bethly

It's not even lying: the cycle is red-green-refactor, not red-green-done. Knowing when to extract functionality (and when to inline functionality) is part of our job, not a management concern.

Anyone can pitch a tent: building a house is harder.

Given that we need a lot of housing and we don't have many home builders, there are a lot of tents floating around this industry.

I worked on code that had been in active development for sixteen years, but it only got to be sixteen-year-old code because it was reasonably well-written and maintained. Code quality and developer skill is what lets software become older than four years old without collapsing under its own weight.

Any team striving for a perfect architecture rather than building something that works and then seeing where the bottlenecks are isn't focused on "code quality": they are ignoring best practice in favor of intellectual satisfaction. The best software architecture is any architecture that solves the problem.

I've had business people tell me I should stop writing tests so I could go faster, even when I'm producing significantly more features than the developers around me who don't write tests. Once you get good at testing, writing code with tests is faster than writing code without tests: the challenge is the four to six year learning curve to get there.

I think we just need to stop telling people that code without tests is an option. It is an integral part of being a software developer, not a nice-to-have afterthought.

Hey, author here! I'd been thinking about how all the advice, discussions and books that have sprung up around white board algorithm interviews makes them seem more inevitable than they actually are. Discussing different kinds of interviews can help candidates do their best, rather than being surprised or knocked off balance when they encounter these recent interview formats.

Even if you don't trust the previous author, you can still use your "go to definition" shortcut and see the real thing: who is it who is coding one file at a time?

You shouldn't be considering the implementation details every time you want calculate the circumference. Even that pi is involved is WAY more information than the caller needs to know. And if you are debugging you need to know where to look for the answer, not the answer itself.

In that they don't think women should have to deal with sexism: damn skippy. Existing in an environment where some people think women should have to deal with sexism sucks. You don't get to be inclusive of both sexists and me, because their "beliefs" are about my actual life. My life is significantly better if the dudes around me will step up when sexism happens and deal with it without me having to even be involved. And it doesn't work if they are only objecting because "this will alienate women!"; it needs to be because they, personally, are hurt by sexism and don't want to be around it.

We also exclude programmers who insist on using nothing but GOTOs to structure their code, and sexism is way more annoying for me to work with than GOTO-filled code. You don't really think I should be inclusive of everyone: you just don't think sexism should be allowed to be a deal-breaker. Don't worry, if you want to work at a company that's inclusive of sexists that's way easier to find.

You can tell my strategies are possibly effective because of all the sexist men in these comments who are apparently deeply concerned that I will never work for them, or perhaps that they might accidentally end up in an organization that isn't sexist. Le horror!

Note that these attempts to drive our discussions and opinions from Hacker's News and deprive us of the opportunity to express our preferences all help maintain discrimination and the economic power a limited workforce provides. Men are so scared that they might have to compete on a level playing field, or possibly even be held accountable for their actions, that they externalize that fear onto women trying to get by in an actively-hostile world.

Dear men who are offended when people acknowledge the reality of sexism: I guarantee you, I have spent more time thinking about this than you. I had to, a fact of which I am already resentful, thanks everso. Consider before you comment that I know what I'm talking about, and that way you've assumed I don't? It is a manifestation of the very problem we are discussing. Every time you express your uninformed opinions that start from the assumption that the world is fair, you are proving that you are unprepared to contribute to this discussion.

I don't want to work in an implicitly or explicitly sexist environment. You want to somehow get to participate in the patriarchal norms, not face any criticism for that, and then still somehow have me want to work with you. As The Man In Black says, get used to disappointment. Your desired outcome isn't one of the options.

Instead of attempting reassuring your existing cultural cognition (http://www.culturalcognition.net/), try listening to the generous gifts of experience the women and marginalized people have offered in response to the question that was asked. You might learn that there is more to the world than you had known. That will make you a better developer, as well as someone I might someday be willing to work with.

Men of color don't get a pass on sexism, but that particular screening tool isn't effective because of the racism many of them face. The cost to them of publicly supporting marginalized people is higher, whereas it doesn't have any cost professionally for white men (citation from the Harvard Business Review: https://hbr.org/2016/03/women-and-minorities-are-penalized-f...) Thus, if white men don't do it, it's a pretty good signal, whereas if men of color don't do it, it doesn't tell me anything.

I suspect part of the disconnect is that not all the ex-mil folks I've worked with were coming from support units: overall the military still has a lower percentage of women than software development does, and it was even lower when some of the people I was working with were enlisted. Also, although as a civilian I would never make such generalizations, it does seem that different branches attract different kinds of people.

