29,438 karma · joined April 10, 2012
@geofft on Twitter and GitHub; also that username at that domain name for email.
I live in Brooklyn, NY, and can generally be found thinking too hard about kernels, linkers, or crypto.
'lizdenys and I make a podcast about practical personal digital security: https://looseleafsecurity.com or search for "Loose Leaf Security" on your favorite podcast app.
(They look cool and I'm enjoying reading about them, but for the things I do for my day job where C is one of the obvious options, I'm not sure they're better options, but usually one of the languages above is a better option.)
Reading from a pointer past the bounds of the object being pointed to is undefined, not implementation-defined. How do you propose to write a size_t strlen(const char *s) that successfully compiles with your proposed compiler?
(Depending on jurisdiction, they also have other advantages like not requiring a license, being able to travel in bike lanes / greenways and avoid traffic congestion, etc.)
(To be clear, I don't see why we actually need to pedal backwards. Assuming we're otherwise comfortable with a bike-by-wire system as the author advocates, a a brake-by-wire lever should be super cheap and has none of the maintenance complexity of physical brakes.)
As an SRE it's my job to make sure the computers I'm responsible for are working properly, and to actively pay attention and gather information about whether that's true. If there's too much load on the site for our current set of servers to handle, even if it's not my job to actually order more, it's certainly my job to say something to the people who can. If we're okay now but we won't be in a month, that's also my job. And if a disk fills up on a server and it stops working, "Oh, I wasn't monitoring disk space, and the program never logged anything" is no excuse. Nor is it an excuse that I'm not good at guessing what problems are likely to happen; that, also, is my job and why they look for people with experience doing this job successfully.
Isn't there all the more reason that someone whose job is looking after people and not computers should figure out if they're overworked or about to be overworked and if they're in comfortable, humane conditions?
But those two things are never going to be perfectly aligned. Enjoyment of the process does not guarantee working results that people want to pay for, and the thing that people want to pay for may be deeply unenjoyable. You can choose to tackle a subset of the problem space that happens to be both enjoyable and worthwhile, but you ought to admit that you are looking at a subset of the problem space.
I do a lot of other things with my time that are also near the confluence of enjoyable and marketable. I sing, for instance, and choir rehearsal and individual vocal practice are both a lot of fun. But any professional musician will tell you that if you want to be world-class at it, you occasionally have to wring the fun out of it. I've chosen the other option: I'm a volunteer singer with my church choir. I sing because it is fun, and only insofar as it is fun. We do new and not-very-difficult music each week, and an hour of effort is inherently an hour of reward. If my choir director said we're going to spend six months drilling some difficult measures and perfecting our German pronunciation, we'd all leave.
(In that sense, a volunteer church choir is rather like hobbyist open-source development. If it produces an enjoyable result for the public, great, and we do find that rewarding. But there is an upper bound on just how hard we're going to try and on the complexity of what we commit to doing.)
If you want an hour of reward after an hour of programming effort, that's fine! But the employment contract I've signed promises a year of payment after a year of effort. Occasionally, I want to work on things that are immediately rewarding because that helps with motivation. And it's important that such work is available. But I have a job that ultimately demands specific products, some of which just aren't going to be immediately rewarding, some of which do actually take a lengthy learning process, and - most importantly - many of which require building on top of other people's work that I could understand if I wanted but there is no business value in doing so up front.
(Incidentally, in the other direction, there are cases where I want to tackle problems that promise years of reward only after multiple years of effort, and figuring out how to get time to work on those problems is actually hard, too.)
We share almost all our code internally and leverage lots of open-source code so that we all have the option of understanding the code if we need it - but we rely on division of labor so that we have the option of building something on top of the existing base today, if we need that, which is more often what we need.
If you want to work on a hobbyist UNIX-like kernel for enjoyment's sake, great, I'm genuinely happy for you. But my servers at work are going to keep running a kernel with millions of lines of code I've never read, because I need those servers to work now.
