of warnings but is otherwise marvelously obfuscated in multiple levels, don’t worry about it. Nevertheless, be sure that the warnings do not constitute a potential “show stopper” compiler problem. Be sure that compilers such as both gcc and clang won’t produce a compiler error and refuse to compile your code: unless for some reason that is what you intend to happen in which case document that too in your remarks.md file. :-)

All other things being equal, a program that must turn off fewer warnings will be considered better, for certain values of better.

To turn off a compiler warning, in your submission’s Makefile, try something such as:

    CSILENCE= -Wno-some-thing -Wno-another-thing

For instance:

    CSILENCE= -Wno-parentheses -Wno-binding-in-condition -Wno-misleading-indentation

If you do add “-Wno-foo” to your Makefile, consider changing:

    CUNKNOWN=

to:

    CUNKNOWN= -Wno-unknown-warning-option

Some compilers have reported this as an error, however, and if you have such a compiler you might want to not add it, or at least note in your remarks.md file, which OS, OS version, compiler and compiler version that has such a problem.

If you need to define something on the compile line, use the CDEFINE Makefile variable. For example:

    CDEFINE= -Dfoo -Dbar=baz

NOTE: just because we offer a default way to use the -D option, we still do not like excess use of -D to get past size limits, especially since you could use #define in your code instead.

If you need to include a file (as in #include) on the command line, use the CINCLUDE Makefile variable. For example:

    CINCLUDE= -include stdio.h

If you need to add other “magic” flags to your compile line, use the COTHER Makefile variable. For example:

    COTHER= -fno-math-errno

NOTE: We only recommend using “magic” flags if BOTH gcc and clang support it.

Again, please note that in macOS, /usr/bin/gcc is actually clang!

When make clobber is invoked, we request that submissions be restored to their original submission state. For example, any temporary files (including the compiled program(s)) created during the build process, or during execution should be removed by the clobber rule. In other words, the only things that should be in the directory after running make clobber is what is in your submission tarball itself.

While people are free to manage their submission under git(1) or even use a GitHub repo, dot-files and dot-directories such as .git are not allowed in a submission.

The mkiocccentry(1) tool will ignore dot-files and dot-directories (such as .vimrc, .bashrc, .git and .github) and not put them in the submission’s compressed tarball. So while you may use such files and directories to help develop your submission, they won’t be included when you run the mkiocccentry(1) tool.

Even if you did manage to get dot files or dot directories in the tarball somehow, txzchk(1) will flag it as an error. When the judges run txzchk(1) on the uploaded submission compressed tarball, if anything is wrong, for instance if you “sneak in” any dot files or dot directories, the submission WILL BE REJECTED for violating Rule 17!

You may use whatever tools you need to develop your submission, including the use of git(1) or gh(1), just be sure that your submission code and your submission Makefile don’t depend on such tools.

Of course if you do use GitHub to work on your submission, you might want to make the repo private so you don’t reveal who you are. :-)

If this is not clear, please do NOT use these tools to help with the clobber rule! For instance, do NOT use git clean! Not only does this depend on the user having git(1) but it also does not account for the submission tarballs. Even worse is when someone does have it in a git(1) repo it will remove files that are not under git(1) control! Instead, see the clobber rule in the example Makefile to see how to manage this.

In other words, for make clobber, do something like:

    clobber:
            ${RM} -f foo bar baz

and NOT something like this:

    # do NOT do this!
    clobber:
            -git clean -f   # WRONG !!!

And do NOT use git for any other Makefile rule either.

NOTE: mkiocccentry(1) will directly run make clobber in the submission directory. And although it is not an error if this fails, if it fails because the Makefile does not have a clobber rule, it will be flagged by mkiocccentry(1) (you may ignore this but it does put you at a risk of violating the rules).

Jump to: top

We VERY MUCH LIKE submissions that use an edited variant of the example Makefile, as described and linked to in the Makefile section, renamed as Makefile of course. This makes it easier for the IOCCC judges to test your submission. And if your submissions wins, it makes it easier to integrate it into the Official IOCCC winner website.

We VERY MUCH LIKE submissions that have some educational value. This does NOT mean that your submission should not be obfuscated but rather that the IOCCC has moved away from the idea of “spoilers”. You should not encrypt or rot13 content in your remarks.md file.

It is VERY MUCH appreciated if your remarks have some educational value. And although educational value is not required, it is an EXCELLENT bonus.

