No, quite often it is slower. Mostly it just depends on the architecture and algorithms you have chosen, but compilers are designed to be much better at optimizing typical code than at optimizing weird code.

Interesting. Didn't expect that. Thanks for the fast reply!

It depends on how you obfuscate it. For example, if you replace a bunch of stuff with defines, then it would be transparent to the compile and run identically. If you replace some things with a bunch of function calls, then it would likely end up slightly slower because of the function-calling overhead.

Well yes, but actually no

The changing of the CASE statements bisquit did at the beginning was faster i think

No it is not.

I absolutely love Bisqwit's replies. They're always so simple, yet they convey more meaning than even some books do. I like your videos Bisqwit, keep ut the amazing work :D

I was high when I watched this XD

+Safdar Mahmood: kzfaq.info/get/bejne/iqmqrNNlqd_ddoU.html

god spoke

I had to replay that bit a couple of times because I couldn't believe how brilliant that one sentence was. It's like the English language equivalent what he's talking about doing in C.

C really isn't as good as everyone implies

@@okie9025 No, C is the only good programming language. Just because it's too hard for you doesn't mean it's bad.

@@hexagonist23 C is probably the simplest, most rudimentary language that exists, and it shows.

@@okie9025 and thats a good thing.

@@hexagonist23 C is great, but isn't for everything. The simplicity can be a good thing when your software need to run very fast or use less space in memory, but when the development is more important than speed they are executed, C can be a headache, and time means money.

I just found a random dude using the same profile picture as me

you're right and i've seen the same too. Something like this? www.makeuseof.com/tag/basic-programming-principles/ i dabble in C# and generally what you describe comes in the form of everything being split into classes and functions neatly, always calling functions to pass values rather than tunneling through multiple classes with reference values ( i've heard this referred to as "agile coding"...?)

I used to code a fair amount for uni, and one of my favorite pastimes would be to align to readability for 99% of the code, then put one absurdly obfuscated line somewhere with no comment. Something like switch ( data>>(2*(3-i))&3 ) {

@@denny141196 good luck when you end up using a module from one project with an obfuscated line in something else three months later and you have to debug it

I think functional programming is what you're talking about

@@thezipcreatoryou do indeed need to write even a minimal comment just to have a bare understanding of what a line does. But I personally do not use a lot of comments as I tend to remember every bit of code I write

do u realize that in business development u've got software/libraries to code obfuscation? xd doing it manually is just waste of time

​@@miszczklasykuw3025I don't think you're understanding the point of code obfuscation in C. When you develop a program in C, compile it, and send the compiled version to customers, they don't get access to the source code. Obfuscating the C source code is a terrible idea from a business perspective, whether you automate it or not: It doesn't make your program harder to reverse-engineer, it just makes it harder for developers to modify later. The sort of obfuscation described in this video is actually done for fun: To show off knowledge of the tiny details of C, or to see just how unreadable code can get while still maintaining functionality. If you automate that, you take the fun out of it.

@@edwardfanboy u dont have obfuscated code in system version control (otherwise u've got it in different branch) - the point is u run obfuscation script after release, so only customers don't have clear code.

@@miszczklasykuw3025 But you don't need to obfuscate your source code. The compiler does that for you simply by converting it into machine code.

@@edwardfanboy not if u write in example in javascript

In a sense, that's exactly what a raw binary code file is! Far as the computer is concerned, nothing ends It just follows instructions in the order given, and jumps to a different place when told to. No lines, just memory addresses, and memory is for all intents and purposes, 1 dimensional. Just my 2 cents

Actually, you can write almost any C program in 2 lines

The compiler definitely does not care about newline in code or white-space. Indeed you can. There is actually a term for this process that java programmers use, a process known as 'gulping'. It places an entire program on 1 line and uniquely identifies all variables, giving them sequential letter renaming (a, b, c... aa, ab, ac...) and combines them on a line (or at a configured maximum of 32768/65536/16384/... characters for parser capability differences, some text editors will crash when a file contains a line above certain limits aswell) Gulping is applicable to C and C++ but no one uses it because we'd like to avoid going insane. It is however a very useful process for creating code art, as gulping was designed to minimize the number of bytes in a php or javascript that are transmitted over the internet for running on the client machine, gulp contains a c and c++ parser and can minimize code to an extent like he does here manually. Its primary focus is to minimize the number of characters the program contains. For some builds of gulp, it will even rework program logic to find a better (and smaller, in bytes) solution.

You can do that easily in any language, like JavaScript

@@okie9025 But for example not in python, where indentation is used instead of braces etc.

