Hacker News Reader

dspeyeron I Knew a Programmer Who Went Completely Insane

At a much higher cost. A company with a dozen customers can drop the worst of them without taking a major hit. An employee with one employer has to risk everything. Especially since companies are often communities, so community membership is at risk too.

dspeyeron Security theater martial law and a tale that trumps every cop-n-donut joke ever

You have to be willing to measure the cost of lives. No one likes doing it, but there's always opportunities to spend a lot of money to save a few lives, and a limited amount of money. If you don't make tradeoffs intelligently and deliberately, you wind up trading many lives for few lives.

dspeyeron Google Outage

If it's a main multimegawatt power cable, it could take a while to plug back in (first the charred corpse needs to be removed). But it doesn't take long to redirect all traffic to another datacenter.

dspeyeron Google Outage

Google is better designed than that. No single cable failure could cause a user-facing outage of more than a few minutes.

dspeyeron Women, Tech Conferences and the Bullshit Surrounding It

Not reject, but kind of squint suspiciously at.

dspeyeron Women, Tech Conferences and the Bullshit Surrounding It

The number being that low is suspicious. In my vague, personal experience, the upper tiers of computing are about 10% female. The odds of 1 in 23 are 0.315. Plausible, but worth investigating.

Are the speaker selectors favoring men? That should be possible to test for. If so, it needs to get clamped down on hard.

Are women less inclined to try to speak at these conferences? If so, why? It's likely that making presenting more female-friendly would be worth it.

Is this subfield just that male? If so, it's not the conference-organizers' place to deal with it, but someone should sniff around for active biases.

In short, this is suspicious enough to justify a real investigation, and I'd be very interested to read the results of one.

dspeyeron Tests Are Overhyped

Principles I've found:

* A test should accept any correct behaviour from the tested code. Anything which is not in the requirements, should not be enforced by the test.

* A test should not use the same logic as the code to find the "right" answer.

* A function whose semantics are likely to be changed in the next refactor should not be invoked from a test.

* Whenever a test fails, make a note of whether you got it to pass by changing the test or by changing the code. If it's the test more than 2/3 of the time, it's a bad test.

* If you can't write a good test, don't write a test at all. See if you can write a good test for the code that calls this code instead.

dspeyeron Tests Are Overhyped

> What I don't understand about comments like this is that a whole section of your code, both runtime / deliverable code and test code had become worthless

Not so. The code demonstrated that the first algorithm wasn't good enough and provided the experience needed to write the second one. The tests (hopefully) made the first algorithm's code maintainable, but it turns out there was no need to maintain it.

dspeyeron Women, Tech Conferences and the Bullshit Surrounding It

Did you have similar problems recruiting make execs? If not, do you have any idea why the difference?

dspeyeron How Microsoft quietly built the city of the future

That's an exaggeration. While Seattle is painfully distant (painful really is the word for the 520 bridge) Bellevue is right next door. Bellevue is a city of 122k people, bigger than Ann Arbor. It's no great cultural center, but it's certainly more than just Microsoft people.

dspeyeron How Microsoft quietly built the city of the future

Thanks.

I didn't realize there were more pages. I just thought it was a very short and fluffy article.

dspeyeron Whose Water Is it Anyway?

Not let, sell. PNW needs to export something.

And the idea that Seattle and Spokane have a strong enough cultural bond to secede together is just silly.

dspeyeron Whose Water Is it Anyway?

Last estimate I saw for a Columbia River to Los Angeles pipeline was $150G. Expensive, but not out of reach for the US government. And there's enough water there to take care of the southwest for a long time.

The nasty things will happen across national borders. Like, when things in California get tight, will Mexico continue to receive any of the Colorado? Or worse, what about the Upper Ganges Aquifer, already shrinking and used by both India and Pakistan.

dspeyeron Poll: How many programmers don't know about bit-wise operations?

That's a code snippet where extra parenthesis might help. The first several times I read it as:

    ((width * bytes_per_pixel)+(3&~3))

and thought "what's the point?"

I don't know why bitwise ops are so low precedence. They've always felt more like multiplication to me.

dspeyeron Poll: How many programmers don't know about bit-wise operations?

    struct header __atribute__((packed)) {
      unsigned int highnibble : 4;
      unsigned int flag_foo : 1;
      unsigned int flag_bar : 1;
      unsigned int flag_baz : 1;
      unsigned int flag_qux : 1;
    }
    
    if (((header)value).flag_bar) {

dspeyeron The Ivy League Was Another Planet

How automatic is automatic?

