HNHacker News
TopNewBestAskShowJobs

roganartu

706 karma · joined April 2, 2014

Infrastructure Engineer at Hudson River Trading

Place where I write about reliability, programming, and other things: https://www.tonylykke.com/

[ my public key: https://keybase.io/roganartu; my proof: https://keybase.io/roganartu/sigs/LyULghKURMpLPigIRAq_rQZWO_NjjkZESA4Mr_TKXYI ]

submissionscomments
roganartu··on John Giannandrea named to Apple’s executive team
I find this absolutely fascinating when I compare it to my personal experience. It begs the question: how frequently does the average person Google something compared to power users? I consider myself to have above average googling skills, and I'm a developer so there's obviously a lot of documentation searching regularly, but even excluding all of that I surely perform a good 10-20 non-programming searches a day and it could easily be an order of magnitude more if I'm actively researching something.

Does the average user perform significantly less searches, and so the novelty or occasional voice search moves the needle 20%, or are they performing as many searches as me but using voice for many of them. I personally only ever found voice search useful for things that are more like questions and not research ("how old is _______" is a classic example) so I find it difficult to believe the latter. The former would be quite the revelation though because I always assumed _everyone_ googled as much as I do but it seems that might not be the case.

roganartu··on Google One
I had to switch back to Spotify a while ago to give my wife access to the premium I was already paying for via the family plan because Google Play Music doesn't support family plans for gsuite accounts.

If they addressed this issue I would probably consider consolidating Dropbox, Spotify, and maybe some other stuff into a single Google One subscription but I suppose grandfathered free gsuite accounts are a low priority for these things.

roganartu··on Robinhood launches 3% checking account
The coin counting machines in the US closed a couple of years ago due to a class action suit for chronic short-changing: https://www.bankrate.com/finance/banking/banks-expel-coin-co...
roganartu··on The CRISPR Baby Scandal Gets Worse by the Day
You appear to be conflating ethics and legality. It is perfectly legal (for now) for social media giants to spy on their US users in increasingly intrusive ways, but many would argue there are lines in there that cross into unethical. Perhaps legislation will follow to enforce this ethical stance over the coming years as has happened in the EU, but the absence of such legislation does not preclude it from being unethical. The objection to Google's interest in the Chinese market is another example where ethics and legality don't match up.

Anecdotally, I personally find it ethically indefensible to abort a child due purely to sex, despite otherwise being pro-choice.

roganartu··on Insomnia: 'No link' between sleepless nights and early death
I had a similar experience after a week long camp where I only slept a couple of hours a night, if that. On returning home I went to sleep and distinctly recall waking up about an hour later, only to discover I'd actually been asleep for ~25 hours.

I never experienced euphoria then, but in my younger years when I could still successfully pull all-nighters I found myself re-energized once the sun rose in the morning and this lasted typically until later that afternoon. This lead to a general rule where if I was planning on staying up that late I always tried to go to bed before the sun came up.

roganartu··on Jason Fung explains why intermittent fasting diets work
This has been posted before, but a good digestible source for the information in this comment is this two-video podcast series: https://www.foundmyfitness.com/episodes/valter-longo

The studies you mention results of (like the 30% immune system culling in a 4-5 day fast) have only been published for mice so far afaik, but they're currently doing human studies.

Also worth noting that one of the points Valter Longo makes quite strongly in these podcasts is that refeeding after the fast is at least as important as the fast itself.

roganartu··on Go hits the concurrency nail on the head
> Finally, Go doesn't do anything to prevent data races, which are the biggest problem facing concurrent code. It actually makes data races easier than in languages like C++, because it has no concept of const, even as a lint. Race detectors have long existed in C++ as well, at least as far back as Helgrind.

While it can't guarantee correctness outside of runtime (and even then, obviously only if whatever you are running actually triggers the race), a race detector has been part of the Go core since 2012[1].

[1]: https://blog.golang.org/race-detector

roganartu··on In Georgia, a legal battle over electronic vs. paper voting
Counting paper votes is a solved problem, and nations that use entirely paper ballots don't have slow preliminary counts. Australia, for example, has a somewhat complicated preferential voting system and relies entirely on paper ballots, and yet they are generally able to provide a conclusive result the night of an election (with televised tracking of counts throughout the country).

This may not make sense intuitively, but counting paper scales quite well by being distributed to the polling locations and physically sorting the paper ballots. The Australian Electoral Commission documents the counting process quite thoroughly [0].

The only time counting paper takes a long time to get a win/lose result is when results are close enough that every vote needs to be counted. This is surprisingly rare in practice.

[0]: https://www.aec.gov.au/Voting/counting/index.htm

roganartu··on Scrapping Contracts: a critique of the Go generics proposal
This is complicated by the fact that a struct method can have a pointer or a non-pointer receiver. They can't be mixed, but consider the following:

    type Foo struct {}
    
    type Bar interface {
        Baz()
    }
You could implement the interface either with a pointer receiver:

    func (f *Foo) Baz() {...}
Or with a non-pointer receiver:

    func (f Foo) Baz() {...}
Pointer receivers are more common for various reasons, but non-pointer receivers are still perfectly valid.