We LIKE submissions that use an edited version of the try.sh example script (and if you have alternate code, the same applies with the try.alt.sh script):

• view example try.sh

• download example try.sh

• view example try.alt.sh

• download example try.alt.sh

Of course, it is quite possible that only one invocation is possible, so it is not necessarily detrimental to your submission if you do not include one, though we do like interesting and creative uses of submissions. See also the FAQ on “submitting try.sh and try.alt.sh scripts”.

You might wish to add ./try.sh to the try rule in the Makefile you submit. If you have alternate code, then you can use the try.alt rule as well.

Doing masses of #defines to obscure the source has become ‘old’. We tend to ‘see thru’ masses of #defines due to our pre-processor tests that we apply. Simply abusing #defines or -Dfoo=bar won’t go as far as a program that is more well rounded in confusion.

Many C compilers DISLIKE the following code, and so do we:

    #define d define
    #d foo             /* <-- don't expect this to turn into #define foo */

In other words, it is a compilation error, and in order to get older IOCCC winning entries that did this to compile, we had to update them to not do this.

When declaring local or global variables, you should declare the type:

    int this_is_fine;
    this_is_not;       /* <-- Try to avoid implicit type declarations */

We tend to like less a submission that requires either gcc OR clang. We prefer submissions that can compile under BOTH gcc AND clang.

We RECOMMEND that the compiler flags you use in your submission’s Makefile are supported by BOTH gcc AND clang.

We DISLIKE the use of obscure compiler flags, especially if gcc and/or clang do not support it. We suggest that you not use any really obscure compiler flags if you can help it.

One side effect of the above is that you cannot assume the use of nested functions such as:

     int main() {
|        void please_dont_submit_this() {
|           printf("The machine that goes BING!!\n");
         }
|        please_dont_submit_this();
     }

On 2012 July 20, the IOCCC judges rescinded the encouragement of nested functions. Such constructions, while interesting and sometimes amusing, will have to wait until they are required by a C standard that are actually implemented in BOTH gcc AND clang.

We DISLIKE submissions that require the use of -fnested-functions.

We prefer programs that do not require a fish license: crayons and cat detector vans not withstanding.

If your submission uses functions that have a variable number of arguments, be careful. Systems implement va_list in a wide variety of ways. Because of this, a number of operations using va_list are not portable and must not be used:

• assigning a non-va_list variable to/from a va_list variable
• casting a non-va_list variable into/from a va_list variable
• passing a va_list variable to a function expecting a non-va_list arg
• passing a non-va_list variable to a function expecting a va_list arg
• performing arithmetic on va_list variables
• using va_list as a structure or union

In particular, do not treat va_list variables as if they were a char **.

We DISLIKE the use of varargs.h. Use stdarg.h instead.

We DISLIKE the use of gets(3). Use fgets(3) instead.

We tend to DISLIKE the blatant use of tarballs in an attempt to simply get around the extra file number limit. We realize there may be cases where a tarball containing a number of extra files may be needed. Such a need for a tarball MUST be explained in the remarks.md file.

On 28 January 2007, the Judges rescinded the requirement that the # in a C preprocessor directive must be the 1st non-whitespace byte.

The exit(3) function returns void. Some broken systems have exit(3) return int; your submission should assume that exit(3) returns a void.

This guideline has a change mark at the very start of this line.

Small programs are best when they are short, obscure and concise. While such programs are not as complex as other winners, they do serve a useful purpose: they are often the only program that people attempt to completely understand. For this reason, we look for programs that are compact, and are instructional.

While those who are used to temperatures found on dwarf planets (yes Virginia, dwarf planets ARE planets!), such as Pluto, might be able to explain to the Walrus why our seas are boiling hot, the question of whether pigs have wings is likely to remain a debatable point to most.

One line programs should be short one line programs: say around 80 to 132 bytes long. Going well beyond 132 bytes is a bit too long to be called a one-liner in our vague opinion.

