Ask HN: What's the most important piece of code you've written?

My first professional job was doing a little bit of work for the Human Genome Project. I've never done anything else nearly as worthwhile.

I wrote an entire e-commerce inventory SaaS platform singlehandledly (frontend, backend, database, cache, nginx, blah blah blah) that had 100 users and about $30k/mo revenue.

Then I got kicked out by the owner high and dry after I refused to take sales calls and support calls on top of my developing responsibilities. :)

Software is still running today. It's probably at $50k/mo.

I guess when I was 12 and had a Commodore 64 in my bedroom. I stayed up late and got up early to learn basic, put sprites on the screen, debug and otherwise get caught up in whatever I wanted to create that day.

I’d like to say thank you to my father for buying it and keeping the chocolate milk and powdered doughnuts flowing.

I currently work for a company that is focused in increasing online privacy. Everything I write now supports this effort. I've never felt better about a job.

Second to that would be the games I wrote on my TI back in high school (90s) when I should have been learning trig. That and some late night script kiddie exploits got me on the path to a career in tech.

It really depends on what you mean by important? Some of my most important code saves a company millions of dollars a year... It's not at all important to me, but it does it's job and it is important to someone.

The most important code to me is a difficult question... I look at code as solutions to problem. What's most important to me, is the code that made me realize none of us are free (weird as that sounds). I can track anyone on the web from their discussions (provided enough comments, i.e. 1000+ words):

https://twitter.com/AustinGWalters/status/104189476543920128...

Expanding on this system I built an investment framework part of which is incorporated in (https://projectpiglet.com). It has helped me make 100% yoy returns since 2013 (including >200% last year, even with the downturn):

https://twitter.com/AustinGWalters/status/107014266501716787...

It's the basis for my startup, but perhaps more importantly, made me realize we have no freedom. I can do this with no funds, in my house, with a computer. I track a couple million people in real-time (public data only)... I wrote a blog post about my introspection:

https://austingwalters.com/the-last-free-generation/

In a sense, it's open my eyes to the dystopia we are building and it's changed my outlook dramatically.

Well, this is important to me, but it isn't exactly a scientific breakthrough or as worthwhile as what others might say, but several years ago at the utility I'm currently working for, we have this stupid method of adding users to a HMI...by manually typing in their user ID, their name and some default password into the user credential file onto the HMI (this is also a crappy HMI software, so the text document isn't even hashed) and then emailing said password to the users.

Anyways, what I did was write a VBA code that

1) Allowed me to take the user ID of the people that needed to be added into said HMI, then get their first name and last name from outlook using an outlook api and automatically adds these users into the credential files.

2) Send the email to the same list of users after I update the file onto the HMI (the HMIs themselves in an air gap network).

Why is this important? Because it got some of my other coworkers to trust that I can find bullshit problems and solve them (this isn't the only bullshit problem I have to do over the years), thus gaining their trust and overtime, got a promotion.

Probably this UI: https://www.facebook.com/live/create

I didn't do it alone of course (probably a dozen people have worked on the UI, and hundreds on the backend), but it was my main project for around 9 months. Lots of videos that people care a lot about go through there.

For me, it's been codes that have been for internal use. One ran a room full of production machines. Another has been in use for a decade, and is the main utility for manufacturing a complex apparatus that requires a lot of adjustment and testing steps.

So each time one of these programs goes through its mainloop, a little bit of money comes into the kitty. ;-)

Possibly a coincidence, but these are also the programs that I was the most disciplined about writing so that they have been in production for years with extremely small and manageable lists of known bugs.

It's not really coding, but I helped with troubleshooting and analysis when Obamacare first rolled out.

There was quite a fire-- the front-end would crash, lines were endless, etc. The government quickly formed a Tiger Team, with contributors from various companies.

I didn't care for the way the legislation was formed, but I am proud of my involvement. I gave it 100%.

wrote an entire CRM platform for a niche which is still my business today. revenue is at ~200,000 $ per month, profit margin is >60%.

