# Most enigmatic Cody Boone Ferguson Twitter: @xexyl The code for this entry can be found in prog.c ## Judges' comments: ### To use: make ./prog ### Try: # IOCCC obfuscated! echo IOCCC | ./prog # IOCCC obfuscated and de-obfuscated! echo IOCCC | ./prog | ./prog # more input fun ./prog - < try.this.txt 2>/dev/null # more obfuscated fun echo testing test tests | ./recode echo testing test tests | ./recode | ./prog - 2>/dev/null echo testing test tests | ./recode | ./prog - 2>/dev/null | ./recode echo testing test tests | ./recode | ./prog - 2>/dev/null | ./recode | ./prog - 2>/dev/null ### Selected Judges Remarks: This code is an enigma. Try to decode it! There is a good deal of useful documentation that is provided with this entry: * enigma.1 A useful man page for this entry. To render, try: man ./enigma.1 * recode.markdown ([recode.html][]) Some useful information about recode.c * [chocolate-cake.html][] Because most of us could use some *Double-layered Chocolate Fudge Cake*! [recode.html]: recode.html [chocolate-cake.html]: chocolate-cake.html ## Author's comments: My remarks about my Heer 'Army' and Luftwaffe 'Air Force' Enigma simulator for the 27th IOCCC. I want to dedicate this entry in part to the millions of poor souls that were lost during the tragedy that is the Second World War. There are so many people who are unknown, unloved, unclaimed and even unidentified. And who knows what they might have gone on to do if their lives weren't - like it is with all other wars - needlessly taken from them? When one considers it was all because the most powerful man in the world at the time felt he should rule all others it makes it somehow even more tragic: there wasn't even a good reason for the war (is there ever?). If one were to write the name of every person who died in the conflict in a book, it would take so many pages that the book would be so heavy, so long, so expensive to print, and it wouldn't bring anyone back to life nor would it erase the suffering. Nothing will. It's horrible. There really are no words that can truly describe the tragedy but I wish to with my Enigma simulator honour each and every soul who was lost, or suffered, or suffered the loss of friends or loved ones, as well as the world itself, in and from one of the greatest tragedies in the history of mankind. I will have more information on this entry at [https://ioccc.xexyl.net/2020/enigma][] after the winning entries have been published. - [IOCCC: An Enigma?](#ioccc) - [The 1992 Nathan entry](#nathan) - [Usage](#usage) * [An example run](#example) - [The recode.c configurator](#recode) - [A parser subtlety that could cause confusion](#parsersubtlety) - [On the source code layout](#layout) - [Obfuscation](#obfuscation) - [Bugs, limitations, differences from the real Enigma machines and general notes](#bugs) - [Adding Kriegsmarine support](#kriegsmarine) - [Portability](#portability) - [Winning thoughts, dedications and thanks](#winning) - [Resources](#resources) ### IOCCC: An Enigma? Is there a better single word that could describe the code that is IOCCC? I'm sure there are some other words that are equally as valid but I believe 'Enigma' is a great word nonetheless. However it's curious that no winning entry has ever been a simulator of the German Enigma machine. I'm actually quite surprised at that because it's such a brilliant piece of engineering of the time even though it had a critical flaw: that no letter could be itself; this combined with cribs (known or suspected plaintext) - which was down to carelessness, capturing an Enigma and sometimes settings sheet, as well as known messages the Germans started out with - allowed for cracking the code. Combining this entry with two of the Morse code entries is particularly interesting because that's what the Germans did: the party sending the message would first set up the machine, write down the letters and then transmit by Morse code the encoded message. Then those picking up on the message would note the Morse code, translate it to the alphabet and then input it into their Enigma machine configured in the same way and thus have the original message. The [1998 ASCII / Morse code translator][] entry (author Dorssel) prints the Morse code of text; and the [2014 Morse audio transcoder][] (author Vik) can be used to allow for the actual sound of it. With my entry we can complete the loop! ### The 1992 Nathan entry The author (Nathan) of the [1992 Worst Abuse of the Rules][] had this to say: This program is