I still don't quite understand it..can someone explain?

There was numb3rs before in array. So think of a number say 60. We add this to each number to the array 44440000000 etc. Then convert it to the ascii. Then it looks like that. Now switch the number we used before with real offsett

@Meowtresting For a computer, all it sees is numbers. Literally everything is numbers. Even characters, they are numbers. Text is numbers. Colors are numbers. Graphics is numbers. Audio is numbers. Video is numbers. Programs are numbers. Sequences upon sequences of numbers. Because all the computer has, is numbers. That’s why they are called _computers._ They compute. To process characters and text and strings, there are agreed conventions that describe how numbers are interpreted as characters. One of these is the ASCII, and it is almost universally used. In ASCII, numbers from 32 to 63 map to symbols !"#$%&'()*+,-./0123456789:;? respectively, numbers from 64 to 95 map to @ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_ respectively, and numbers 96 to 126 map to `abcdefghijklmnopqrstuvwxyz{|}~ respectively. In other words, 65 is the same as 'A', 66 is the same as 'B', 67 is the same as 'C', and so on. So, when in C you have a string of characters such as "HELLO", what you _really_ have is an array of numbers: 72, 45, 76, 76, 79. It’s just _represented_ as characters for your convenience. But it really is just a sequence of numbers. One implication of this is that when you _need_ an array of small numbers, you _can_ represent them as a string of characters. It’s not a good idea for real-world programming, but obfuscated programming is specifically about _not_ good ideas, so it’s perfect there! If the numbers you have don’t map to useful characters, you can use mathematical formula to change them. For example, if your array of numbers is 0,0,0,4,4,4,8,8,8, those numbers don’t really map into printable characters. But if you add 67 to those, you get 67,67,67,71,71,71,75,75,75, which maps to "CCCGGGKKK". Then you can just subtract 67, when reading the string of characters (which, again, is really just an array of numbers), to get the original value.

check this out fro those want graphical explanation: kzfaq.info/get/bejne/adZpermKkr2vcqs.html

@@Bisqwit is there any chance u can help to add my video link to your video description? It is java version

Its because he has a finnish accent

ChrisMC wait, you are everywhere :o

three

In C, you can compress the entirety of your code on a single line (except for preprocessor instructions), by just not using newline characters.

@@LoCloud7 Now imma try coding in c now.

Uhh, i really shouldn't be reading comments before ending to watch the video

+TazeTSchnitzel Yeah, I know. I was trying to figure out a way to incorporate them into my video, but I couldn't. Well, you see a fair share of macros in the IOCCC showcases that flash near the beginning of the video :)

+TazeTSchnitzel I personally prefer no macros to see how far the language allows you to go without them

I think that preprocessor macros are considered kind of a cheap trick to "obfuscate" the code, requiring little to no talent to use. The only "legitimate" use of them is to use them to shape your code into some kind of ascii art, or to shorten the code, if that's a goal.

I did use Chrono Cross in some of my videos, but not in this one! Thanks for commenting. Although there is a little part from Chrono Cross at the end, because that is the music from the video in the lower-right corner, a Chrono Cross TAS: kzfaq.info?annotation_id=annotation_3207288859&feature=iv&list=PL95656014200AC3E7&src_vid=rwOI1biZeD8

It was so tiny and my youtube quality was so low i didn't see it in that corner, i apologise jajajajajaja Oh good! Im plaing it for the thrid time!

Yes, you are right in that this video covered the topic way too narrowly.

Perhaps a second video on the topic is in order? ;) ;)

One of the most entertaining of those was a similar contest, where you had to generate the largest error message with as little code as possible.

CrimSun Compiler error or any error?

Nice. I do believe you need at least one semicolon though!

No. For some short programs you won't need any semicolon at all. However not every program can be written like that.

Please give an example of a program that does not need a single semicolon and that does something other than just terminate. Even if your program is just a single return statement, it still needs a semicolon. (Note: Using compiler -D options to create a #define that contains a semicolonis not a valid solution.)

Well, well. I will surely give you an example. Say you want one small program that just says "Hello there! I'm Bisqwit", your program's main logic should look like this: if(printf("Hello there! I am BIsquit")){} This is taking advantage of 'if' structure where no semicolon is required. You can extend this to a more bigger program by taking user input and based on that print something! I've decided to make a video on it and put it in KZfaq.

I stand corrected! I am ashamed of myself that I could not figure that one out (using for/if/while which do not need a semicolon, but another statement which can be a compound block).

