ioccc/next/guidelines.md

29th International Obfuscated C Code Contest Official Guidelines

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

All Rights Reserved. Permission for personal, education or non-profit use is granted provided 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.

Guidelines Version

These IOCCC guidelines are version **29.01 2025-11-29**.

The | symbol indicates an **change** from the previous IOCCC.

Because the **IOCCC29** guidelines was a substantial rewrite, only **important changes** from **IOCCC28** have been marked.

Be sure to read the IOCCC rules.

# Guidelines for [Rule 2 - Size restrictions](rules.html#rule2-size-restrictions)

Guidelines for Rule 2 - Size

You should pay attention to our observations on [code size and Rule 2 remarks for IOCCC28](../2024/index.html#code-size-and-rule-2).

Your submission must satisfy BOTH the Rule 2a Gross Size AND Rule 2b Net Size rules.

To check your code against Rule 2 - Size restrictions, use the iocccentry(1) tool. Your submission must satisfy BOTH the maximum size Rule 2a Gross Size AND the Rule 2b Net Size.

To check your code against Rule 2 - Size, 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.

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

Make sure iocccsize does not flag any issues with your prog.c.

There a much less need to #define C reserved words in an effort to get around the size limits of Rule 2 - Size restrictions because of the iocccsize(1) algorithm. by the current version of iocccsize(1) is preferred by the IOCCC judges. Make sure iocccsize(1) does not flag any issues with your prog.c.

There a much less need to #define C reserved words in an effort to get around the size limits of Rule 2 - Size because of the iocccsize(1) algorithm.

Yes Virginia, the previous guideline sentence is an important hint!

Jump to: top

# Guidelines for [Rule 3 - Register for the IOCCC](rules.html#rule3-register)

Guidelines for Rule 3 - Register

See FAQ on "What is an author handle?".

Jump to: top

# Guidelines for [Rule 7 - Original Work](rules.html#rule7-original-work)

Guidelines for Rule 7 - Original Work

While it is NOT required, you are allowed to use tools or services such as:

• Code generators
• Static and runtime code analysis tools
• Machine learning tools
• Natural language models
• Large language models (LLMs)
• Code copilot tools
• So called "AI services"
• Etc.

The IOCCC has a rich history of remarkable winning entries created by authors who skillfully employed various techniques to develop their code. 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.

In short: You created it, you own it, its an original work, or you have explicit permission to the work.

See FAQ on "May I use AI, LLM, Virtual coding assistants, or similar tools to write my submission?".

Jump to: top

# Guidelines for [Rule 11 - Legal rule abuse](rules.html#rule11-legal-rule-abuse)

We do realize that there are holes in the IOCCC rules, and invite submitters to attempt to apply creative rule interpretation. A rule abuse though, even if not rewarded, may result in a subsequent rule change for future contests.

We sometime award 'Best abuse of the rules' or 'Worst abuse of the rules', or some variation, to a submission that creatively attempts to exploit a holes in the We sometimes award 'Best abuse of the rules' or 'Worst abuse of the rules', or some variation to a submission that creatively attempts to exploit a holes in the IOCCC rules.

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. :-)

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 abuse of the IOCCC rules is NOT an invitation to violate the IOCCC rules (especailly Rule 17 - Packaging ). 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, making sure to note in the one that does not abuse the rules that this is another version, so that the judges do not assume you uploaded it by mistake.

If you intend to abuse the IOCCC rules, indicate so in your remarks.md file. You MUST try to justify, plead, beg, why you consider your rule abuse to be allowed under the IOCCC rules. Humor 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 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, making sure to note in the one that does not abuse the rules that this is another version, so that the judges do not assume you uploaded it by mistake.

If you do bypass the mkiocccentry(1) warnings about Rule 2a - Gross Size and/or about Rule 2b - Net Size Rule 2a Gross Size or about Rule 2b Net Size 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.

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, balloning 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

# Guidelines for [Rule 15 - GNU Makefile](rules.html#rule15-gnu-makefile)

You may use git, or services such as GitHub to develop and maintain your submission. However, we DISLIKE submissions that require such services 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 GNU make(1) tool, your Makefile, as well as tools found in the Single UNIX Specification environments, and commonly found on systems that reasonably conform to the Single UNIX Specification in the building and compiling of your submission.