My brother went to Harvard, partly because their financial aid offer was best, and I recall my father cursing the required forms. He's a lawyer, so ordinary legalese doesn't phase him, but he found those troublesome. Now I imagine a family without a JD in it, or even without a BA in it, trying to fill them out...

dspeyeron The Ivy League Was Another Planet

The verb ape is not simply a verbed version of the noun. It means "shallowly imitate".

dspeyeron Google Maps: “Treasure Map” mode

Last year's was 8-bit

dspeyeron Can We Stop Drawing Trees on Top of Skyscrapers?

Still, at least consider if climbing vines on trellises might serve better before putting a ton of effort into trees per se. They provide most of the same greenery benefits, can be made to fit the space, don't have dangerous roots, and are in many ways more robust. Plus they can have flowers or fruits near floor level.

dspeyeron Adria Richards, PyCon, and How We All Lost

I always use probabilistic logic for real-world questions. Certainty is just too rare.

I understand and agree with your analysis except for one point: as I understand it, the exact text of the joke has not been published, sadly preventing post mortem evaluation.

dspeyeron The Myth of Focus

Multiple projects mean that when one gets blocked on an outside obstacle, you can shift to another while you wait. However, if the queue of blocked projects grows without bound, then you need to pull back. I've been there.

dspeyeron Adria Richards, PyCon, and How We All Lost

My arrows are for causality. It's simpler this way. Evidence flows the other way. If sexism causes offence, then offence is evidence for sexism. How strong evidence depends on the likelihood of the consequence in absence of the cause. Bayes FTW!

As for concluding !S, most statements aren't sexist. In the absence of meaningful evidence for S, we should conclude !S (or rather, p(S)<.1).

dspeyeron Adria Richards, PyCon, and How We All Lost

No, that's not what I'm saying.

The causal network looks like this:

       S->O
       |
       V
    D->A
    |
    V
    H
    
    S=sexism
    O=other people being offended
    A=Adria being offended
    D=Adria being dishonest
    H=Adria being hypocritical

We know that A and H are true. We don't have data for the others. O would allow us to conclude S. H suggests D, which explains A, causing A not to be evidence for S.

It's like how flying saucer cultists being visibly crazy is evidence against flying saucers if their statements were the evidence in the first place, but if you had actual radar tracks, then it wouldn't matter.

dspeyeron Would it be a good idea to have test coverage in with the tests?

I've seen codebases that suffered from too much testing. It overcomplicated the code and made it brittle without catching actual bugs. Higher quality tests that would have been useful and allowed refactoring would have had lower coverage metrics, and that wasn't acceptable.

Granted, too little testing is probably the more common error.

dspeyeron Adria Richards, PyCon, and How We All Lost

With that phrase, you could find others to vouch for its offensiveness, so she would be far less relevant.

dspeyeron Adria Richards, PyCon, and How We All Lost

Wikia is a for-profit supplier of wiki hosting. bigbangtheory.wikia.com is a (probably not-for-profit) fan group that uses wikia's product.

dspeyeron Adria Richards, PyCon, and How We All Lost

Breathing is a bad point of comparison. Stopping breathing is much, much more costly to you than making a certain sort of joke.

I think if someone said "Could you stop using words containing "e"? Such words hurt my brain." I think I would.

If someone said "Quit your "e" words, you bad guy!" I'd be less inclined to. "Words with the letter "e" are bad words, and you need to stop using them" even less (the difference between these two being hypocrisy).

dspeyeron Adria Richards, PyCon, and How We All Lost

And, as usual with ad hominems, we must be careful not to throw out too much.

She argues that the joke was offensive. Her evidence was that she was offended. Evidence that her concept of offendedness is flawed does argue against the offensiveness of the joke.

Had she asked three people sitting near her if they were offended, and they had said yes, then she would no longer be a part of the argument and her hypocracy would not be relevant. But she didn't so it is.

dspeyeron Adria Richards, PyCon, and How We All Lost

This bit of definition only matters if there's someone involved without power.

Presumably Adria is claiming to be without power. She has the power to get someone fired with a single tweet. She has so much power she's gone mad with it.

dspeyeron How to Hire More Women at Your Startup

It's not about culture. It's about assumptions.

If my employer tried to support my masculinity by giving me beer, I would be offended.