26
votes

Everyone knows of the old adage that the best programmers can be orders of magnitude better than the average. I've personally seen good code and programmers, but never something so absurd. So the questions is, what is the most impressive feat of programming you ever witnessed or heard of?

You can define impressive by:

  1. The scope of the task at hand e.g. John single handedly developed the framework for his company, a work comparable in scope to what the other 200 employed were doing combined.

  2. Speed e.g. Stu programmed an entire real time multi-tasking app OS on an weekened including its own C compiler and shell command line tools

  3. Complexity e.g. Jane rearchitected our entire 10 millon LOC app to work in a cluster of servers. And she did it in an afternoon.

  4. Quality e.g. Charles's code had a rate of defects per LOC 100 times lesser than the company average. Furthermore he code was clean and understandable by all.

Obviously, the more of these characteristics combined, and the more extreme each of them, the more impressive is the feat.

So, let me have it. What's the most absurd feat you can recount? Please provide as much detail as possible and try to avoid urban legends or exaggerations. Post only what you can actually vouch for.

Bonus questions:

  1. Was the herculean task a one-of, or did the individual regularly amazed people?
  2. How do you explain such impressive performance?
  3. How was the programmer recognized for such awesome work?
6
  • 5
    I once saw a guy bench press 2 KLocs... (sorry, could not resist)
    – DevSolo
    Nov 12, 2010 at 12:53
  • "Speed e.g. Stu programmed an entire real time multi-tasking app OS on an weekened including its own C compiler and shell command line tools" - Erm..you might want to keep it real.
    – Jas
    Nov 12, 2010 at 13:05
  • Based on the answers so far, some people are easily impressed. I've done a bunch of the 'feats' listed here.
    – Steven Evers
    Nov 12, 2010 at 16:57
  • Hm, I was thinking about almost duplicating this one exactly yesterday ;) Many thanks for asking it!
    – mlvljr
    Nov 13, 2010 at 15:03
  • @DevSolo Hey, if this is punched cards or rolled paper we're talking about, that could be a physical feat.
    – Mark C
    Nov 19, 2010 at 5:11

21 Answers 21

Reset to default
39
votes

While I can't officially vouch for it, I have always been impressed with Chris Sawyer developing Rollercoaster Tycoon almost entirely in assembly language.

2
  • 4
    that's crazy!!!
    – nanda
    Nov 12, 2010 at 14:08
  • 1
    -.- +27 for something we've discussed at least twice on this site already.
    – Mark C
    Nov 19, 2010 at 5:38
31
votes

Fast inverse square root. How someone can come up with something like that is completely beyond me.

16
  • 1
    Hum... That might be (extremely) clever, but it's like 6 lines of code and only a tiny building block of a much larger app. I don't understand how it is extreme in any of the dimensions I listed.
    – David Reis
    Nov 12, 2010 at 19:59
  • 11
    David Reis, this is most likely optimizing a crucial bottleneck in an ingenious way. Do not confuse the number of lines with the amount of work necessary to write them.
    – user1249
    Nov 12, 2010 at 21:22
  • 2
    @David, I don't know why you insist that programming feats must have been done in a very short amount of time to be a feat?
    – user1249
    Nov 13, 2010 at 7:12
  • 2
    Yep, which is more of a feat, 600 lines of code written in one day vs. 6 lines of code crafted for 100 days, if both do the equivalent thing? I think it's the latter, but your mileage may vary.
    – Joonas Pulakka
    Nov 13, 2010 at 7:39
  • 6
    @David, I disagree that any problem can be solved in infinite time. You may want to read joelonsoftware.com/articles/HighNotes.html. "The real trouble with using a lot of mediocre programmers instead of a couple of good ones is that no matter how long they work, they never produce something as good as what the great programmers can produce."
    – user1249
    Nov 14, 2010 at 9:29
16
votes

Early arcade games.

Completely restricted by memory, processing power, I/O ...

What those programmers achieved in this environment was amazing.

E.g. Pacman - Perfect. Still playable after all these years.