Your Makefile MUST be compatible with GNU make and we suggest you use Makefile.example as a template, renamed as Makefile of course.

Jump to: top

# Guidelines for [Rule 17 - Use mkiocccentry](rules.html#rule17-mkiocccentry)

Guidelines for Rule 17 - Packaging

We STRONGLY recommend you do install the most recent release of mkiocccentry toolkit As use of using older versions of the mkiocccentry toolkit to are HIGHLY LIKELY to be REJECTED!

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 invalid JSON files, PLEASE document and clearly explain this in your remarks.md file.

An _extra file_ is defined as a file that is not `prog.c`, `prog.alt.c`, `try.sh`, `try.alt.sh`, `remarks.md`, `Makefile`, `.auth.json` and `.info.json`: unless it's not in the top level directory, in which case it **IS** an _extra file_.

The non-extra files (`prog.c`, `prog.alt.c`, `try.sh`, `try.alt.sh`, `remarks.md`, `Makefile`, `.auth.json` and `.info.json`) are considered _free files_, and to not count towards the [Rule 17 - Use mkiocccentry](rules.html#rule17-mkiocccentry) limit of **31** _extra files_. These files (`prog.c`, `prog.alt.c`, `try.sh`, `try.alt.sh`, `remarks.md`, `Makefile`, `.auth.json` and `.info.json`) do _not_ count towards the [Rule 17 - Packaging](rules.html#rule-17---packaging) limit of **31** _extra files_.

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.

The mkiocccentry generated tarball does NOT reveal who you are! For this reason, the tarball MUST be a v7 format tarball.

You should NOT use a tarball as an extra file for a test-suite unless you have a very good reason for this: if you do include a tarball as an extra file, then you MUST specify why in your remarks.md file.

If you DO include a tarball an extra file, and the build process or the program extracts said tarball(s), the make clobber rule MUST remove the extracted files.

Instead of including a large test-suite that requires a lot of files as part of your submission, and your submission doesn't require the test-suite to be available to run, then in your remarks.md a URL where such a test-suite may be downloaded from and how to use the test-suite to test your submission.

If you DO include a tarball then make sure it uses the v7 format. See the FAQ on "extra files" for details on how to make a v7 tarball.

See the FAQ on "What should I do with the try.sh and try.alt.sh scripts?".

See the FAQ on "What is a .auth.json file?".

See the FAQ on "What is a .info.json file?".

See the FAQ on "How can I validate my submission directory?".

See the FAQ on "What permissions may my files be and what if I need different permissions?".

See the FAQ on "What are the detailed recommendations for a submission Makefile?".

See the FAQ on "What are the details behind Rule 17 - Use mkiocccentry?".

Jump to: top

Guidelines for the mkiocccentry toolkit

The IOCCC submit server, and related services use the most recent version of the mkiocccentry toolkit: submissions that package using older versions of the mkiocccentry toolkit are LIKELY to be REJECTED!

You REALLY SHOULD use the latest mkiocccentry toolkit are LIKELY to be REJECTED! So you REALLY SHOULD use the latest mkiocccentry toolkit to package your own submission.

The mkiocccentry(1) tool runs a number of checks by way of these mkiocccentry toolkit tools:

• iocccsize(1)
• txzchk(1)
• fnamchk(1)
• chksubmit(1)
• chkentry(1)

Do NOT use txzchk(1) or fnamchk(1) command line options that are labeled "for TESTING purposes" in their respective man pages.

Do NOT use fnamchk(1) command line options that are labeled "for TESTING purposes" in the fnamchk(1) man page.

Do NOT use chkentry(1) command line options that are labeled "for the use by the IOCCC judges only" in the chksubmit(1) (or chkentry(1)) man page(s). Similarly do NOT use chkentry(1) command line options marked "for the use by the IOCCC judges only" in the chkentry(1) man page.

To view mkiocccentry toolkit man pages, while your current directory is the top of the source tree:

man soup/man/man1/mkiocccentry.1
man soup/man/man1/iocccsize.1
man soup/man/man1/txzchk.1
man test_ioccc/man/man1/fnamchk.1
man soup/man/man1/chksubmit.1
man soup/man/man1/chkentry.1

The above man(1) commands will show the toolkit man pages without having to install them first.