a hello world text encrypter/decrypter. It uses an enigma (I think) style encryption algorithm, where the encryption key character is modified by a value, determined from the previous character. Non-printable characters (those with ASCII values < ' ' or > 0x7e) are passed unaltered, thus any kind of file may be successfully processed, but if the original is printable, the processed file will be too. The input is read from stdin, and the output presented to stdout. The key, a text string, is presented as a command argument. This is optional, and if omitted, the file is self-{de,en}crypted. To specify decryption, a "-" should be given before the key. (Actually encryption and decryption proper inverse operations, so you can use decrypt to scramble and encrypt to descramble, if you're perverse.) But the Enigma algorithm is more complex than that and my entry can encipher and decipher text in the manner of the Enigma machines of the Army (Heer) and Air Force (Luftwaffe); the Kriegsmarine 'Navy' was more complex (see section [Adding Kriegsmarine support](#kriegsmarine) for more information on that. My entry should also work as a non-military grade Enigma as it by default has an empty plugboard (they had none). I explain how to set up the simulator later. Yusuke Endoh asked the judges and me if the real 1992 entry source was available. Simon Cooper didn't have access to it because it was before his time in the IOCCC; but before the other judges could answer Ilya Kurdyukov sent a link to the original 1992 code. However because I do not know the current status of the regulation (though it seems unlikely that it still applies) I am only citing an old version of the IOCCC website: nathan - obfuscation by (trivial) encryption (source not exported) Brian Westley decided to just ``screw the (US) government'' and make the real nathan.c source available. We also note that Nathan Sidwell who lives outside the US made his own source available as well. Anyone who wants to find it I am sure can but I leave that to them. ### Usage If you run the program by itself it'll use the default settings and prompt you for input (end with EOF). At that point it'll process the input string and print the result (after finishing the enciphering). If you run it like: ./prog - It will prompt you for the settings of the machine (example below). The first group is in the order of rotor (number), ring setting (letter) and ring (starting) position (letter). It will do this for the first, second and then third for each. Then it will prompt you for the reflector (number). The first one is reflector B and the second is reflector C. Finally it'll prompt you for the plugboard pairs. After that you can input the string and it'll go from there. #### An example run BTW: There's a much more entertaining (and delicious) challenge or exercise in [recode.markdown][] (involves [chocolate-cake.html][]). These however show the general program as well as how to use the two winning entries of the Morse code that I referred to earlier. $ ./prog - Ring 1: 3 Setting 1: I Position 1: O Ring 2: 1 Setting 2: C Position 2: C Ring 3: 5 Setting 3: C Position 3: Z Reflector: 2 Plugboard pair 1: CO Plugboard pair 2: DE Plugboard pair 3: SN Plugboard pair 4: XY Plugboard pair 5: MS Plugboard pair 6: AT Plugboard pair 7: VF Plugboard pair 8: UR Plugboard pair 9: KL Plugboard pair 10: BJ IOCCC ^D CFDRG Notice that all output except the deciphered/enciphered text is sent to stdout. If I didn't specify the `-` by e.g. `echo IOCCC | ./prog` then it will use the default settings; as the judges suggest this invocation I will not show it here. You can also type it out like this though: $ ./prog IOCCC UUMMX The IOCCC comes from stdin; the UUMMX is stdout. If I compile the 2014 entry and copy it to my local directory as `vik` I might do: $ echo TEST|./prog |./vik | mplayer -demuxer rawaudio - And you would hear Morse code of the Enigma output of 'TEST' (i.e. KCWV). Or perhaps not in this case. I'm not sure: I discovered a bug in that entry (Fedora, CentOS and macOS all affected). For example I can use the syntax provided in the winning remarks along with my entry to write it to a file. Say: $ echo TEST|./prog |./vik > test.raw $ ./vik e < test.raw | ./prog LEST Why did that happen? This shows more details: $ ./vik e < test.raw SCWV However as can be seen the first char is wrong: $ echo TEST|./prog KCWV And piping the original output to my entry itself you get the original input: $ echo TEST | ./prog | ./prog