6
  • 2
    Pacman's very famously not perfect. It glitches pretty badly when the level counter overflows one byte. ;)
    – Mason Wheeler
    Nov 12, 2010 at 20:04
  • Excellent! According to Wikipedia: "Pac-Man has a famous kill screen, often referred to as the Pac-Man Bomb Screen. The game's level counter was a single 8-bit byte, and could therefore store only 256 distinct values (0–255). Reaching the 256th level causes a counter that is used while drawing the fruit to overflow to zero, causing 256 fruits and seven blank spaces to be drawn. It can, however, be fixed with a patch."
    – Conor
    Nov 16, 2010 at 14:30
  • 1
    Early space shuttles perhaps???
    – Job
    Nov 18, 2010 at 0:04
  • Ms. Pac-Man was far superior
    – Tim Goodman
    Nov 18, 2010 at 15:08
  • 1
    @Mason, you can even get it on a T-shirt! errorwear.com/shirt-pacman.php
    – user1249
    Jan 13, 2011 at 7:24
15
votes

There is this one programmer where I work who has done the following on more than one occassion.

The most recent example was a project in our group that was assigned to 3 people. It was a complex LOB app with a 6 month schedule (that was an extremely aggressive timeline).

The week coding was to begin, 1 of the 3 people quit (for personal reasons) and another left on a very sudden medical leave. The lone remaining programmer went home to start programming and returned 3 months later with the completed application.

EDIT

Just to add some clarification. The resulting code was nearly perfect, spot-on to the requirements and our QA group was only able to document 3 defects.

12
  • He/she had to be pulling 100+ hour work weeks to pull that feat off. I bet he/she went through a ton of caffeine.
    – Ali
    Nov 12, 2010 at 13:08
  • 3
    The mythical man month in action. Take 2 people off the project == halve the timeline! Though it does make some sense that one focused developer can get things done quicker than trying to coordinate 3 people.
    – CodexArcanum
    Nov 12, 2010 at 14:41
  • If that person ever leaves the company on bad terms (no knowledge trade-off period), the company is going to be in big trouble...
    – Onion-Knight
    Nov 12, 2010 at 15:03
  • 3
    @bjarkef - the code is top notch. This coder is our company's best programmer and the very short time frame did not reduce the normal quality we typically get from him.
    – Walter
    Nov 15, 2010 at 19:32
  • 1
    @Walter, you cannot estimate axcurately until you have actually done it which contradicts when you need the number. This makes estimates east when you've done the task before, and hard when it is unknown to you. As long A's you don't underestimate the client Will be happy it ended up being cheaper than expected.
    – user1249
    Nov 15, 2010 at 20:50
13
votes

Maybe I'm just showing my age, but I think some of the truly impressive feats of programming are being ignored.

Steve Wozniak, Apple Disk II/RWTS

Steve designed the hardware and software together, using really cool tricks in the software to eliminate a lot of complexity (and cost) in the hardware. Normal floppy disk drives used an LED and photosensor near the hub of the disk to shine through a hole punched in the disk substrate. The output from the photosensor was connected to a processor interrupt so the processor would know when to start its timing routine to wait for the correct sector on a track to be under the read/write head (though a few used "hard-sectored disks" that had a hole to signal the start of each sector instead of just one for the start of the track).

Steve eliminated that hardware by designing the software to encode the data in a way that would let you start reading from an arbitrary spot on the disk, and not only decode the data itself, but figure out where (logically) in the track you were. The Disk II drive didn't have the LED/photsensor setup, and completely ignored the hole(s) in the floppy disk.

Gordon Letwin, HPFS

Gordon Letwin was an architect (and coder) on the OS/2 team. At least as I've heard the story, at one point he went on vacation to get away from things, and spend three weeks (or so) sailing around on his yacht (yes, early MS employees could afford things like that...) ...but to keep from getting too bored, he decided to take along his laptop.