Art is not supposed to be practical... or make sense at all.

True. In this video I tried to focus on techniques that don’t rely on the compiler accepting non-compliant code (implicit ‘int’ return type was permitted in C89, but not since C99).

+Delighter Thanks!

Yes -- it is possible to do that infinitely --, although it would not discover new algorithms. It would just keep applying the same formulas. For example, make an interpreter. Then convert that interpreter into its own language, and add an interpreter. Then convert that interpreter into its own language, and add an interpreter.

It's called Joe's Own editor, I think. It can be modified.

Everything can be modified... Did you mean, it's open source? I was also curious to know more about his editor. It seems to be supporting UTF-8 and Unicode! Wow! I wasn't expecting this, and there's so many features. Unfortunately, you can't use it on windows unless you are using Cygwin.

@@Reth_Hard What so special about it supporting unicode? Is it because it's in the terminal? If so, then vim supports unicode, emacs supports unicode, nano supports unicode...

It is

As someone who knows C and uses it as a main programming language, this looks like black magic

People sometimes criticize Python for its mandated use of whitespace, but it prevents you from the sort of descent into madness demonstrated in this video. If you want good, readable C code, clear usage of whitespace might as well be mandatory.

I can surprisingly understand it, and I wouldn't even say I know C. I understand the basics, but not much more than that. You just need to know how strings work in C, and know about ASCII. If you knew that, it's not hard to piece it together.

@@BlazertronGames I mean, when he does this, I understand it completely. Black magic is the ability to come up with this

Cheating implies rules. Which rules are you contesting?

The international contest mentioned 0:10, but I was just joking :D No joka tapauksessa, nää videot on mukavaa katseltavaa.

You can find the rules of the (as of now) most recent IOCCC at www.ioccc.org/2015/rules.txt . To be honest I cannot entirely ascertain how submissions that require supplied include files are treated as for the size limits.

Of course, this will totally throw portability out of windows. In IOCCC, you would have to explain this in the submission and justify why you are doing it.

The comma operator outside e.g. function parameters is really quite a nasty hack / language quirk. If I had to label the worst design choice in C language, it would be the comma operator. I mean, the semicolon should do the same, and the for-statement should be redesigned.

Yeah, when we ran into the comma operator as a bug it blew our minds. i couldn't believe it was an intentional part of the language. GOTO gets a bad reputation, but the comma operator is the nastiest hack by far. great for obfuscation though, haha

I just do that.

@@Bisqwit oh okay haha

Default VGA font, hardware-supplied (or in this case, emulator supplied because I am running DOSBox in the video).

Bisqwit Thanks for your reply. I'll go and find the default font or a replica for my editors.

It is basically a substring that starts from the middle of the other string. But being a pointer, it owns no data nor clones it, just points to it. In Python, you would do that (without the sharing) with e.g.: s = "something" s2 = s[4:] print s, s2 The equivalent in C is: const char* s = "something"; const char* s2 = s + 4; printf("%s %s ", s, s2); Here, s2 is the "custom" character pointer (as opposed to one defined by the compiler by assigning it the start-address of a string constant). But the equivalescene in Python breaks if you try to make changes to s2. In Python, if you do that, changes to s2 will not be reflected in s. In C, changing the data pointed by s2 will also change data pointed by s, as long as they point to (sections of) the same data anyway. In this video though, data is not changed. And in fact, changing the contents of a string _constant_ would be undefined behavior.

Bisqwit thank you so much for the assistance. That all makes sense. I just have two questions now: 1. Wouldn't printing w+25 output "Good %s! ..." but also the other greetings too? How do those get removed? 2. I see that w[tm_hour] would get either "A" , "#", "+", "/", or "9". What do those pointers actually do? (And also, what is the "\0" doing in between the greetings?)

1) That’s what the \0 is there for. Printf and other C language string functions do not really know the length of any string, because a “string” is really just a pointer to (numeric address of) the start of some memory area that ostensibly contains characters. Instead, they just read forward as long as needed until they hit the first \0, and then they stop. Anything in the memory that follows \0 will be completely ignored, because it might just as well e.g. be some completely other type of data. 2) I think my answer also covers this question.

Bisqwit oh I get it now! Thanks for explaining all that

Yeah, C strings are just char pointers. Very, very, dumb and rudimentary.

It was not very legible at small sizes.