TEST I'm not sure what causes that or if there's a way to fix it but that's the idea anyway. Neither do I know if the original transcoding is correct. What if I use the input string 'LEST' instead? In that case it worked fine: $ echo LEST | ./prog | ./vik > lest.raw $ ./vik e < lest.raw | ./prog LEST But there's one obvious question, right? Let's try the most important one: $ echo IOCCC | ./prog | ./vik > ioccc.raw $ ./vik e < ioccc.raw UUMMX $ ./vik e < ioccc.raw | ./prog IOCCC Phew! The program redeemed itself after the test failure! :) (Technically above it didn't print a newline after the output and this is how it is with some of the things I pasted in the other file and maybe others here too. Notice also how it showed the same output that my program gave it - UUMMX.) What about the other Morse code entry? $ echo IOCCC | ./prog | ./dorssel ..- ..- -- -- -..- ### The recode.c configurator I have included a supplementary program 'recode' ([recode.c][]) that can do a number of things which acts as a kind of wrapper and configurator to my entry. By default it prompts you for the settings (rotors/rings, ring settings, positions, reflector and plugboard pairs), validating input as you go. After that it prompts for text that you can pass to the Enigma simulator; alternatively you can specify a file to read in. It has a randomise mode and it can save/read the settings to/from a file to allow for duplicity (an archaic term for the state of being double but obviously for Enigma it has to also mean deceitful). I recommend using this in a pipeline instead of the simulator itself if you want to change the configuration of the simulator because it offers a much richer experience (and because you can reuse the configuration without having to type it out again every time). It does take the effort to try and prevent invalid input (as above) though maybe I didn't think of everything. My entry was meant to be a simulator only as far as the ciphering goes but I thought this would make it much more interesting: make it more flexible by a wrapper program. For examples using it (and a delicious challenge) see [recode.markdown][]. See also [recode.1][] and [enigma.1][] man pages. ### A parser subtlety that could cause confusion The parser is in a sense rather rudimentary but allows for both interactive and automatic (e.g. via the recode program) input. This does mean that something like the following can happen: $ ./prog - Ring 1: 1AB Setting 1: Position 1: Ring 2: 2CD Setting 2: Position 2: Ring 3: 3 Setting 3: A Position 3: B Reflector: 1A Plugboard pair 1: BC Plugboard pair 2: .. Plugboard pair 3: .. Plugboard pair 4: DE Plugboard pair 5: FG Plugboard pair 6: HI Plugboard pair 7: JK Plugboard pair 8: LM Plugboard pair 9: NO Plugboard pair 10: PQ IOCCC WD SIRDK First of all: what is that WD? I didn't type that but it appeared it seems due to too much input that triggers this (it happens when I hit ctrl-d to end input). If everything is input correctly it would show nothing except the actual Enigma output (maybe at times ^D - I have noticed this at least when I don't fill in all plugboard pairs - which might or might not be possible with the Enigma machines that had plugboards, I really do not know). I don't view this as a bug because the parsing of input is meant to be simple and in fact you can expect the same output from the same input even if there are errors in the input. Anyway as you can see the first '1AB' was the first ring, setting and position. After this it wants Ring 2 and I did similar; for ring 3 I did it each variable by itself. However for the reflector I did 1A and it only expects 1 char! So will that mean that the plugboard pairs are off? Yes it does seem to be so. This is buffering at play I believe but it's useful for allowing the recode program to easily configure the Enigma machine. In the other markdown file I give a hint as to how this could be fixed but the caveat is it would necessitate a need for rewriting recode.c. There's another thing to be aware of and that's the way the ranges are enforced. There are five rotors which obviously are in C 0-4 but in 'natural language' 1-5 (they were labelled I, II, III, IV and V though). There are two reflectors and the same applies: in C 0-1 but in human it's 1-2 (technically these were reflectors B and C in the Enigma machine which I display by name in recode.c just like with the rotors). For more information see [recode.markdown][] or [recode.html][]. BTW: If you need a reminder to