We tend to DISLIKE programs that:

• are very hardware specific
• are very OS version specific (index(3)/strchr(3) differences are OK, but sockets/streams specific code is likely not to be)
• dump core or have compiler warnings (it is OK only if you warn us in your remarks.md file)

• won’t compile or run in a Single UNIX Specification environment

• depend on a utility or application not normally found in systems that conform to the Single UNIX Specification

• abuse the build file to get around the size limit
  

• obfuscate by use of ANSI trigraphs

• obfuscate by use of digraphs

• are larger than they need to be

• have more lines than they need to have

• are “blob-ier” (just a pile of unformatted C code) than they need to be

• are rather similar to previous winners :-(

• are identical to previous losers :-)

• that mandate the exclusive use of a specific Integrated Development Environment (IDE)

In order to encourage submission portability, we DISLIKE submissions that fail to build unless one is using an IDE. For example, do not mandate that one must use Microsoft Visual Studio to compile your submission.

The program must compile and link cleanly in a Single UNIX Specification environment. Therefore do not assume the system has a windows.h include file:

    #include <windows.h>  /* we DISLIKE this */

Unless you are cramped for space, or unless you are entering the ‘Best one liner’ category, we suggest that you format your program in a more creative way than simply forming excessively long lines.

At least one judge prefers to maintain the use of the leap-second as part of the world’s time standard. If your code prints time with seconds, we prefer that your code be capable of printing the time of day during a leap-second where the value in seconds after the minute mark is 60.

The Makefile should not be used to try and get around the size limit. It is one thing to make use of a several -Ds on the compile line to help out, but it is quite another to use many bytes of -Ds in order to try and squeeze the source under the size limit.

Your source code, post-pre-processing, should not exceed the size of Microsoft Windows. :-)

Other windows, on the other hand, might be OK: especially where “X marks the spot”. Yet on the third hand, windows are best when they are “unseen” (i.e., not dirty). :-)

The IOCCC judges, as a group, have a history giving wide degree of latitude to reasonable submissions. And recently they have had as much longitudinal variation as it is possible to have on Earth. :-)

You should try to restrict commands used in the build file to commands found in Single UNIX Specification environments and systems that conform to the Single UNIX Specification.

You may compile and use your own programs. If you do, try to build and execute from the current directory. This restriction is not a hard and absolute one. The intent is to ensure that the building of your program is reasonably portable.

We prefer programs that are portable across a wide variety of UNIX-like operating systems (e.g., Linux, GNU Hurd, BSD, UNIX, etc.).

You are in a maze of twisty guidelines, all different.

There are at least zero IOCCC judges who think that Fideism has little or nothing to do with the IOCCC judging process.

Don’t forget that the building of your program should be done WITHOUT human intervention. So don’t do things such as:

    prog: prog.c
        #echo this next line requires data from standard input
        cat > prog.c
        ${CC} prog.c -o prog

However, you can do something cute such as making your program do something dumb (or cute) when it is built ‘automatically’, and when it is run with a human involved, do something more clever. For example, one could put in their Makefile:

    prog: prog.c
        ${CC} prog.c -DNON_HUMAN_COMPILE -o prog
        @echo "See remarks section about alternate ways to compile"

and then include special notes in your remarks.md file for alternate / human intervention based building.

We want to get away from source that is simply a compact blob of bytes. REALLY TRY to be more creative than blob coding. HINT!

Please do not use things like gzip(1) to get around the size limit. Please try to be much more creative.

We really DISLIKE submissions that make blatant use of including large data files to get around the source code size limit.

We do not recommend submitting systemd source code to the IOCCC, if nothing else because that code is likely to exceed the source code size limit. This isn’t to say that another highly compact and obfuscated replacement of init would not be an interesting submission.

Did we remember to indicate that programs that blatantly use some complex state machine to do something simple, are boring? We think we did. :-)

All generalizations are false, including this one. – Mark Twain

Given two versions of the same program, one that is a compact blob of code, and the other that is formatted more like a typical C program, we tend to favor the second version. Of course, a third version of the same program that is formatted in an interesting and/or obfuscated way, would definitely win over the first two! Remember, you can submit more than one submission. See the IOCCC rules for details (in particular, Rule 9).

We suggest that you avoid trying for the ‘smallest self-replicating’ source. The smallest, a zero byte entry, won in 1994.

Programs that claim to be the smallest C source that does something, really better be the smallest such program or they risk being rejected because they do not work as documented.

Please note that the C source below, besides lacking in obfuscation, is NOT the smallest C source file that when compiled and run, dumps core:

    main;

Unless you specify -fwritable-strings (see COTHER in the example Makefile, described in the Makefile section), do not assume this sort of code will work:

    char *T = "So many primes, so little time!";
    ...
    T[14] = ';';    /* modifying a string requires: -fwritable-strings */

Even so, one should probably not assume that this is universally accepted.