When he returned, he had HPFS written, debugged, and working -- entirely in Intel 386 assembly language. The code he wrote was eventually sold as the "HPFS386" that was exclusive (at least at the time) to the LAN Manager Edition of OS/2. Another team then spent something like six months writing a version in C that became the "normal" HPFS included in the normal editions of OS/2 (and after IBM and MS broke up, IBM updated and re-compiled the C code to get their "HPFS386"). Although it's been modified and update since, if you were being fair about things, NTFS would probably be called "HPFS 2.0" (or maybe 3.0) -- there's no question that Microsoft's best current file system is still closely derived from what he designed.

Burroughs B220 tapes

These embodied (at least IMO) the real beginnings of object oriented programming. Where IBM tapes (for one example) had "labels" to tell about the format of data on the tape, Burroughs tapes developed a convention (I don't think anybody knows for sure who started it) of putting a small set of routines on the tape that would understand the data and how it was formatted, so you could manipulate the data correctly without knowing the details of how it was formatted. In other words, the first few "blocks" on the tape were basically a vtable in persisted form. You'd basically read the vtable into memory, then use the methods it defined to manipulate the data on the rest of the tape. All very neat and simple (if just slightly less than completely portable).

11
votes

John Carmack regularly impresses me with his feats of epic programming. According to one story, he was bored in his hotel room, so he wrote the Doom 3 lighting engine on his laptop over a weekend.

1
  • 2
    Pity he didn't write an actual game to go with it.
    – DeadMG
    Oct 1, 2012 at 14:15
10
votes

In spite of its questionable purpose, I'm still in awe about a captcha solver userscript for MegaUpload. It's a GreaseMonkey script written in JavaScript that has a bitmap decoder and a neural network that recognizes the letters in the captcha image, and I repeat, all of this is made using JavaScript.

1
  • 1
    Something to show the next person who moans to you about not being able to do good work because of the restrictions placed on them.
    – Jon Hopkins
    Nov 12, 2010 at 14:12
10
votes

Miguel de Icaza - the Mono project.

(He's also incredibly funny, witty and entertaining - at least from the brief talk I saw him give and in listening to some podcast interviews)

1
  • +1 The guy is obviously twins or perhaps triplets, all going under his name. There is no way that one person is involved in all the sh*t that he is, and on such a detailed level.
    – Dan Rosenstark
    Nov 15, 2010 at 14:32
8
votes

Anders Hejlsberg: the Turbo Pascal language and the compiler. I still think after all these years TPC is a masterpiece of software engineering. It is the most compact and the fastest native-code (starting from certain versions also optimizing) compiler I've ever seen. In the Borland Pascal IDE there was no "compilation" essentially. You modified your code, hit F9 and watched your program running straight away, and that was on a 8MHz or 12MHz machine. TPC had later evolved into Delphi, then Hejlsberg joined Microsoft and co-authored C#.

I'd love to take a look at the sources of earlier TPC versions, but unfortunately despite numerous requests Borland never opened them.

Hejlsberg's compiler was inspired by this book: Niklaus Wirth, Algorithms + Data Structures = Programs

1
  • Delphi still has a really fast compiler. I guess that's why it gets away with not supporting modify-and-continue.
    – Joeri Sebrechts
    Nov 17, 2010 at 21:27
6
votes

My training supervisor at Gemplus (now Gemalto) wrote the first JavaCard interpreter and OS during his own 3 month undergraduate training period. He single-handedly produced code from the spec. To produce the v.2, Gemplus put up a 30+ men team, without the guy, which worked for more than a year. When done but unable to get decent performance, they called him for help, and in a few days he and a colleague of his pointed out dozens of bottlenecks.

Java Card is a subset of Java intended to run in very small devices. His implementation ran in smartcards, which have a few MHz clock and 2 or 3 KB RAM.

Just for fun: another personal everyday feat of his was diving into a work trance, answering none of the questions we trainees yelled in the room, and then, suddenly 2 hours later popping alive again when his work was done and answering the 5 or 10 questions in a row.