go to the gym just do your Enigma ABCs and it should help you remember (though not at this time in our world it might help you later on?). This might become even more useful after the challenge is accepted, accomplished and made use of! :) $ echo ABC | ./prog GYM $ echo ABC | ./prog | ./prog ABC ### On the source code layout The layout of recode.c isn't too significant (but see below) but I want to mention prog.c especially: I did think of a more artistic layout but reflecting on it I think that the blankness of it can be another way of honouring all the many people who died in - like all other wars - what was a useless conflict because - again like all other wars - a powerful human being decided that they should rule others. The blankness is a reflection of the millions of people who died unknown, uncared for, unloved, without ever being acknowledged in their far too short lives, some even unidentified; it's the presence of absence. I find it somehow fitting that the day I have finished this is 30 April 2020, which was exactly 75 years from the day that the person responsible for the Second World War, Adolf Hitler, ended his life. (I have since then made a bug fix but I consider this inconsequential). As for recode.c though I really love this: /\ / This is *not* what you think: ...because it seems so wrong in C code at first glance. ### Obfuscation It's an Enigma simulator! Isn't that proof that it's obfuscated itself?! :) Silliness aside [obfuscation.txt][] has some of the ways I think this entry is obfuscated. To decipher try: ./recode -Robfuscation.key -fobfuscation.txt | ./prog - 2>/dev/null > obfuscation.markdown To encipher: ./recode -Robfuscation.key -fobfuscation.markdown | ./prog - 2>/dev/null > obfuscation.txt The file [obfuscation.key][] is the key to decipher/encipher [obfuscation.txt][]. For the lazy [obfuscation.markdown][] has the deciphered version. I am afraid I'm not so inclined to do that for the cake recipe: the idea there is to make it a fun exercise that when solved unlocks a wonderful double-layered chocolate fudge cake recipe. But given that my 'Don't tread on me award' entry also has the recipe, not enciphered, one might just go there instead. Still it's a fun way to explore this entry. As for the [obfuscation.key][] file if you observe the contents you'll find the word OBFUSCATION: 2OB5FU1SC2ATIONCDEFGHJKLMPQRYZ OB FU SC ATION ### Bugs, limitations, differences from the real Enigma machines and general notes Ways that my simulator differs from the real thing. Not all of these are bugs and some could be considered features; others are just differences. * Does not check for using the same rotor more than once. The real Enigma machines this would not work because they had to choose from a set of rotors and there were no duplicates. This shouldn't be a problem here however except that it won't be a possible configuration of the real Enigma machine. The `recode` program *however does validate this* (except when reading in settings via the `-R` option which I explain in the [recode.markdown][] / [recode.html][] file). * The way the plugboard - for the machines that had them - is if you connect A to B then no other letters can connect to A or B. Earlier there wasn't proper detection and this caused invalid output. Why only 10 pairs when the German alphabet has 26 letters? It's theoretically possible to have 13 but the Enigma typically had 10 pairs. It's easy enough to update my entry to do this once you de-obfuscate it but I'm not helping there (and in fact I don't think I could any more even if I wanted to - at least not easily!). * It does not obviously have the clicking sound associated with the moving parts of the real machine. I think it would be annoying to have this as it processes each letter. However with the Morse code audio transcoder entry you can at least have the output played in Morse code as the Germans did; I think that is really cool (I think it's really cool even if there's a bug in it)! That it pays homage to that (and another) entry is a bonus here. * The simulator doesn't report errors but tries to gracefully deal with them: if one chooses a value out of range it moves it into the range (if > the max then the max, if < min then the min). If invalid chars are input for the plugboard pairs then it's not added to the plugboard etc. * The real Enigma machines had only letters: no digits, no punctuation. My entry prints `!isalpha()` chars untouched. The Germans had different systems for