Military hiring, of taking all comers over certain transparently-published bars and then turfing out bad actors and training the rest, is a strategy I'm not sure any tech company has tried. I've long wanted to compare our interviewing strategies to "random" to see if our screening actually contributing anything to outcomes.

It's not that there is nothing that could be learned, but as a woman I wouldn't use someone's veteran status as signal on whether or not they are likely to be sexist.

This comment is an excellent example of why screening for racism is helpful in avoiding working for sexists like this.

As David Morrison said, the standard you walk past is the standard you accept. "How to signal you are on the side of non-sexists instead of the sexists" is probably a different thread, though: the original poster wanted to know what strategies marginalized developers had found effective and I shared mine.

It is unsurprising to me that effective strategies end up rubbing people the wrong way. The reason social media posts work as a filter is because they are a minorly-costly signals to send, and the suggestion that people be expected to do extra work to appear trustworthy, rather than being assumed to be trustworthy even though probability isn't on their side, is going to feel threatening. Privileged men feel entitled to the benefit of the doubt, even when it is women who bear the costs when they fail to live up to it.

This isn't just "my preferred topic": this is my life. This is whether I have a fair shot at promotion. It's whether I have to be 150% sure with a giant pile of citations by dudes before I say anything, or people trust my technical judgement. It's whether I can be most productive at my job, or I have to waste a bunch of time managing the anxieties and insecurities of the dudes I work with instead. It's about whether I have to spend time and energy sending Comey-like memos every week to document the sexual harassment against my female coworkers until they all quit or are fired anyway. It's about whether my coworkers will refuse to speak to me for six months because I made work "not fun" by not wanting to hear about his pick up artistry, and still get promoted. After that last one, now I only work places that don't prioritize catering to sexist developers over catering to me. There are plenty of options out there; there is no reason to settle for anything less than an actively anti-sexist workplace.

Many men treat this like an intellectual exercise: it's not. These days, I only work for the ones who already understand that.

All of life is tradeoffs. It doesn't matter if people "value" diversity in some abstract sense; it matters how they will value it when it comes into conflict with something else they value. If they can't think of anything they are willing to prioritise below diversity, they are almost certainly going to end up with a homogeneous team.

Also, I want to know that they are willing to tell people who only accept universally-applied, context-blind systems to take a hike. It is impossible to fix a cultural problem of disparate experiences with anything pretending to "objectivity".

Relatedly, the strategies I've used also change over time.

At this point in my career, I rely more on preemptive screening by employers. I list my talks about Cyborg Feminism and Feminist Software Development on my resume, along with my talk at Lesbians Who Tech on why imposter syndrome is a rational reaction to a lack of reliable feedback. I posted a blog post about how the decline of women in computer science is related to the rise of Internet culture and the disparate treatment experienced there. I regularly tweet about racism and sexism in tech. In my personal statement last time I was looking I said explicitly I was looking for an "intersectionally-feminist team".

This isn't a technique I would have used when I was starting out and needed Any Job. Now that I have a skill set in high demand, though, when people who don't want me around rule themselves out it just saves both of us time. And on the flip side, when a founder is excited that I name-drop Donna Haraway, it's always at least going to be a fun conversation.

That was not my experience working at a defense contractor, and it was obviously not because I do drugs. I haven't found military service a relevant signal for levels of sexism, either positive or negative.

Oh, the other thing is that what you are looking for will change over the course of your career. Early on, you need to work with people who can effectively mentor you. Your direct boss and the technical leaders on your team will matter, but the rest of the company can be kind of terrible and it will probably be fine. Once you start moving into technical leadership or architect roles, the additional work you probably need to do to have an impact in dysfunctional, sexist organization will be a much bigger drag on your productivity and moral. On the up side, by that point you will have the skills and resume to pick and choose better organizations.

I believe that noticing when that transition happens and swapping companies is key to persisting in the industry.

Being willing to fire people for "just" sexism is a great sign, and one many Big Name companies fail to manage.

I look for existing diversity, and of underrepresented minorities not just gender diversity (IME this is often a better signal.) Especially in upper levels: the % of Software Engineer 1s matters a lot less than the % of architects and tech leads.

Second (and I kind of hate to give this one away), I ask "what do you sacrifice when recruiting to ensure diversity?" The answer should be either "it takes us longer to recruit, because we ensure a balanced pool" or "we have changed our process to allow many types of candidates to shine". If the answer is just "we spend lots of money on sending people to Grace Hopper!" it means they aren't willing to inconvenience or piss off overrepresented engineers. You can't fix culture problems by spending money, and when "diversity" is seen as separate from "recruiting" in general it's a clear sign of a problem.