3
  • 1
    I think I overshot the 200-guys thing... I can't remember too well, it was more than 10 years ago. I mailed him the question anyway!
    – Gabriel
    Nov 12, 2010 at 21:04
  • 1
    Damn, that guy must be a cyborg..
    – mlvljr
    Nov 13, 2010 at 15:09
  • Right, they were 30+ (=
    – Gabriel
    Nov 16, 2010 at 11:09
5
votes

I remember it very clearly; on highschool a classmate has coded a full-blown driving videogame in a couple of lessons with QB45 (Quick Basic 4.5). Vertical scrolling, stopwatch, lifes, levels; I was totally shocked.

1
  • 1
    I saw a guy do that with a text adventure game in Lisp. Not impressive in and of itself, but in the time he did it in... wow.
    – Michael K
    Nov 12, 2010 at 13:29
5
votes

Git and/or Linux

Linus Torvalds wrote an operating system kernel and a version control system from scratch. I don't know of anyone with a similar productivity.

7
  • 16
    A popular as Linux is, it's easy to overblow the difficulty and/or originality involved. It's almost entirely a clone of an existing system. Compared to (for one obvious example) Dave Cutler having written RSX/11, VMS and Windows NT from scratch, each a new and unique piece of work (and, for those who care, the VMS file system included version control) Linus hasn't even made it onto the map yet.
    – Jerry Coffin
    Nov 12, 2010 at 14:43
  • 1
    Kudos for VMS, but with the same reasoning you have to discount NT. Wikipedia says "[...] led by Dave Cutler to build Windows NT, and many elements of the design reflect earlier DEC experience with Cutler's VMS and RSX-11.". Anyway, i find Git impressive.
    – LennyProgrammers
    Nov 12, 2010 at 14:58
  • @Lenny222: while it's certainly true that you can find substantial similarity between VMS and Windows NT, it's not even close to being a clone like Linux is. I think pretty highly of Git as well -- but if it didn't exist, I can think of at least two or three others that would work fine in its place.
    – Jerry Coffin
    Nov 12, 2010 at 15:41
  • 5
    Didn't Linus work alone on Linux for like a month or so and then turned it a very sucessfull community project? Great work indeed, but I would not classify as an individual feat of programming, but rather as a very successful collaborative open source project.
    – David Reis
    Nov 12, 2010 at 20:10
  • 1
    I think what we're getting at here is the the fact that we now have X users using linux at home...I suppose you have a VMS box in your house? Grant it wasn't a single work from a single author...conversely if he hadn't done it, were those that are critical lined up to do it instead?
    – RobotHumans
    Nov 12, 2010 at 20:23
5
votes

Richard M Stallman comes to the top of my mind. This one man started giving life to GNU Emacs, GCC, GDB and many more notable programs. He even went on to author the original GPL. I have heard lore about him saying, he goes into a cave for 18 months and returns with over 150,000 lines of code that make the C compiler of GCC suite. Those 150,000 LOC gave the bed rock for g++ later on. In his own words, Emacs started as text editor, then it became a way of life, and now it is a religion for some of us. Even today, Emacs' design principles are used in many successful programs.

Donald E Knuth's TeX system is another program that comes to the mind when thinking about programming marvels.

2
  • You should try Reading the TeXbook. Is is not for mere mortals.
    – user1249
    Nov 15, 2010 at 20:54
  • Authoring the GPL isn't an impressive feat. It'd be like crediting yourself with Singletons.
    – DeadMG
    Oct 1, 2012 at 14:17
4
votes

In my degree class my classmate created an OS (albeit very basic but it booted from a floppy and did the necessary routines) as a project requirement in under 30 days. The duration is inclusive of both the necessary reading/research and actual coding.

I remember os.com mailing lists guys discouraging him saying he wouldn't be doing it in that few days and I also have seen him walking through the dorm at 5 am in the morning in a semi conscious state talking gibberish. :-)

He is an awesome programmer.