punctuations, numbers etc. (one of the documents I link to has these details). * My entry preserves case of letters. The real Enigma had only one case as it was just a keyboard and they didn't need case-sensitivity. * Later models introduced a fourth rotor. The Kriegsmarine procedures were also far more complex than the Heer and Luftwaffe. I find this particularly interesting because it was Hermann Goering who was the Supreme Commander of the Luftwaffe and the intelligence agency Forschungsamt (FA) was his; it provided Hitler with invaluable information so one would think the Luftwaffe system would be more complex! See [adding Kriegsmarine support](#kriegsmarine) for more information. * I only include reflectors B and C. To have included more would have used too many bytes (each reflector and rotor is 26 characters!) and I could not include them for this reason; I include 5 rotors and 2 reflectors and that's a total of (`5 * 26 + 2 * 26` == 182 chars) but there are as you can see many other strings too. It is however feasible to update the rotor and reflector options and you could even make your own versions if you wanted to. Doing this is an exercise to the reader however (the idea of trying this makes me cringe because of how I obfuscated things)! There's one thing I'm unsure of in my implementation. Because of all the possible settings I used some other simulators as a starting point and more importantly to validate that my code worked right. These other simulators had bugs though and so one thing looked odd and I don't know if it is right or not. An online simulator appears correct but I don't see the option to change that part. If you notice in the Q() function where I set up specific pointers and other data there is a call to the macro 'q' but I only refer to the first ring setting (of which I no longer even know which element it is! Or was it position? I don't even know that now for certain). That's because the Python code did that (though very differently of course) but I also had to bug-fix the Python in another part. I do not know if this is correct but as I show somewhere my entry does match the online simulator so I presume it is correct. It might be that the Germans tended to have it at A but again I do not know. ### Adding Kriegsmarine support I said that one could add the additional rotors used by the Kriegsmarine but looking at this again it appears that it's not as simple as adding the rotors to the code and updating the references to the array etc. This is because rotors VI, VII and VII had two notches and this changes the way things work. The Kriegsmarine M4 also had an additional change which would complicate matters more. I looked into trying to do this for another version after winning but some of the obfuscation techniques dramatically complicates even attempting this so I haven't even tried. It is feasible that you could add your own rotors however but I leave this as an exercise to the reader. I will say that the five rotors I, II, III, IV and V can be found in the source code verbatim and you could confirm it by looking them up; however much more would have to be done in order to get them to work, than to just put the strings in. Again the way I set things up complicates this greatly. ### Portability I have tested it under macOS, Fedora and CentOS Linux and all seems okay under these platforms. When I was in the final stages of obfuscation I made an error somehow (or else maybe the vim formatting caused a problem - I have seen this before) and I had to go back and undo some of the obfuscation to figure out what was going on (I had a hunch where and I noted it in the obfuscation I believe) but this made me wonder on if I made any other mistakes since it appeared to be working before that. I don't know if EBCDIC will work but I think it might because it appears that the alphabet is a contiguous sequence. I don't believe endianness matters but I no longer remember what I've done in full and I've no way to test it. If there are any inconsistencies in input/output one also has to be certain that input was put in correctly because as soon as a letter differs the entire sequence will be off. The results of my test script (not included here but I will probably have it on my website) are the same under Fedora, CentOS and macOS. ### Resources #### A note on Wehrmacht, its branches and other agencies and their Enigma machines The Wehrmacht was the defence force in full which included the Heer (army), Luftwaffe (air force) and