Initialized char arrays are OK to write over. For instance, this is OK:

    char b[] = "Is this OK";
    b[9] = 'k';     /* modifying an initialized char array is OK */

There are more than 1 typos in this very sentence.

X client submissions should be as portable as possible. Submissions that adapt to a wide collection of environments will be favored. For example, don’t depend on a particular type or size of display. Don’t assume the use of a particular browser. Instead assume a generic browser that forms to a widely used W3C standard. Don’t assume a particular sound sub-system or video driver is installed in the OS. Instead, make use of a well known and widely available open source program (one that actually works) to display audio/visual data.

X client submissions should avoid using X related libraries and software that are not in wide spread use.

As of Red Hat RHEL9.0, the X.org server is deprecated. See the FAQ on “Xorg deprecation”” for more details. This does not mean that a submission using this will necessarily be rejected, but it would be better if it can support Wayland in some way or another.

This could be the only guideline that contains the word fizzbin.

However, do you know how to play fizzbin? You do?!? (Except on Tuesday?)

OK, there are actually 2 guidelines that contain the word fizzbin, unless you count this one, in which case there are 3. :-)

We DISLIKE submissions that use proprietary toolkits such as the M*tif, Xv*ew, or OpenL*ok toolkits, since not everyone has them. Use an open source toolkit that is widely and freely available instead.

NOTE: The previous guideline in this spot has been replaced by this guideline:

X client submissions should try to not to depend on particular items in .Xdefaults. If you must do so, be sure to note the required lines in the your remarks.md file. They should also not depend on any particular window manager.

Try to avoid submissions that play music that some people believe is copyrighted music.

While we recognize that UNIX is not a universal operating system, the contest does have a bias towards such systems. In an effort to expand the scope of the contest, we phrase our bias to favor the Single UNIX Specification.

You are well advised to submit code that conforms to the Single UNIX Specification Version 4.

To quote the IOCCC judges:

You very well might not be completely prohibited from failing to not partly misunderstand this particular guideline, but of course, we could not possibly comment! :-) Nevertheless, you are neither prohibited, nor are you fully required to determine that this or the previous sentence is either false and/or perhaps misleading. Therefore, it might be wise for you to not fail to consider to not do so, accordingly. Thank you very much.

Any complaints about the above guideline could be addressed to the Speaker of the House of Commons, or to the speaker of your national parliament should you have one.

We LIKE programs that:

• are as concise and small as they need to be
• do something at least quasi-interesting
• are portable
• are unique or novel in their obfuscation style
• make use of a NUMBER OF DIFFERENT TYPES OF OBFUSCATIONS! <== HINT!!
• make us laugh and/or throw up :-) (humor really helps!)
• make us want to eat good chocolate

Some types of programs can’t excel (anti-tm) in some areas. Your program doesn’t have to excel in all areas, but doing well in several areas really does help.

You are better off explaining what your submission does in your remarks.md file section rather than leaving it obscure for the IOCCC judges as we might miss something and/or be too tired to notice.

Please avoid this specific individual guideline, if it at all possible.

We freely admit that interesting, creative or humorous comments in your remarks.md file help your chances of winning. If you had to read so many twisted submissions, you too would enjoy a good laugh or two. We think the readers of the contest winners do as well. We do read your remarks.md content during the judging process, so it is worth your while to write a remarkable remarks.md file.

We DISLIKE C code with trailing control-M’s (\r or \015) that results in compilation failures. Some non-UNIX/non-Linux tools such as MS Visual C and MS Visual C++ leave trailing control-M’s on lines. Users of such tools should strip off such control-M’s before submitting their submissions. In some cases tools have a “Save As” option that will prevent such trailing control-M’s being added.

One should restrict libcurses to portable features found on BSD or Linux curses.

Rule 13 no longer discourages the use of UTF-8 characters in C code.

It is a very good idea to, in your remarks.md file, tell us why you think your submission is obfuscated. This is particularly true if your submission has some very subtle obfuscations that we might otherwise overlook. <<– Hint!

Anyone can format their code into a dense blob. A really clever author will try format their submission using a “normal” formatting style such that at first glance (if you squint and don’t look at the details) the code might pass for non-obfuscated C. Deceptive comments, and misleading formatting, in some cases, may be a plus. On the other hand, a misleading code style requires more source bytes.