I also use the Internet: I look on LinkedIn for people who have left the company, see how long women stayed and reach out to find out why they left if they did. Backchannel mailing lists are great ways to find a vouch. I check the social media profiles of their prominent engineers and search with keywords like "feminism", "women" and "she". See how they talk about women in the field, whether they follow women on Twitter, whether they posted angry anti-Hillary memes. I've found that a lot more effective than relying on direct questioning, because there are companies out there that will say whatever they think will let them add you as a shield against the accusations of sexism they are facing. Especially my boss: if my boss is going to be a white dude, he had better have publicly condemned sexism in a way that feels constructive and genuine to me.

One other internet trick is to look at where the women you respect are working and apply there. At the very least, you would get to work with technical mentors you admire and can learn from.

You can find good spots here. Seriously, though, trust your instincts, don't be afraid to walk away and don't be afraid to take an "unsexy" job at an old-school company with an HR department.

It's the total, and reflects all shares whether or not they are likely to have vested by then; vesting schedules vary widely, and hopefully are something people will be told about when getting their offer.

This tool reflects that profit by assuming a current strike price based on average valuation changes between rounds. Instead of saying that you get the full value of the stock you buy, the calculation takes into account that you also have to purchase the stock.

This tool doesn't explain the dynamics going into the estimated profits (though it links to things that do), but it helps people avoid over-estimating the value by missing dynamics like strike price.

You can also choose not to exercise and just walk away, in which case the value is still 0. Though I imagine a lot of people would have trouble doing that even if they didn't think the company was doing great.

It assumes that the total amount of money raised over all rounds will be $41 million, rather than that the seed round valuation is $40 million. All the investors, whether they came in before or after the employee joined, will still get their money back before the employee starts to profit.

Yep! That's part of what changes with the round: we assume a relatively-average dilution of 35% per round.

In the worst case, an option is going to be worth 0 and in the best case it's going to be worth > 0, so the value of an option is always positive until the company actually dies.

Lottery tickets, unlike options, cost money up front. You still can't lose more than you pay, so the payoff curve is similarly non-linear, but unless you can trade options for cash options will always beat lottery tickets.

Lottery tickets are easy to diversify in that you can buy a variety of numbers. (The times when lotteries have become a net-positive buy the buyers takes advantage of this fact.) Employee startup options are more like buying one set of numbers over and over again. On the other hand, the odds of getting any payoff from options are somewhere around 15-30%, whereas the odds of getting any payoff from a California SuperLotto Plus ticket are ~4.3%. Because you can diversify ticket numbers, you could get the same odds of getting any payoff by buying 14 tickets with different Mega numbers, which would earn you $1 to the $14 you spent. If you could buy stock in many different startups, you would be called a "venture capitalist" and those folks on average do much better than people who play the lottery.

Finally, the maximum payout of a lottery ticket is capped and known ahead of time. The largest lottery win in the US was $656 million. On the other hand you don't know going into a startup what the payoff for that particular startup is going to be, and the largest exit of all time was Facebook at $104 billion. Just like with the lottery you don't know how many ways you are splitting the payoff, but unlike the lottery it's going to be based on the decisions of the board/founders, rather than random.

Basically, the lottery is a lot simpler than a startup, with few sources of actual uncertainty, and so there's no real risk involved. It is just gambling: you can do the math to figure out what edge the house has and figure out for sure that you shouldn't do it. Startup options, on the other hand, reflect actually-unknown unknowns, and so are more valuable to those who hold more-optimist-than-average beliefs about the probability of that particular startup succeeding.

Thanks for the feedback!

Preferred shares are reflected in the current payoffs; that's what's responsible for the "low exit" portion of the slider not paying off at all. Are you talking about participation preferences that let VCs double-dip? Those are a whole different kettle of fish, but also not as standard as preferred shares.

Ah, the glories of default html input behavior!

If you ask the company should tell you what the "the total shares outstanding on a fully-diluted basis" is. If they don't tell you when you ask it's cause for concern; otherwise there isn't a way to tell what your compensation might be worth.

Thanks! I really appreciate you letting me know; it's fixed now :)

(For the curious, I added a click mask to the modal contents and forgot to update the links to fire anyway.)

Thanks! I really appreciate the feedback.

The price of the options eating into returns is reflected in the number we present (we assume a consistent valuation growth by stage and at exit), but taxes aren't and those can 40%+ in the US, which people don't necessarily expect. Limited exercise windows are one of the things we have in the list of ways this can go horribly wrong, but you are right that it is something that you can plan for if you know it's coming.

Thinking about it, that and early exercise might both be candidates for some sort of "list of questions to ask" tool...