Kriegsmarine (navy). As I said the Kriegsmarine procedures were more complex: they had additional rotors and one of their models had a different reflector system (thin instead of thick - some of the documents below discuss this). The military intelligence agency, the Abwehr, had their own system too. Non-military grade Enigma machines didn't have the plugboard. And so on. However one of the documents below distinguishes the Kriegsmarine from the Wehrmacht; when you see this you can read it as the Kriegsmarine versus the rest of the Wehrmacht rather than the Wehrmacht and the Kriegsmarine. [Enigma Message Procedures][]: The Heer, Luftwaffe and Kriegsmarine Enigma procedures. [Interactive Enigma Machine][]: Shows wiring of the Enigma machine based on the settings and input (unfortunately uses Flash). This was what I used to verify output (as best I could find). To set this up to use the default settings in my simulator make sure to use Rotors I, II and III and the starting positions at B, B and C respectively (as in left, centre, right). To do this click the button at the top right that looks like: o-> <-o And then select the rotors. Next click the button again at the left of the rotor selection 'screen'. Then use the arrow buttons above each rotor to set from left to right B, B and C. After this you can type your message in the input field and watch the wiring of the rotors change and show the path it takes. If I made use of the Enigma code in the recode.c program I would maybe consider doing something like that but I intentionally do not use it there: as cool as it might be it would give a better idea of how the algorithm works (unless I were to make it just as obscure as I do in prog.c but that's also a bit risky). I don't believe I could do it now anyway. I'm uncertain how to change the rotor settings itself in this simulator to determine if the ring settings being different from ABC would cause a problem (this is the thing I mention I'm uncertain about in the bugs section). Going back to the configuring the online simulator try inputting the text IOCCC and you'll see the resulting text being: UUMMX Which is exactly as my entry shows! So at least for the starting position BBC and an empty plugboard and the other defaults in my entry (which appear to be consistent with the online simulator too) I can verify it works. [Tony Sale's pages on The Enigma cipher machine][]: This is many pages on the Enigma machine by the late Tony Sale who was the original curator of the Bletchley Park Museum. There is a wealth of information in these pages on how the Enigma worked and this includes some challenges to do to see how well you understand the algorithm. Fun fact: the abbreviation for the Signals Intelligence is that of our wonderful SIGINT. [Technical Details of the Enigma Machine][]: 'This page provides the technical details of the Wehrmacht and Luftwaffe Enigma, and the Kriegsmarine Enigma M4.' It is quite a nice document with diagrams, pictures of actual Enigma rotors and other parts and more. [The German Enigma Cipher Machine][]: The history of Enigma, its development, use during the war and cryptanalysis. [How the Enigma was Set Up and Operated][]: This is the third page in a document series on the Enigma that explains the initial set up of the Enigma machine for each day. This is a good introduction to the set up. There is one thing that I was not aware of (or if I was it did not register): the operator would after setting the positions also put in three letters before the actual message (there were several parts to each message but this was another layer of complexity). I did not implement this in the recode program but what would it be anyway? If the user here wants to they could add three letters to their message beginning. I don't think there need be any addition to the recode program - at least not that. [1992 Worst Abuse of the Rules]: http://www.ioccc.org/1992/nathan.hint [1998 ASCII / Morse code translator]: http://www.ioccc.org/1998/dorssel.hint [2014 Morse audio transcoder]: http://www.ioccc.org/2014/vik/hint.text [Enigma Message Procedures]: http://users.telenet.be/d.rijmenants/en/enigmaproc.htm [Interactive Enigma Machine]: http://enigmaco.de/enigma/enigma.html [enigma.1]: enigma.1 [recode.1]: recode.1 [recode.markdown]: recode.markdown [chocolate-cake.html]: chocolate-cake.html [recode.c]: recode.c [obfuscation.markdown]: obfuscation.markdown [obfuscation.txt]: obfuscation.txt [obfuscation.key]: obfuscation.key [Tony