If you do elect to use misleading formatting and comments, we suggest you remark on this point in your remarks.md where you talk about why you think your submission is obfuscated. On the other hand, if you are pushing up against the size limits, you may be forced into creating a dense blob. Such are the trade-offs that obfuscators face!

We prefer code that can run on either a 64-bit or 32-bit processor. However, it is UNWISE to assume it will run on an some Intel-like x86 architecture**.

We believe that Mark Twain’s quote:

Get your facts first, then you can distort them as you please.

… is a good motto for those writing code for the IOCCC.

The IOCCC size tool source is not an original work, unless you are Anthony C Howe, in which case it is original! :-) Submitting source code that uses the contents of iocccsize.c, unless you are Anthony C Howe, might run the risk of violating Rule 7.

The IOCCC submission tarball validator source is not an original work, unless you are Cody Boone Ferguson, in which case it is original! :-) Submitting source code that uses the contents of txzchk.c, unless you are Cody Boone Ferguson, might run the risk of violating Rule 7.

In addition to the above tools, none of the mkiocccentry tools, including, but not limited to mkiocccentry, chkentry, fnamchk and location, nor the JSON parser and library, jstrencode, jstrdecode or any of the other jparse tools (see the jparse repo), are original works, unless you are Cody Boone Ferguson or Landon Curt Noll, in which case they are original! :-) Submitting source code that uses the content of any of these tools or library, unless you are Cody Boone Ferguson or Landon Curt Noll, might run the risk of violating Rule 7.

And unless you are Landon Curt Noll, the code in dbg and dyn_array are not original works. Submitting source code that uses the contents of those libraries, unless you Landon Curt Noll, might run the risk of violating Rule 7.

Rule 7 does not prohibit you from writing your own obfuscated versions of these tools, unless of course you are Landon Curt Noll, in which case you probably won’t win since IOCCC judges should not enter the IOCCC! :-) However, if you do write your own version, you might wish to make it do something more interesting than simply implementing the IOCCC tools’ algorithms.

Even so, we do not recommend you try and submit a JSON parser due to the fact it will likely exceed the source code size limit and because you likely can’t beat flex and bison in obfuscation. This isn’t to say that the so-called JSON spec is not obfuscated, but unless you have some really clever way to compact and obfuscate a JSON parser more than flex and bison you will likely not win, either because of the source code size limit or because it is not as obfuscated as the flex and bison generated code of jparse.

While programs that only run in a specific word size are OK, if you have to pick, choose a 64-bit word size.

If the IOCCC judges are feeling ornery we might choose to compile your program for running on an Arduino or a PDP-11. Heck, should we ever find an emulator of 60-bit CDC Cyber CPU, we might just try your submission on that emulator as well :-)

If your submission MUST run only on a 64-bit or 32-bit architecture, then you MUST specify the -arch on your command line (see ARCH in the example Makefile, described in Makefile section). Do not assume a processor word size without specifying -arch. For example:

    ARCH= -m64

Note, however, that some platforms will not necessarily support some architectures. For instance, more recent versions of macOS do NOT support 32-bit!

If there are limitations in your submission, you are highly encouraged to note such limitations in your remarks.md file. For example if your submission factors values up to a certain size, you might want to state:

This submission factors values up 2305567963945518424753102147331756070. Attempting to factor larger values will produce unpredictable results.

The IOCCC judges might try to factor the value -5, so you want to might state:

This submission factors positive values up 2305567963945518424753102147331756070. Attempting to factor large values will produce unpredictable results.

However the IOCCC judges might try to also factor 0, so you want to might state:

This submission factors values between 1 and 2305567963945518424753102147331756070. Attempting to factor values outside that range will produce unpredictable results.

Moreover they might try to also factor 3.5 or 0x7, or Fred, so you want to might state:

This submission factors integers between 1 and 2305567963945518424753102147331756070. Attempting to factor anything else will produce unpredictable results.

You submission might be better off catching the attempt to factor bogus values and doing something interesting. So you might want to code accordingly and state:

This submission factors integers between 1 and 2305567963945518424753102147331756070. Attempting to factor anything else will cause the program to insult your pet fish Eric.

The IOCCC judges might not have a pet fish named Eric, so you might want to state:

This submission factors integers between 1 and 2305567963945518424753102147331756070. Attempting to factor anything else will cause the program to insult your pet fish Eric, or in the case that you lack such a pet, will insult the pet that you do not have.