There are a huge number of languages where you pass strings along with their length. There are also plenty of languages where you know (either via compile-time analysis and an insistence on writing code in an analyzable way, as in Rust, or via mild runtime overhead, as in Swift or Haskell or Java) that the memory remains still valid when you reference it. But those languages are not a subset of C. They either need more information about a program than a C program contains, or they disallow operations that the C abstract machine can perform.
So, yes, I think that's a defensible position, and if you wanted to characterize my decision to vote as evidence that I believe that my government as an institution is salvageable and not inherently hopeless the way I believe the police as an institution are, I don't think you'd be wrong.
The other thing is that you wouldn't really do away with elections, you'd do away with the whole organization (the government) that you're electing people to. Which is, in fact, the anarchist position, and I think that position is also defensible.
I really don't think it is either well-defined or relevant whether the individuals are "good" or "bad." As an analogy that is hopefully clear at the risk of being a little inflammatory: almost all of the 18-year-old conscripts of any random country you might be at war with have absolutely zero influence over the political direction of that country. But that doesn't change the fact that people handed them a gun and told them (even, perhaps, forced them) to shoot at you. They can be utterly kind and friendly individuals (see e.g. the World War I Christmas truce), but if they're about to shoot at you, that's a problem regardless of whether they're kind. When someone says "The XYZ army is about to invade us," it makes no sense to say "Well most of the XYZ soldiers are good people!"
And no, the distinction is not whether they're actually shooting at you. If they're lined up on the border ready to invade, if they're in training camps practicing shooting, if they're not assigned to be a footsoldier but they are assigned to build infrastructure and implement logistics for the invasion, etc., they're contributing to the invasion, and you'd still say that the entire invading army needs to be stopped. Saying that most of the invading army hasn't seen active combat yet and therefore isn't personally culpable is also nonsensical.
Or put another way - in gang-ridden territories or in post-war countries ruled by warlords, it's often the case that gang members or soldiers will do things to genuinely benefit the people who live there, running essential food and medicines across blockades, keeping people safe from other gangs or warlords, etc. That hardly justifies the gang or the warlord as a good government, though!
Every single person who works for a police department, without exception, works for an organization that is doing bad things. If, somehow, the specific things they do as individuals are good, when they do them on behalf of the organization, it serves to increase the legitimacy of that organization to gain more power and do more bad things in the future. If they are good people who haven't incurred personal moral culpability for doing anything wrong, great! I'm not in the business of judging souls for heaven or hell. I'm in the business of having opinions, as a citizen of a republic, about how that republic should be constituted.
As mentioned upthread, this story of the small town that has turned into a front organization for a police force that, were they not operating under the label "police," would be a gang of highwaymen, is a pattern that repeats itself a lot! When you see something that happens repeatedly from many different individuals who all happen to be in the same circumstance, Occam's Razor demands you at least consider investigating the circumstance and not the people.
There are almost certainly a large number of genuinely good and kindhearted and well-intentioned people in the employ of various police departments. They should leave their jobs, lest they stop being good and kindhearted and well-intentioned and become dangerous to society. Ideally, they would do that of their own free will; alternatively, we can defund their payroll. I have no problem with the individual former police officer merely by virtue of their former employment, and I am more than happy to judge them as an individual on their own merits.
There is clearly a need for some governmental mechanism for public safety: it's also clear that the police, as we know it are the wrong starting point, and if at all the existence of the police prevents crime, it's by coincidence.
(I also have no problem being partisan about my political views. They're political views! If you could get everyone of every political view to agree on something, it's probably ineffective. Standing up for specific strong viewpoints is a good thing.)
The big difference between Genshi and Jinja (which ended up winning) is that Genshi templates are valid XHTML using the "py:" XML namespace for control flow, and Jinja templates use their own syntax (and therefore can template things other than XML).