So that you do not have to repeatedly answer all the mkiocccentry(1) questions, use mkiocccentry -a answers ... to create a file containing the answers you first give, so later enter mkiocccentry -i answers ... to reuse those same answers.

The mkiocccentry -A answers ... does the same thing except that the answers file will be overwritten.

The mkiocccentry -i answers ... will still be required to confirm "Yes/No" questions.

Use mkiocccentry -Y -i answers .. to force a yes answers with great caution because it might force a yes answer to something you do NOT want, such as a change in your code that violates a detected rule. The mkiocccentry -i answers .. will still be required to confirm "Yes/No" questions. While mkiocccentry -Y -i answers .. will force yes answers, use with great caution, because it might force a yes answer to something you do NOT want.

To help with not having to repeatedly enter a UUID, put your UUID into a file and use mkiocccentry -u uuidfile ..., or use mkiocccentry -U UUID ... to give the UUID on the command line.

If you wish to test that your submission passes the mkiocccentry(1) tests without having to type 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

Do NOT submit a tarball formed by mkiocccentry -d .. OR by mkiocccentry -s seed .. OR by using the test UUID!

The mkiocccentry(1) tool creates an xz(1) compressed v7 format tarball using a tar(1) command of the form:

cd workdir/UUID-slot_num && tar --format=v7 -cJf submit.UUID-slot_num.timestamp.txz UUID-slot_num

In many places the mkiocccentry(1) tool will prompt you to verify what you input, allowing you to correct details as you go along.

Pressing the "enter key" at the mkiocccentry(1) tool prompts will allow you to proceed with a default, however make sure the default is the CORRECT answer.

The mkiocccentry(1) tool will use code from iocccsize(1) which detects a number of issues (such as Rule 2 - Size) that you may ignore, however ignoring such issues comes with a significant level of risk! issues (such as Rule 2 - Size) that you may ignore, however ignoring such issues comes with a significant level of risk as noted else where.

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

The use of mkiocccentry -W .... is highly discouraged as this may ignore warnings about a problem that may cause your submission to be REJECTED!"

If mkiocccentry encounters an error, the program will exit and the xz compressed tarball will NOT be generated: this is a feature, not a bug!

If you encounter a bug with the mkiocccentry toolkit PLEASE run the bug_report.sh script to help us out here!

While you don't have to install the mkiocccentry toolkit, then in order to run a tool outside the repo directory you will have to specify mkiocccentry(1) command line options to locate required tools such as chkentry(1) (mkiocccentry -C chkentry ..), txzchk(1) (mkiocccentry -T txzchk ..), and fnamchk(1) (mkiocccentry -F fnamchk ..).

The mkiocccentry(1) tool will ignore any filename or directory that begins with a period (.).

The mkiocccentry(1) tool will ignore all files that begin with COPYING, COPYRIGHT, and LICENSE, regardless of filename case.

The mkiocccentry(1) tool will reject any of the following files if they are found directly under the top level directory (topdir):

• README.md
• index.html
• prog
• prog.alt
• prog.alt.o
• prog.orig.c
• filenames matching the [0-9][0-9][0-9][0-9]_*.tar.bz2 file glob

The files listed in the above list are allowed to exist in sub-directories: just NOT directly under the top level directory (topdir).

The mkiocccentry(1) tool evaluates all filenames and directory names in a case independent way: thus FOO, Foo, and foo are considered the SAME file even if your filesystem doesn't do the same.

Having two or more filenames that differ only by case WILL lead to problems because of the previous statement.

When the mkiocccentry(1) tool ignores a file, that file is NOT included as part of your submission.

When the mkiocccentry(1) tool ignores a directory, that directory, and everything under that directory is NOT included as part of your submission.

The mkiocccentry(1) tool will ignore all symlinks.

Creating a directory directly under the top level directory (topdir) that is normally considered to be a regular file, such as a directory named Makefile, prog.alt.c, prog.c, remarks.md, try.alt.sh, or try.sh is NOT a good idea and will likely cause the mkiocccentry(1) tool to abort with a fatal error.

The required Makefile and remarks.md files may NOT be empty. This does not mean just 0 bytes but also that it MUST have content, and content that is relevant (e.g. a Makefile with just comments is NOT allowed). However, with the exception of the Makefile, the mkiocccentry(1) tool will ONLY check file size.

