Nothing wrong with the manager taking the team out after a long day, scheduling 3 months of long days is the problem.

There's nothing wrong with something like that once in a while....

Maybe I'm not clear with my complaint: some places have beer in the fridge, and a few select places have beer on tap. These kinds of places tend to longer hours.

We have beer in the fridge. We sometimes have beer on tap. (I think there's a service that provides a kegerator - it's not always there, but sometimes it is)

The office is a ghost town after 5pm any day, after 3pm on fridays. I've never seen anyone in the office on a weekend - the few times I've tried to get into the office to get my laptop or some other stuff, I've found it locked up.

FWIW my last 2 jobs before this one were beer and sometimes kegs in the office also. One of them was like my current one - every left on time, no work on weekends. The other had a few over-workers. That one was a high frequency trading company and the quantitative mathematicians tended to work odd, and sometimes high, hours. But they also got a piece of the pie in a very real way - if their models made lots of money, they made lots of money. Some of them were more driven by that than others

Job before that, I dunno, I worked remotely. All my jobs before that were startups and varied a lot. In the first startup I worked for, all those guys were my best friends, we hung out a lot (this did not include management, but also, there wasn't much management). Next startup was very professional and I only ever hung out with one guy outside work. Pretty sure no beer anywhere on premises. 3rd startup was actually the same as the first one but a few years later - less of a startup I guess. Still beer in the fridge but I don't remember if anyone ever drank at work.

I know a few in downtown Austin have built-in taps.

I can drink with the best of them but 1/2 beer and I'm no longer able to focus on programming. I've programmed while getting trashed a few times, but that's only on my side-projects.

RustyBrooks,

How do you break into the HFT firms as someone who isn't a math Olympiad or didn't go to a top 10 school?

I do a competitive on the side for fun but I'm not even that good. I'll consider myself good once I start placing high enough to win those free t-shirts and hoodies!

When you say long hours, do you mean people are hanging out after work and socializing? Does that count as your work hours? I hang out after work sometimes and drink. Been doing it less though since I have gained some weight . Darn free snacks and beers! *Shakes fist*

I did not go to a top school, and I am not really the type who is into programming competitions or puzzles or stuff like that. Some of the people who worked there *were* like that, and there was probably a bias towards those kinds of programmers. Getting hired was exhausting - programming exercise up front, which they are *super* serious about, and whiteboarding all day.

My skills, and the reason I think I got hired, are in system simplification. A lot of the work that I did while I was there was replacing complex systems with simpler ones. Sometimes the simpler ones were faster, or enabled speed improvements, and sometimes they just freed up brain space for developers/it/devops/etc. I was not really one of their hot **** developers.

It probably helped that I knew a few people who worked there, from previous jobs and from a school I went to. That probably bought me the benefit of the doubt - they were one of those places that will pass on anyone if someone in the interview process is not 100% on board.

The industry seems to really be contracting right now. I started there in like 2010. In 2013 or 2014 there was a round of layoffs. There was a 2nd round of layoffs in like... 2015? And I was part of those, along with around 20 other people. 10 of those people weren't "really" laid off, they went to work for a spinoff.

This year the HFT company got bought by a competitor, presumably because they were having trouble making money and wanted to consolidate costs. The spinoff is limping along but I think it's in trouble.

At some point the industry might consolidate enough that it can be profitable again. A big problem was that it is easy for an individual HFT company to become a significant portion of the market. I never verified this but I was told that on some days our firm was about 10% of the US equities market. It is really hard to expand by trading more because eventually there are no counter parties. If your HFT competitor are as good as you and they are also 10% of the market, then if there are more than a few of you, you all lose. It's a little like how a few decent players at a live high rake poker table will all lose money, even if each individually has an edge.

Access to the markets is pretty expensive, so the consolidation makes them money by removing a competitor, but also by removing millions/year in subscription costs.

Recall that I don't have a no bull**** detector. From their statement, I'd guess they are saying actually doing work until midnight, further evidenced by me witnessing an office of people working at 8pm during an interview. But if they mean "just hanging out and drinking beers," I wouldn't know. If they are paying flat salary, the terminology becomes more ambiguous.

Rusty, are you aware of the hacker homes in Austin and the type of people who reside in them? I didn't see many of these types at meetups, but perhaps my experience is shaded by the types of people I met out there. There was quite a few stock traders at those meetups. Seemed that a lot of people were dedicated to the long work hour. Just trying to gauge the difference between typical and atypical.

I'm aware of hacker houses but I doubt I know anyone who lives in one. I'm 40, all my friends own houses etc.

4chan found my game lol. Not a problem yet idk. I can't wait to get hit by LOIC though.

Well, I have a contract that went completely sideways. Not saying there wasn't any signs with this one, but they were a tad more subtle than usual. I was willing to tolerate a bit of headache due to the tech involved, but lines were definitely crossed on this one.

for my project course I had to write a file system for a small sized disk, ~4kb, and tested for all these gnarly edge cases and even put some try/catch clauses in places where I thought in case I missed some weird bug, I'd catch an array index out of bounds and I wouldn't fail the project automatically.

professor emails us yesterday and said they're not really testing for certain edge cases I put a lot of work into, and that if our program crashes, we'll have a chance to test one line of input at a time.

doh

This is encouraging to read, because this is me. I don't really have this gnarly passion for programming or doing hackathons or the **** that almost everyone I know seems to love, but I believe I am a good programmer with some marketable skills.

I'm trying really hard to program more, but it just isn't there for me. Maybe I need to consider a different career.

Did he email after your submission?

My guess: the professor looked at your code (and perhaps a couple other students), and decided to clarify about the "edge cases". That's the first time I've heard that term, "edge cases", btw.

"edge cases" is a pretty standard term, for annoying things that crop up a very small percentage of the time, but that nonetheless need to be handled. It can often be the case that a simple 10 line algorithm can balloon into hundreds of lines of spaghetti to take care of edge cases.

When designing a complex UI, I feel like half my job is "Edge cases".

If it's only half then you're lucky. General rule is you're gonna spend about the same amount of effort programming as designing. If it's simple to design it's simple to program and vice versa. I'd say I usually spend 80% of my programming time on edge cases. Generally the more user-facing the more edge cases.

more like 80+%

Yesterday my commits added 30 lines of code and removed 300+, refactoring bad code is such a satisfying feeling.

The code had been hard coding in IDs, which made them completely not modular, now they are components that accept props and the same component can run everywhere. I feel like react makes stuff like that so easy.

I think for really large applications, it's more like 99+% for some definition of "edge cases" - think about something like Google Search or Facebook - the real core of what it does is not that complicated, but they have millions of lines of code because there's a truly long tail of features that really only make sense in specific cases. Or Linux Kernel - it's millions of lines of code not because OS kernels are that complicated but because there's a long trail of devices that approximately no one uses that nevertheless need to be supported.

That's exactly what happened. Particularly, I was concerned about what would happen with read/write operations when the disk was full. Lots of off by one errors and stuff. This led to the exposure of several weird bugs that wouldn't let me fill my disk, including one really frustrating bug that had to do with how I implemented my bitmap that indicates where the open blocks are.

Another way to think about this is that computer science is about the general ideas and software engineering is about the edge cases

i like this.

also, you could almost take as a definition of elegant refactoring: "re-framing your problem so that the many edge cases vanish in the new solution." you see this everywhere: sentinel values, the null object pattern, monads in general, on and on.

and this isn't just about software -- see, eg, maxwell's equations, many entire branches of math, etc