The immediate thing that caused me to start looking for a new job, most recently, was that I had a performance review / bonus conversation that went poorly (for me). It's a little hard to distinguish whether the thing that made me unhappy was getting the feedback of "We don't value the things you want to do with your career" or getting the feedback of "and, therefore, we will dock your pay." (In fairness to them, it's not clear the things I wanted to do with my career were in fact valuable to them.)
However, once I started looking, I discovered I was being seriously underpaid, and I really should have already been looking for new jobs. There's just a bootstrapping problem here - you don't know your market value until you start getting offers, and you have to start looking in order to start getting offers.
With new norms about talking about pay (this very website is a good spot, for software engineers) and sites like levels.fyi, I suspect I would have come to this conclusion a little earlier, on the basis of money. It's hard to estimate your market value but you now have a better idea of what the ballpark is.
(I suppose the other option is to continually be "looking," whether or not you have an active desire to leave. I know some people do that. But I personally find that exhausting.)
https://www.nature.com/articles/d41586-020-02550-8 is a brief Nature post from Peter Tsai, who invented this electrostatic filter. In it he says that the charge are "permanent," and adds:
> They wanted to sterilize the masks for reuse without damaging them or destroying the electrostatic charge. I knew that heat would not alter the charge, but that alcohol would erase it.
The N95DECON project says that masks can be reused if stored at room temperature and do not lose efficiency: https://www.n95decon.org/publications#time
I see some studies that find that various decontamination methods (high heat, various disinfectants) can cause the charge to dissipate, e.g., https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7724761/ - but I'm not advocating doing any of those, I'm just advocating leaving it hanging on a hook by your door when you get home. And, interestingly, those studies show that while filtration may drop below 95%, it's still pretty high, so I think that yes, a discharged N95 is still more effective than a cloth mask. But it'd be good to find a direct study.
I'm having trouble finding a source that says the filtration performance drops after a day. I did find something that says that the N95 standard requires it must keep working for at least 24 hours, but that's the other direction.
(And, anecdotally... the masks I ordered are shipped stacked together, not individually wrapped, in a cardboard box. They've presumably been in the box for a while at the warehouse. While it's true that cardboard is an insulator, so is air, and I'm not sure that the mask hanging by my door is losing charge any faster than the mask sitting in the box.)
One thing that does seem to be true is that getting the mask wet can cause the charge to dissipate. If I'm wearing my mask in humid conditions and biking (I usually leave my mask on if I'm stopping frequently), I'll notice that it gets damp, and I do change it when that happens. If your home is extraordinarily humid, maybe put your mask in a brown paper bag or something.
Price vary, but they run about $2 a mask give or take a bit. You can generally reuse them for a while, assuming you don't work in a hospital - the usual reason to dispose of them is if pathogens are on them, but if you're only using it for protection against covid, the virus dies quickly when outside a host. I don't go out a whole lot because I work from home so I usually use one mask for at least a week, often several weeks. If you're going out frequently, I've seen a recommendation to rotate through a couple of active masks, giving the virus a few days to die off.
The so-called "a.out format" is the format used by the very first tools to generate a.out files. Compilers still generate files named a.out if you don't specify a specific name, but they're in some other more useful format these days. https://en.wikipedia.org/wiki/A.out
https://astralcodexten.substack.com/p/ivermectin-much-more-t... looks at a bunch of ivermectin studies, and reaches the weird but entirely believable conclusion that ivermectin improves outcomes in parts of the world that are heavily affected by parasitic worms - the sort of thing that ivermectin is actually well-documented as treating!
Perhaps the phrasing that many present-day pundits would find recognizable would be - the people whose lives were saved by ivermectin weren't dying of covid. They were dying with covid, of worms.
This is an excellent reason to fight hard for increased ivermectin access to anyone who might be at risk of parasitic worms, covid or no covid, and to look to other countermeasures (pharmaceutical and non-pharmaceutical) against covid.