If the optional file try.alt.sh, or try.sh exists and is an empty file, the mkiocccentry(1) tool will abort with a fatal error.

If the mkiocccentry(1) tool encounters anything that isn't a file, isn't a directory, and isn't a symlink (which the tool ignores), the mkiocccentry(1) tool to abort with a fatal error.

See FAQ on "How do I report bugs in an mkiocccentry tool?".

See Rule 11 - Legal rule abuse for warnings about ignoring mkiocccentry(1) tool warnings.

See FAQ on "What is the fnamchk tool?".

See FAQ on "What is mkiocccentry in simple terms?".

See FAQ on "What is the mkiocccentry tool, how do I obtain it and how do I use it".

Jump to: top

# Makefiles

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)`.

## Default compiler flags

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](#optimiser) for details for changing optimiser flags.

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

## Default compiler

**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!

## C standard

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

## Default optimisation level

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

## Compiler warnings

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](#disabling-warnings).

### The `-Weverything` option

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](../faq.html#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](../faq.html#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](../faq.html#forced_warnings)" and the FAQ on "[-Weverything](../faq.html#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.

## Defining macros in the Makefile

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.

## Include files in the Makefile

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

## Magic in the Makefile

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`!

## The clobber rule

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.

# General Guidelines

General Overall Guidelines

These guideline are hints and suggestions, NOT IOCCC rules.

While submissions that violate guideline are allowed, submissions that remain within the IOCCC guidelines are preferred.

You are encouraged to review the following FAQs:

• How to enter the IOCCC

• Obtaining and compiling the most recent mkiocccentry toolkit

• How to upload your submission

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

The official locale of the IOCCC is C.

You are encouraged to examine the winners of previous contests.

Because rules change from year to year, some winning entries of the past may be rejected this year.

What was 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 (such as clang-format(1))
• 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 as 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.

Given two submissions, we are more inclined to pick the submission that does something interesting when it's executed.

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.

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

OUR LIKES AND DISLIKES:

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".

Even if it only has one invocation we still **LIKE** the try scripts.

If you do include a `try.sh` then **PLEASE** remove the `try` rule in the Makefile.

If you do include a `try.alt.sh` then **PLEASE** remove the `try.alt` rule in the Makefile.

If you don't have a prog.alt.c, then **PLEASE** remove 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. Hint!

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.

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.

Using a mass of `goto`s to obfuscate your code has become 'old' and is unlikely to make it through the final rounds, if it even gets that far.

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.

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.

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.

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. :-)

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.).

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.

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

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 8 - Submitting requirements).

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 */

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.

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.

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:

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.

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 both BSD and Linux curses.

If you do `#include ` make **CERTAIN** you link in curses (i.e. `-lcurses`) and not ncurses (i.e. `-lncurses`).

Rule 12 - UTF-8 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.

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

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.

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 4 - Required files, and the 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

Fun😄damental Guidelines

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

At least 2 of several reasons for selecting the prime, 2503, for Rule 2b - Net Size may be found on the "Curios!" about 2503.

The 2503 value for Rule 2b - Net Size is a decimal anagram of the number of the kernel disk pack of one of the judge's BESM-6.

There may or may not be fewer than 2^7+1 reasons why these IOCCC guidelines seem obfuscated.

Excessively "dotty" use of the chksubmit(1) or chkentry(1) commands suggests that either you like silly things, or that you "RTFS" with some level of care.

Some people might question what actually constitutes a valid JSON file, given the state of the "so-called" JSON spec.. :-)

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

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

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.

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.

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. :-) 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 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.

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

There are more than about 1 typos in this very sentence.

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. :-)

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

> 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 or partly 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 (allegedly).

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.

Please try to avoid this specific individual _guideline_, if it at all possible.

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.

We do not recommend submitting [systemd](https://systemd.io) source code to the IOCCC, if nothing else because that code is likely to exceed [Rule 2 - Size](rules.html#rule2-size). This isn't to say that another highly compact and obfuscated replacement of `init(8)` would not be an interesting submission.

The [mkiocccentry toolkit](https://github.com/ioccc-src/mkiocccentry) is not an original work, unless you one of the [several mkiocccentry toolkit contributors](https://github.com/ioccc-src/mkiocccentry/graphs/contributors). in which case it is original! :-)

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 :-)

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

Jump to: top

FOR MORE INFORMATION:

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