When all other things are equal, a submission with fewer limitations will be judged better than a submission with lots of limitations. So you might want to code accordingly and state:

This submission attempts to a factor value of any size provided that the program is given enough time and memory. If the value is not a proper integer, the program might insult a fish named Eric.

Do not fear if you’re not 100% sure of the significance of 2305567963945518424753102147331756070 as it is not of prime importance: or is it? :-)

We DISLIKE the use of ASCII tab characters in markdown files, such as in the required remarks.md file.

We don’t mind the use of ASCII tab characters in your C code. Feel free to use ASCII tab characters if that suits your obfuscation needs. If is perfectly OK to use tab characters elsewhere in your submission, just not in markdown files as this tends annoy us when it comes time to rendering your markdown content as it complicates the process.

If you do use ASCII tab characters in your non-markdown files, be aware that some people may use a tab stop that is different than the common 8 character tab stop.

PLEASE observe our IOCCC markdown guidelines when forming your submission’s remarks.md file. And if your submission contains additional markdown files, please follow those same guidelines for those files. See also Rule 19 and our FAQ on “markdown”.

We LIKE reading remarks.md files, especially if they contain useful, informative, and even humorous content about your submission. Yes, this is a hint. :-)

We RECOMMEND you put a reasonable amount effort into the content of the remarks.md file: it is a required file for a reason. :-)

Try to be even more creative!

Jump to: top

Legal abuse of the IOCCC rules is somewhat encouraged. Legal rule abuse may involve, but is not limited to, doing things that are technically allowed by the IOCCC rules and yet do not fit the spirit of what we intended to be submitted.

Legal rule abuse is encouraged to help promote creativity. Rule abuse submissions, regardless of if they receive an award, result in changes to the next year’s IOCCC rules and IOCCC guidelines.

Legal abuse of the IOCCC rules is NOT an invitation to violate the IOCCC rules. A submission that violates the rules in the opinion of the IOCCC judges, WILL be disqualified. RULE ABUSE CARRIES A CERTAIN LEVEL OF RISK! If you have a submission that might otherwise be interesting, you might want to submit two versions; one that does not abuse the IOCCC rules and one that does.

If you intend to abuse the IOCCC rules, indicate so in your remarks.md file. You MUST try to justify why you consider your rule abuse to be allowed under the IOCCC rules. That is, you must plead your case as to why your submission is valid. Humor and/or creativity help plead a case. As there is no guarantee that you will succeed, you might consider submitting an alternate version that conforms to the IOCCC rules.

If you do bypass the mkiocccentry(1) warnings about Rule 2a and/or about Rule 2b or any other rule and submit a submission anyway, you MUST try to justify why the IOCCC IOCCC judges should not reject your submission due to a rule violation, and you would be wise to do this towards the top of your remarks.md file.

Abusing the web submission procedure tends to annoy us more than amuse us. Spend your creative energy on content of your submission rather than on the submission process itself.

We are often asked why the contest IOCCC rules and IOCCC guidelines seem strange or contain mistakes, flaws or grammatical errors. One reason is that we sometimes make genuine mistakes. But in many cases such problems, flaws or areas of confusion are deliberate. Changes to IOCCC rules and IOCCC guidelines in response to rule abuses, are done in a minimal fashion. Often we will deliberately leave behind holes (or introduce new ones) so that future rule abuse can continue. A clever author should be able to read them and “drive a truck through the holes” in the IOCCC rules and IOCCC guidelines.

At the risk of stating the obvious, this contest is a parody of the software development process. The IOCCC rules and IOCCC guidelines are only part of the overall contest. Even so, one might think the contest IOCCC rules and IOCCC guidelines process as a parody of the sometimes tragic mismatch between what a customer (or marketing) wants and what engineering delivers. Real programmers must face obfuscated and sometimes conflicting specifications and requirements from marketing, sales, product management and even from customers themselves on an all too regular basis. This is one of the reasons why the IOCCC rules and IOCCC guidelines are written in obfuscated form.

Jump to: top

Submissions are judged by Leonid A. Broukhis and Landon Curt Noll.

Each submission submitted is given a random string and subdirectory. The submission files including, but not limited to prog.c, Makefile, remarks.md, .info.json, .auth.json as well as any data files that you submit, are all placed under their own directory and stored and judged from this directory.