Tech stack: MySQL, Java EE, JQuery, Bootstrap and the like. No big frameworks. All hosted on Google Cloud.

The most important code I ever wrote was probably the control system for a submersible robot we developed in grad school. It was important not so much from a money standpoint, but because a whole group of graduate students got their Masters' and Doctoral degrees using it, it was sort of my bequest to humanity. And boy was it a hack...

I had decided to use a new single board computer on the robot that used a compact PCI bus, which at the time was a brand-new standard. It was very expensive - $25,000 - which was a whole lot of money for a university lab, but the computer had specs that we just couldn't beat with other existing single-board computers at that time.

There were no available compact PCI motor controller boards, so we had to use a motor controller board that was built for a different bus standard, and then convert from the Compact PCI system to the other board using a bridge chip. The particular motor controller board we chose was based on an 8-bit motor controller IC, the LM629. This particular chip uses memory-mapped 8-bit registers, and in order to communicate with it you have to write and read the registers in a very specific order. If you do anything in the wrong order, or you try to write to a read-only register, or vice-versa, the chip generates an error.

I was a decent low—level C programmer at that time, and was able to crank out the code in two days. But it didn't work. Whenever we tried to communicate with the chip, it threw an error. I went over the code with a fine-toothed comb, and I was absolutely certain it was all correct. I had no idea what was wrong. I was looking pretty bad to my advisor; I was the C stud, and I couldn't even write this simple device driver. And worse, I had recommended that we use this particular computer system, which cost $25,000, far more expensive than any other SBC we had ever bought, and now I couldn't make the thing work.

Finally, after banging my head against it for a week and making no progress, we threw up our hands and asked the motor controller board vendor if we could bring our system to their facility and get their help debugging it.

We arrived at the vendor and set up. Their programmer checked my code, and he couldn't find anything wrong with it either. After two days the owner took pity on us and asked his best engineer, a digital logic expert, to help us. He carted in a $20,000 digital logic analyzer and hooked it up and had me run my code. What he discovered was that when I had issued an eight-bit read, the chip saw a 16-bit read, which it wasn't expecting, so it threw an error, because the high-order byte was getting read from a write-only register. But the code was clearly issuing an 8-bit read. So where was the 16-bit read coming from?

It turned out the bridge chip had a bug. When it saw an incoming 8-bit read request on one bus, it translated it into a 16-bit read on the other, then threw away the most significant byte. We called the manufacturer, and were told "that's known, documented behavior - it's clearly spelled out in the manual." And when we checked, sure enough, it was - it was mentioned on page 48 in the third footnote, in 8-point type.

The solution we eventually came up with was to cut all of the address bus lines on the motor controller board and shift them to the right by one, and then take the least significant bit line and connect it to most significant line on the address bus. That way, access requests to any odd 16-bit memory address would map into unmapped register space so the LM629 wouldn't see them. Then I rewrote the code to only use even memory addresses. Worked like a charm. But I still feel sorry for the grad students who had that robot after I graduated. There was no way they ever figured out what I had done. Or why.

I'm an IT/SysAdmin type guy so I don't really code beyond web stuff at a hobbyist level, though I did do CompSci at uni way back when. However when I followed my partner overseas for a few years, the only job I had while away was working in a computational research group at a university.

I was brought on to provide a nice looking web UI for an online system that scientists could use to submit jobs to a server running molecular dynamics simulations. After finishing the job queue and interfaces though, I had plenty of time to spare. So the boss asked me to try and run some necessary manipulations on the uploaded PDB (protein data bank) files, before they would be sent to run in the simulation.

It's still the only time I have ever actually used direct material from my undergrad degree at work - using Dijkstra's algorithm and standard deviation to discover protein structural features, then manipulating them. It was also the only time I ever used the Python I spent just one semester learning a decade before. With all those libraries, I can understand why it was the language of choice my boss' field.