Sale's pages on The Enigma cipher machine]: http://www.codesandciphers.co.uk/enigma/index.htm [Technical Details of the Enigma Machine]: http://users.telenet.be/d.rijmenants/en/enigmatech.htm [The German Enigma Cipher Machine]: http://users.telenet.be/d.rijmenants/en/enigma.htm [How the Enigma was Set Up and Operated]: http://www.ellsbury.com/enigma3.htm [https://ioccc.xexyl.net/2020/enigma]: https://ioccc.xexyl.net/2020/enigma ## Author's dedication: ### Winning thoughts, dedications and thanks Once again I'm honoured to win along with some of the other veterans who I have a huge amount of respect for; there's Don Yang, Dave Burton, Yusuke Endoh and Edward Giles all of whom are amazing programmers. I'm especially proud to win besides Dave Burton and Yusuke Endoh but Don Yang and Edward Giles too. But there are some others who I don't remember like Ilya Kurdyukov (who's offered a lot of interesting thoughts on Snake and who plans to have his own version after the entries are published), Nathan Otterness, Nicholas Carlini (whose entry is an absolutely brilliant abuse of printf - and I totally agree with the judges comments that your entry is so novel that it'll be worth special mention in the future Best of IOCCC list!) and tsoj too. Well done and congratulations to all of you again! Thank you for submitting the wonderful entries and I'm honoured to have my two entries beside yours! To Yusuke Endoh, Dave Burton and Nicholas Carlini I very much appreciate your wonderful comments on this program. Your comments Yusuke on this entry specially and also Dave's comments on my double-win are both extremely special to me - it truly takes my pride to another level! I would like to dedicate this entry to my wonderful mum Dianne Ferguson whom I love with all my heart and soul <3 Thank you for being the best mother imaginable! You've been there for me in the darkest of my hours without question without complaint and with all the love, compassion, empathy, sympathy and devotion anyone could ever ask for. Bless you again and again and again dear mother. I also want to dedicate this to Vicky Wilmore who has given me a lot of love the past few years and who was there for me in some very dark times. Thank you Vicky sweetheart from the bottom of my heart and soul. Although we have parted ways, whether for now or for ever, you will always always always have a place in my heart and soul xxx Once more besides those two I want to also dedicate this to all the many millions of people whose lives were taken from them in what amounts to - like all other wars - a horrible, pointless conflict and an utter waste of life, something precious that is far too often taken for granted. As I noted before I found it fitting that I finished this on 30 April 2020 - exactly 75 years after Adolf Hitler, he who started the Second World War, ended his life. The blankness of the layout is another way of honouring the poor souls who lost their lives, many of whom we will never know even their name let alone what they might have accomplished if it wasn't for their tragic and pointless deaths. Always remember that life is precious and never ever take it for granted! Believe in yourselves. Everyone. Do not let anyone ruin that for you ever. I would like to thank my dear friend Martijn Schoemaker for encouraging me in my programming over the years and who I owe a great deal to. Thank you for believing in me and my programming abilities (and other abilities) even when I couldn't believe in myself! Very much appreciated especially coming from an amazing programmer who I have huge admiration for! Finally I want to thank Leo Broukhis, Simon Cooper and Landon Curt Noll for continuing to hold the contest after all these years - and for having selected my entries. It's a huge honour; thank you! I also happen to love your comments as well as the award titles. And yes indeed 'most of us could use *[Double-layered Chocolate Fudge Cake][]*!' [Double-layered Chocolate Fudge Cake]: chocolate-cake.html ----------------------------------------------------------------------------------------------------- (c) Copyright 1984-2020, [Leo Broukhis, Simon Cooper, Landon Curt Noll][judges] - All rights reserved This work is licensed under a [Creative Commons Attribution-ShareAlike 3.0 Unported License][cc]. [judges]: http://www.ioccc.org/judges.html [cc]: http://creativecommons.org/licenses/by-sa/3.0/ -----------------------------------------------------------------------------------------------------