Any information about the authors is not read by the IOCCC judges until the judging process is complete, and then only from entries that have won an award. Because we do not read this information for entries that do not win, we do not know who did not win.

The above process helps keep us biased for/against any one particular individual. Therefore you MUST refrain from putting any information that reveals your identity in your submission.

Now some people point out that coding and/or writing style might reveal the information about the authors. However we consider this to be simply circumstantial and outside the scope of the above paragraph.

Some people, in the past, have attempted to obfuscate their identity by including comments of famous Internet personalities such as Peter Honeyman. The IOCCC judges are on to this trick and therefore consider any obfuscated source or data file claiming to be from Honeyman to not be from Honeyman. This of course creates an interesting paradox known as the “obfuscated Peter Honeyman paradox”. Should Peter Honeyman actually submit to the IOCCC, he alone is excluded from the above, as we will likely believe it’s just another attempt at confusion. This guideline is known as the “Peter Honeyman is exempt” guideline.

BTW: None of the submissions claiming to be from Peter Honeyman have ever won an award. So it is theoretically possible that Peter Honeyman did submit to the IOCCC in the past. In the past, Peter had denied submitting anything to the IOCCC. Perhaps those submissions were submitted by one of his students?

Hopefully we are VERY CLEAR on this point! The rules now strongly state: PLEASE DO NOT put a name of an author, in an obvious way, into your source code, remarks.md, data files, etc., the above “Peter Honeyman is exempt” notwithstanding.

We seemed to have digressed again … :-) Returning to the judging process:

We prefer to be kept in the dark as much as you are until the final awards are given. We enjoy the surprise of finding out in the end, who won and where they are from.

We attempt to keep all submissions anonymous, unless they win an award. Because the main ‘prize’ of winning is being announced, we make all attempts to send non-winners into oblivion. We remove all non-winning files, and shred all related printouts. By tradition, we do not even reveal the number of submissions that we receive.

During the judging process, a process that spans multiple sessions over a few weeks, we post general updates from our Mastodon account.

Make sure you reload the feed every so often because unless you are mentioned you will NOT get a push notification!

Jump to: top

Judging consists of a number of elimination rounds. During a round, the collection of submissions are divided into two roughly equal piles; the pile that advances on to the next round, and the pile that does not. We also re-examine the submissions that were eliminated in the previous round. Thus, a submission gets at least two readings and often more than two.

Jump to: top

A reading consists of a number of actions:

• Reading prog.c, the C source and reviewing the remarks.md information

• Briefly looking at any supplied data files

• Passing the source thru the C pre-processor skipping over any #included files
• Performing a number of C beautify/cleanup edits on the source
• Passing the beautified source thru the C pre-processor skipping over any #included files
• Compiling/building the source
• Running the program
• Doing other things that only IOCCC judges know about :-)

In later rounds, other actions are performed including performing miscellaneous tests on the source and binary.

This is the very guideline that goes, BING!

Okay, technically that was the second guideline that goes BING, and this is the third. :-)

Until we reduce the stack of submissions down to about 25 submissions, submissions are judged on an individual basis. A submission is set aside because it does not, in our opinion, meet the standard established by the round. When the number of submissions thins to about 25 submissions, we begin to form award categories. Submissions begin to compete with each other for awards. A submission will often compete in several categories.

The actual award category list will vary depending on the types of submissions we receive. A typical category list might be:

• best small one line program (see above about one line programs)
• best small program
• strangest/most creative source layout
• most useful obfuscated program
• best game that is obfuscated
• most creatively obfuscated program
• most deceptive C code (code with deceptive comments and source code)

• best X program (see OUR LIKES AND DISLIKES and the FAQ on “Xorg being deprecated”)

• best abuse of ISO C or ANSI C standard (see above about compilers)
• best abuse of the C preprocessor
• worst/best abuse of the rules (or some variation)
• (anything else so strange that it deserves an award)

We do not limit ourselves to this list. For example, a few submissions are so good/bad that they are declared winners at the start of the final round. We will invent awards categories for them, if necessary.

In the final round process, we perform the difficult tasks of reducing the remaining submissions (typically about 25) down to to about half that number: declaring those remaining to be winners.

Often we are confident that the submissions that make it into the final round are definitely better than the ones that do not make it. The selection of the winners out of the final round, is less clear cut, however.

