The International Obfuscated C Code Contest
Obfuscate: tr.v. -cated, -cating, -cates. 1. a. To render obscure.
b. To darken. 2. To confuse: His emotions obfuscated his
judgement. [LLat. obfuscare, to darken : ob(intensive) +
Lat. fuscare, to darken < fuscus, dark.] -obfuscation n.
How it was started:
The original inspiration of the International Obfuscated C Code
Contest came from the Bourne Shell source and the finger command as
distributed in 4.2BSD. If this is what could result from what some
people claim is reasonable programming practice, then to what depths
might quality sink if people really tried to write poor code?
I put that question to the USENET news groups net.lang.c and
net.unix-wizards in the form of a contest. I selected a form similar
to the contest (Bulwer-Lytton) that asks people to create the worst
opening line to a novel. (that contest in turn was inspired by disgust
over a novel that opened with the line "It was a dark and stormy
night.") The rules were simple: write, in 512 bytes or less, the worst
complete C program.
Thru the contest I have tried to instill two things in people. First
is a disgust for poor coding style. Second was the notion of just how
much utility is lost when a program is written in an unstructured
fashion. Contest winners help do this by what I call satirical
programming. To see why, observe one of the definitions of satire:
Keen or energetic activity of the mind used for the purpose
of exposing and discrediting vice or folly.
The authors of the winning entries placed a great deal of thought into
their programs. These programs in turn exposed and discredited what I
considered to be the programmer's equivalent of "vice or folly".
There were two unexpected benefits that came from the contest winners.
First was an educational value to the programs. To understand these C
programs is to understand subtle points of the C programming language.
The second benefit is the entertainment value, which should become
evident as you read further!
Suggestions on how to understand the winning entries:
You are strongly urged to try to determine what each program will
produce by visual inspection. Often this is an impossible task, but
the difficulty that you encounter should give you more appreciation
for the entry.
If you have the energy to type in the text, or if you have access to
a machine readable version of these programs, you should next consider
some preprocessing such as:
sed -e '/^#.*include/d' program.c | cc -E
This strips away comments and expands the program's macros without
having things such as macros clutter up the output. If the
entry requires or suggests the use of compile line options (such as
-Dindex=strchr) they should be added after the '-E' flag.
The next stage towards understanding is to use a C beautifier or C
indenting program on the source. Be warned that a number of these
entries are so twisted that such tools may abort or become very
confused. You may need to help out by doing some initial formatting
with an editor. You might also try renaming variables and labels to
give more meaningful names.
Now try linting the program. You may be surprised at how little lint
complains about these programs. Pay careful attention to messages
about unused variables, wrong types, pointer conversions, etc. But be
careful, some lints produce incorrect error messages or even abort!
Your lint may detect syntax errors in the source. See the next
paragraph for suggestions on how to deal with this.
When you get to the stage where you are ready to compile the program
examine the compilation comments above each entry. A simple define or
edit may be required due to differing semantics between operating
systems. If you are able to successfully compile the program,
experiment with it by giving it different arguments or input.
You may also use the makefile provided to compile the program.
Keep in mind that C compilers often have bugs, or features which
result the program failing to compile. You may have to do some
syntax changing as we did to get old programs to compile on strict
ANSI C compilers.
Last, read the judges' comments/spoilers on the program. Hints
for `foo.c' are given in `foo.hint'. Often they will contain suggested
arguments or recommended data to use.
If you do gain some understanding of how a program works, go back to
the source and reexamine it using some of the techniques outlined above.
See if you can convince yourself of why the program does what it does.
About the judges:
As of 1988 the contest had two judges: Landon Curt Noll (contest
founder) and Larry Bassel (judge since 1985). Landon works as a
systems programmer for Amdahl Corporation and Larry works as an systems
programmer for Sun Microsystems. In real life, both judges strongly
dislike obfuscated code.
Regarding the source archive:
Each sub-directory contains all the entries for a single year. Often
the file names match one of the last names of the author. Judges'
hints are given in files of the form ``*.hint''. The makefiles
given are set up for a System V based machine. You may need to
tweak this makefile to get everything to compile correctly.
Read the hint files for suggestions. The rules for a given
year are given in the file named ``rules''. The last year
in an archive contains a copy of the rules for the upcoming
Regarding the distribution of sources:
All contest results are in the public domain. We do ask that you observe
the following request:
You may shar these files with others, but please do not prevent them of
doing the same. If some of these files and/or contest entries are
published in printed form, or if you use them in a business or classroom
setting, please let us know. We ask that you drop a line to the
'judges' Email box. As of 1988, it is:
[this could change from year to year, so consult the current rules]
Some final things to remember:
While the idea for the contests has remained the same through the
years, the contest rules and guidelines vary. What was novel one year
may be considered common the next. The categories for awards differ
because they are determined after the judges examine all of the
The judges' hints assume that the program resides in a file with the
same username as the author. Where there is more than one author, the
first named author is used.
Some C compilers are unable to compile some of these programs. The
judges tried to select programs that were widely portable and
compilable, but did not always succeed. As of 1988, only ``K&R''
compilers were used. Due to the timing of the ANSI C standard, ANSI C
issues were not addressed until 1988 at all (and in 1988 there were
just a few comments in the hint files). Often only a simple edit is
needed to get a new C compiler to accept the source file.
The contest rules are posted in early March. The winners are announced
at the Usenet BOF of the Summer Usenix conference. Later they
are posted to the net.
The rules are posted to the following Usenet news groups:
As of 1988, the winners were posted to the following Usenet news groups:
People are strongly encouraged to wait until the new contest rules
have been posted before sending entries. The rules, and sometimes
the contest Email address itself, change from time to time.
The typical start date for a contest is March 15. The typical
end date for a contest is May 20.
Last, PLEASE don't code in the style of these programs (unless you
are submitting a contest entry of course!) It is hoped that you will
gain an understanding that bad style destroys an otherwise correct
program. Real programmers don't write obfuscated programs that other
people have to use!
Landon Curt Noll ([email protected])
Larry Bassel ([email protected])