4
votes

I wasn't going to say anything, but someone piped up with "Ruby is a feat", so

John McCarthy with the original RECURSIVE FUNCTIONS OF SYMBOLIC EXPRESSIONS AND THEIR COMPUTATION BY MACHINE, ie the paper in which he defined Lisp back in 1960, when if was a concept people were just beginning to think seriously about. Talk about being ahead of your time.

4
  • nice reference. i was the one that piped up with ruby, for elegance over convention. this is ahead of time as well
    – RobotHumans
    Nov 12, 2010 at 20:27
  • 1
    McCarthy did a math article. It was only later that it was found that it could be converted into actual code.
    – user1249
    Nov 12, 2010 at 21:18
  • 1
    @Thorbjørn Ravn Andersen - I was careful to use "defined" as opposed to "implemented".
    – Inaimathi
    Nov 13, 2010 at 0:58
  • @Thor That is somewhat like denying Arthur Clarke credit for discovering (or inventing) the circular geosynchronous orbit simply because he couldn't build a rocket to put a satellite there, or because the technology did not exist yet.
    – Mark C
    Nov 19, 2010 at 5:37
4
votes

Walter Bright's Symantec/Zortec C++ compiler keeping up with compilers implemented by huge teams of programmers. Later, his design and implementation of D.

2
votes

Terry Winograd's SHRDLU.

Doug Lenat's Automated Mathematician.

2
  • SHRDLU has always amazed me!
    – user1249
    Nov 17, 2010 at 20:10
  • @Thorbjørn: Me too. That's part of that Minsky-lab spirit I soaked up, for better and for worse :)
    – Mike Dunlavey
    Nov 17, 2010 at 20:15
1
vote

Linus Trovalds writing Linux

0
votes

8 bits video game programmer on computer from the 80's (ie: Commodore 64), some pretty good game were in machine language, and their size were often smaller than a bitmap for a Windows icon ;-) Optimization at it's finest

Geos (a WYSIWYG OS for Commdore 64) was very impressive for it's time

-2
votes

I'm not sure that this counts really...but the fellow that developed ruby. I just can't get past the elegance of the language. And I believe it was the first language to support dynamic typing, although I am not the guy that started with punchcards and writing assembly so I could be mistaken.

1
  • 3
    It wasn't the first language with dynamic typing, and no, you shouldn't have started with punchcards to know this.
    – Jas
    Nov 12, 2010 at 13:06
-3
votes

David Heinemeier Hansson creating Rails.

The first time I installed Rails and set up a test site against a database and it did all the framework and CRUD stuff automatically, set up the test site and it was like my eyes were open for the first time to what can be done before you start programming. That was really impressive and I've seen ideas from that appearing all over the place in other languages/platforms since then.

7
  • How is this a feat of programming?
    – David Reis
    Nov 12, 2010 at 20:06
  • the fact that it does a good job at laying out tables with simple has, belongs to, habtm...further it abstracts database connectivity into a db agnostic classful structure where it almost seems like magic and all you have to do is a rails new -d (dbtype) bundler generate and rake db:create db:migrate command set. The generate scaffold for a base to work from is nice. it looks like someone doesn't like ruby/rails as these get downvoted pretty hard while they were new approaches to elegance over convention
    – RobotHumans
    Nov 12, 2010 at 20:20
  • after some reflection, i think it's because we were impressed by the artistic nature of the thing as opposed to technical prowess
    – RobotHumans
    Nov 12, 2010 at 20:32
  • It was a totally different way of thinking- making everything totally oriented towards what would be ideal rather than what was sufficient. It was like all of us web developers had been driving around in Model T fords and suddenly rails turns up and it's a modern BMW. It was a massive change in simplicity and user-friendliness in terms of developers as users and I've seen so many other platforms built on those ideas in the years since.
    – glenatron
    Nov 12, 2010 at 21:40
  • 1
    Fair enough, but notice that your question does not mention the guy's name, what he did, when he did it, etc. It just says: "Ruby is nice". Well Puppies are also Nice, but that' not an answer to this question.
    – David Reis
    Nov 15, 2010 at 0:11

Not the answer you're looking for? Browse other questions tagged or ask your own question.