Sometimes a final round submission is good enough to win, but is beat out by a similar, but slightly better submission. For this reason, it is sometimes worthwhile to resubmit an improved version of a submission that failed to win in a previous year, the next year. This assumes, of course, that the submission is worth improving in the first place!

Over the years, more than one IOCCC judge has been known to bribe another IOCCC judge into voting for a winning submissions by offering a bit of high quality chocolate, or other fun item.

One should NOT attempt to bribe an IOCCC judge, unless you are an IOCCC judge, because bribing an IOCCC judge by a non-judge has been shown to NOT be effective when the person attempting the bribe is made known to the IOCCC judges (i.e., they are not anonymous) AND/OR the bribe is otherwise associated with a submission to the IOCCC.

With the previous guideline in mind: anonymous gifts for the IOCCC judges that are NOT ASSOCIATED WITH a submission to the IOCCC may be sent to the IOCCC judges via the IOCCC Amazon wishlist. It has been shown that receiving anonymous gifts provides the IOCCC judges with a nice dopamine boost, and happy IOCCC judges help make the IOCCC better for everyone. :-)

See the FAQ on “supporting the IOCCC”.

More often than not, we select a small submission (usually one line) and a strange/creative layout submission. We sometimes also select a submission that abuses the IOCCC guidelines in an interesting way, or that stretches the contest rules that while legal, it nevertheless goes against the intent of the rules.

Nevertheless, see Rule 12.

In the end, we traditionally pick one submission as ‘best’. Sometimes such a submission simply far exceeds any of the other submissions. More often, the ‘best’ is picked because it does well in a number of categories.

In years past, we renamed the winning submission from prog.c to a name related to the author(s)’ names. This is no longer done. Winning source is called prog.c. A compiled binary is called prog.

Jump to: top

The IOCCC judges will toot initial announcement of who won, the name of their award, and a very brief description (award title) of the winning entry from the @IOCCC Mastodon account.

We recommend that you follow us on mastodon but please make sure to refresh the feed every so often (if not more often) because unless you are mentioned or someone boosts your post you will not get a push notification.

Jump to: top

The current status of the IOCCC will change from judging to closed .

The contest_status in the status.json file will change from judging to closed as well.

When the above happens, the winning entries have been selected by the IOCCC judges.

The IOCCC judges will begin to prepare to release the source code of the new IOCCC winners.

The IOCCC judges will commit the winning source to the IOCCC winner repo which will update the Official IOCCC website.

The IOCCC news will also contain an announcement of the winners.

Jump to: top

The IOCCC no longer uses twitter. Winning IOCCC entries will be announced by a git commit to the IOCCC entries repo that, in turn, updates the Official IOCCC website.

In addition a note is posted to the IOCCC Mastodon account.

We will also post a notice to the IOCCC discord server.

Consider joining the IOCCC discord community via this link: https://discord.gg/Wa42Qujwnw

Jump to: top

It is pointless to ask the IOCCC judges how many submissions we receive. See How many submissions do the judges receive for a given IOCCC?.

Often, winning entries are published in selected magazines from around the world. Winners have appeared in books (‘The New Hacker's Dictionary’, ‘Obfuscated C and Other Mysteries’, ‘Pointers On C’, others) and on t-shirts (sometimes by the author(s) themselves). There have been multiple classes that have focused in part, on IOCCC winning entries. And more than one winner has been turned into a tattoo!

Last, but not least, winners receive international fame and flames! :-)

Jump to: top

For questions or comments about the contest, see Contacting the IOCCC.

Be SURE to review the IOCCC Rules and Guidelines as they may (and often do) change from year to year.

PLEASE be sure you have the current IOCCC rules and IOCCC guidelines prior to submitting to the contest.

See the Official IOCCC website news for additional information.

For the updates and breaking IOCCC news, you are encouraged to follow the IOCCC on Mastodon. See our FAQ on “Mastodon” for more information. Please do note that unless you are mentioned by us you will NOT get a notification from the app so you should refresh the page even if you do follow us.

Check out the Official IOCCC winner website in general.

Consider joining the IOCCC discord community via this link: https://discord.gg/Wa42Qujwnw

See the FAQs on “obtaining, compiling, installing and using the mkiocccentry tools” and the FAQ on “how to enter the IOCCC” as that FAQ has important details on how to register as well as how to upload your submission for the IOCCC.

---

Jump to: top