(One of the things you get in Delaware and many many other meatspace jurisdictions is the doctrine of contra proferentem, i.e., ambiguous clauses are read to the detriment of the person who wrote the contract and made the other person sign it. In this case, there is clearly an ambiguous clause, at least in the sense that the parties disagree on it, and clearly one party wrote it and the other party had no ability to negotiate it. This rule incentivizes people to a) write unambiguous contracts and b) not play stupid games where they try to invent interpretations of their own contract so they don't have to do the thing they said they would.)
Anyway, to the meta-level of the arbitration / escrow - if I'm reading right, the only reason it went to arbitration is Kleros voluntarily put up 50 ETH, the entire amount under dispute, into a new cryptographically-binding arbitration contract, and the two parties had a verbal agreement that the result of that arbitration would end their dispute?
Which is to say, if Kleros said "We're happy to go to arbitration but we're not going to lock up 50 ETH while it's being disputed," the process would be stuck? And especially if Kleros said "There's no reasonable argument here, we're right," arbitration wouldn't actually be binding on them? (It seems they only entertained the dispute to show off their arbitration product....)
And also, if it went to arbitration and the claimant lost, nothing would prevent the claimant from saying "I never agreed to this, and I still have a dispute with you"?
Can you encode "and we have recourse to the Delaware courts" into a smart contract? I thought you could only encode statements about things that happen on the smart contract's blockchain, but I'm not very knowledgeable about this.
If you do not have access to real-world courts, in the case of fraud, how do you recover your money? That is, suppose that investor A invests a bunch of cryptocurrency in pseudonymous founder B according to a smart contract, for a certain project, and B uses the funding to hire engineer C to help with the work. However, B and C are actually the same entity, who had no intention of doing any work at all, and B pays a generous "salary" to C and then comes back to A a few years later and says "Yeah we ran out of runway, we're gonna shut down, sorry" and A says "Yeah it happens to the majority of startups, it's fine." Then A hears a rumor about what really happened. How does A go about substantiating that rumor and getting their money back?
The only solution I can see involves B putting up collateral in the smart contract equal to the amount A invests in B, and that can only be released when some decision is made some years later that indeed B was non-fraudulent (let's imagine, for the sake of argument, there was a smart-contract court of chancery that could evaluate whether B was working in good faith and had subpoena powers, although I'm also curious about how that could work). That way, in the case of fraud, A could recover the amount they invested because exactly that much money is locked into the contract for this very purpose. But the market for venture capital where the founders already have as much money as they are raising, and are willing to lock up use of those funds while they work on their startup, is extremely small - such founders are necessarily capable of bootstrapping themselves.
So I'm genuinely trying to understand how it can be done or better yet whether anyone has done it. (Smart contracts are generally public, right? So if this has been done, it should be visible?)
To be clear - I absolutely agree that these are all important and these powers need to be exercised wisely. But without democratically controlling them, every single one of them is controlled by management/HR. Why do you feel that management/HR is going to do a better job of controlling those things?
So you need to separate the more conceptual parts of unions - that maybe we should value sustainable and secure employment over shareholder value, that laborers should be able to negotiate collectively and not individually, that you can do good work that is beneficial to the public by prioritizing happy laborers - from the very specific implementation of unions that we happen to have. Without our current model of for-profit businesses, the implementation of unions would not be a coherent concept (and admittedly they are a bit less coherent now and when applied to e.g. tech companies than they were in the past for the businesses for which they were designed - which isn't to say they're the wrong answer, just that they're an awkward fit). But the values are still valid.
As a sibling comment said, worker-owned cooperatives are one model here. Another similar option is very small companies: the concept of a solo founder unionizing makes no sense, but the solo founder of a bootstrapped "lifestyle business" has most of the benefits one generally wants from unions, and communities of such businesses can choose to federate for a better voice in the market. There are certainly a lot of other options besides the idea of a large business owned by its founders or publicly traded.