+Fred M This tool: bisqwit.iki.fi/source/inputter.html

+Bisqwit Thank you :)

+Dawid Pietrzak (ca1ek) Without looking it up, I presume it uses self-modifying code? I.e. it "mov"s into its own instruction space, replacing instructions to be executed. Otherwise I cannot see how it could possibly do e.g. unbounded loops.

I stand corrected. Looks like it does not do self-modifying code. Impressive. For an explanation, see this: kzfaq.info/get/bejne/iJ11eNKIzJypm58.html To be fair, this hardly qualifies as a statement of the condition of x86. With look-up tables and stuff (that movfuscator uses), the same could be done on ARM thumb code. Maybe even on 6502...

If you take the character value '4' and subtract the character value '0' from it, the result is the integer value 4. This holds true for all decimal digits. It similarly also works for alphabet, and gives the character distance (e.g. 'E' - 'C' = 2).

Oh yes I understand now, I had just forgotten that there is no more table of numbers! Thank you for your response

+juliocrivera Here are the clues necessary to piece the idea together: char*s = "abcdefgh"; printf(s+5); // Prints "fgh" char t = 5; char*s = "abcdefgh"; printf(s+t); // Still prints "fgh" char*s= "\5bcdefgh"; char t = *s; // Gets '\5', which is the same as 5 //char t = s[0]; // Same as above printf(s+t); // Still prints "fgh", because t is 5 char*s = "abcdefgh\0ijklmn"; printf(s+5); // Still prints "fgh", because printf stops reading string when it encounters \0.

+Bisqwit Oh Wow, thank you very much! that was super easy to understand haha

+Bisqwit Btw if you are so kind again how do the custom characters work? i mean the part Where it says AAA### I understood the numbers but not that

+juliocrivera Maybe this helps? upload.wikimedia.org/wikipedia/commons/thumb/1/1b/ASCII-Table-wide.svg/2000px-ASCII-Table-wide.svg.png 'A' is 65, '#' is 35, and so on.

+Bisqwit ah haha duh, thanks

In C, string is just a array of characters, so you can access it as a char array. For the " - '0' " part, it is used as an offset because character '4' is not equal to integer 4

@@hawariam ah i get it now, thanks

I'm neither planning to make nor planning to not make. Though in retrospect I understand there's plenty I forgot to cover in this video. In this video I only dealt with code compaction, which is just one of obfuscation techniques.

I was a bus driver. In the regional traffic 3.30 am is when the day changes. Before that it's good night, after that it's good morning.

+Mike Scott I can't honestly remember whether I have submitted to IOCCC or not. If I have, it was not recently. And even if I did submit, I did not win.

no, it gets more sinister later

It works because the digits 0…7 in ASCII are 00110000…00110111. In other words, only the lowest three bits matter. And I only needed to deal with numbers 0…4. Had this not been the case, I could have chosen a different representation for my numbers.

“to use a graphic libraries to make windows simple […], and not so much dedicated to the programming of games” Sorry, but I am having trouble parsing your comment. I think you are asking about a graphics library for making GUI applications. I think you might like wxWidgets. www.wxwidgets.org/downloads/ If you need a graphics rendering library, try SDL libsdl.org/ or SFML www.sfml-dev.org/ .

That’s true, and you could also remove the “return” statement, but doing these generates eyesore compiler warnings.

“not trying to be rude”, says the person with “FUCK OFF” as their name. You can’t eat money, nor can you keep any money with you when you die. You only need a certain amount in your life. People who see money (or power) as the ultimate pursuit in life are the cause of most of the world’s problems. That being said, there are more recent comments by me as for my employment status that you haven’t probably read.

thanks for responding, lots of respect to you, you're an inspiration

I kept thinking whether saying “you can’t eat money” was a mistake from my part since I lost my train of thought behind that particular sentence, but now I found this. i.imgur.com/lF3kUe8.png

What would this artistic code be like in your opinion? How do you suppose the compiler/parser would define art?

will i asked a stupid question? to you about building a software that can just a make normal code to a obfuscated code , similar to bison, "you're the best programmer i ever seen you're skills are awesome".

You asked about a program that generates _artistic code_. Art is subjective. It depends on human psychology. I hoped that you would clarify your views on what would qualify as “artistic code” when generated by a computer program. If you meant to say “obfuscated code”, that is much easier. For Javascript language, you can find dozens of Javascript obfuscaters in the Internet by a simple search. For C++ it’s a bit more difficult than for Javascript, mostly because there is no eval(), but some similar approaches work, such as reducing all user-defined symbols to their minimal length, removing all whitespace, removing parenthesis were not necessary, and finally using #defines to compress identical subsequences of parser tokens.