I made a log of the jobs as well, and in the 5 year since then, over 10,000 jobs have been submitted, helping scientists from dozens of countries all over the world. It's also the only code I've written since my uni days which isn't just typical web app type stuff. It's just this pokey little online service but I'm really proud of it, and I hope I've contributed to humanity's progress in some tiny way.

In 1987 I wrote about 150 lines of 68000 assembler that applied a low-pass filter to every sound heard by the towed-array sonar system on the Los Angeles class fast-attack submarines. I was an intern (co-op) and nobody even looked at my code. But it fixed the problem that had been plaguing the ANN-BQQ-5D system, preventing it from going into trials and 'management' just shipped it once the tests started passing.

Compared to other things (including some of the ID3 tag processing code that's still in Windows today, and OLEView) this seems pretty important.

A ringbox mailbox multi-threaded implementation.

Whilst I'm sort of proud of the 'word at a time', always contiguous and zero copy aspects of it... the important bit is the metadata.

All the optimization was really to allow me to record metadata around...

* When was a message posted?

* When did it start to be handled?

* When was it finished being handled?

* What was the correlation ID? (ie. the ultimate causal source of the events chain)

...and then extract this data into a sqlite DB and either query that directly or to create message sequence diagrams with the mscgen tool.

Why is this important?

Because I have lost count of the really hard, gnarly real time bugs I have solved using this tool.

Machine vision applications to inspect inkjet printheads during manufacturing. You've probably used a cartridge my apps inspected.

Nothing.

The most important code I ever wrote was as a phd student. No one uses that code, but they use the concepts behind it every single day.

The Q was how does one manage large numbers of heterogeneous containers (VM sprawl was already considered a problem, but we said if containers took over the sprawl would be worse). At the time the primary container mechanism was Solaris Zones and people uses ZFS clone functionality to create per zone container file systems. However, this only really helps if your containers are homogeneous. I built a system that created container file systems out of composable layers. One would define template file systems as a collection of layers that are stored in local and remote layer repositories and these layers would be instantly unioned together on demand when one provisioned new containers. We also demonstrated that since heterogeneous containers are now easy to manage, we can not just use containers as lower overhead but similarly long lived VM replacements, but also enable them to be used ephemerally, opening up many new use cases.

You probably know of this by a different name today.

Adventures in Game Development is the perfect example of the risks of giving assholes your money for crowdfunded projects.

The guy that runs it spent years bragging about doing drugs and drinking to all hours of the night with nothing to show for it and then shits on his backers for asking about the progress of the project (none).

The first Flash game that I released [1].

I had been programming for about 6 years but had never finished anything of substance. The game took about 6 weeks of working on it between classes (and with the help of a friend with the polish and graphics). It went on to be played millions of times and made several thousand of dollars in ad revenue. Even 10 years and a PhD later, this is one of my biggest accomplishments.

It was one of those moments that makes me feel like now I can actually do this (this feeling seems to repeat every few years...).

[1] http://www.mindjolt.com/stay-up.html

Like others, it depends upon what you mean by important.

In grad school I helped another PhD student by writing control software for a device used to assess the locality and severity of brain lesions. I got a tee shirt for doing that work.

I’d have to say the headless server infra/features for Windows server 2003. The previous dev quit saying it was impossible, but as a new engineer I didn’t know any better.

I wrote a data collection and messaging framework for devices with a few MBs of memory, that had to stay powered-on and connected to a TCP socket server for about a year before they were going to be accessible again for maintenance etc. A bug in the code that could require a reset of the device could cost the company a big client. I couldn't find any ready-made solution that would satisfy all of the requirements. So I wrote, tested, documented and deployed it in a month almost single handedly. Two and a half years later, that code is still serving them.

The most important code I “wrote” was merged with code from my wife and produced a tiny human executable.