If you force interface arguments to be prefixed with a * then you end up with confusing usage depending on how you initialize the variable being passed and whether the interface was implemented using pointer receivers or not.

eg, for your ReadStuff example above, there are the following possibilities:

    reader := SomeReader{}
    pointerReader := &SomeReader{}
    
    // SomeReader implements io.Reader with a pointer receiver
    ReadStuff(&(&reader)) // This isn't actually valid, compiler will throw "cannot take the address of &reader"
    ReadStuff(pointerReader)
    // SomeReader implements io.Reader with a non-pointer receiver
    ReadStuff(&reader)
    ReadStuff(*pointerReader)
roganartu··on Air-conditioners do great good, but at a high environmental cost
By "purely in the room" I imagine they meant within a closed system. By having a hose connecting your units to the outside world your room is no longer a closed system (ignoring things like drafts and imperfect insulation etc that already prevented it from being considered closed).
roganartu··on Spyware Company Leaves ‘Terabytes’ of Selfies, Messages, Location Data Exposed
If we take the Equifax breach as an example here, and every person affected sued for even a very modest amount (say $10) and everyone wins their cases then we come to something like $1.5bn in damages, plus legal costs. Sure, that will probably work in that $10 is a gross underestimate, Equifax's market cap is ~$15bn, and their legal costs defending that many cases would be significant.

However, consider that it is very likely a small minority of those ~150m affected people are actually in a position to spend the time, money, and effort in actually suing and you end up in exactly the position you are now: Equifax doing fine and suffering no penalty for their actions. Class action suits aren't really a better suggestion either because they are typically settled for pennies-on-the-dollar, with the lion's share going to the lawyers anyway.

Suing might make sense where there's a small number of affected people, or where the damages per person are much higher, but when we're talking less than $1,000 damages per person it's really just not worth each individual's time or money to do so. This is _exactly_ the kind of thing regulation is good at protecting against.

roganartu··on Reddit Security Incident
The SMS interception via social engineering of telecom support staff, as others have pointed out, seems most likely, but consider another approach: an app on the users phone with message read permissions. Most people are not diligent enough to perform an audit of the permissions requested by every app they install and I could also believe a determined attacker might install an app on an unattended and unlocked phone given the opportunity.
roganartu··on Australian firm Meriton fined $3M for misleading consumers on TripAdvisor
You'll find that replacements for change of mind in Australia (and perhaps other countries with strong consumer protections) are less friendly. I have definitely noticed there are more stores in the US (especially big box retailers that are competing with Amazon) that have flexible return policies in the short term for changing your mind, while my expectation for any store in Australia is that you are likely unable to return something for change of mind if it has been used.

The trade off for that is that you get very long periods of fitness for use, where you can comfortably buy something that is advertised for a purpose and know that you can return it after a reasonable time if it isn't actually fit for that purpose. A good example is smartphones and laptops, where Australian-sold iphones effectively have a two year warranty and macbooks have a three year one iirc because of these strong consumer protection laws.

It's these kinds of faulty goods returns to smaller retailers that I've had most trouble with in the US (and where I think consumer protection makes the most sense). In Australia the consumer would simply file a complaint online to the Department of Fair Trading (or whatever it is for their state) and these are usually resolved within a couple of weeks or less and usually in the consumer's favour. I'm not aware of any equivalent in the US, and if it does exist I'm sure it varies significantly state-to-state.

roganartu··on Australian firm Meriton fined $3M for misleading consumers on TripAdvisor
The ACCC and strong consumer protection laws are something I didn't realise I took for granted until I moved to the US recently. I had a pair of pants develop a hole within a few weeks of purchase so I took them back to the store and it was difficult to convince them that pants should last longer than that and get a replacement. This is in contrast to my Fitbit, for example, which I had replaced about three times in two years without question in Australia because of a fault in the band.

It's interesting that capitalism comes into it somewhat, insofar as companies like Amazon and REI have consumer-friendly return policies as a sort of competitive advantage. I know I've bought things from REI despite a higher price just because I know I won't have any problems returning it for any reason within 12 months.

roganartu··on Mea culpa: there is a crisis in the humanities
I see what you're saying, but I think it's safe to say they were suggesting it is expensive to the individual if they get sick, not to their entire society. The two are very different problems.

Yes, the burden of healthcare costs in places with universal healthcare are spread across all taxpayers, but for that you get a society that can be reasonably confident that getting sick won't bankrupt them (especially if your socialised healthcare nation also has good welfare).

It's easy to dismiss as a "won't happen to me" kind of thing, but in reality it can happen to anyone. A couple of years ago my wife was suddenly admitted to hospital for a couple of weeks and then physical therapy for 6+ months and it didn't cost us anything except her lost income. In the US that could've easily run us several hundred thousand dollars with bad or no insurance. There's a reason medical bills are the leading cause of bankruptcy in the USA.[0]

[0] https://www.fool.com/retirement/2017/05/01/this-is-the-no-1-...

roganartu··on Content Snare: Leaving a job to build a SaaS business
> average rents in Brisbane/Sydney/Melbourne are about AUD$2600/month (~USD$1900)

Melbourne and Sydney maybe, but Brisbane is drastically cheaper. When I was studying in Brisbane a few years ago my rent was in the ballpark of $300-400 AUD per week for 2-3 bedrooms, even at the higher end of $400/w that's only $1700AUD/m. You can even buy a two bedroom apartment for under $300k in Brisbane, compared to the $1m or more the same would run you in Sydney or Melbourne.

In contrast, when I graduated and moved to Sydney for work we lived a similar distance from the city as we did in Brisbane and yet were paying $500/w for 1 bed and much less total room.

I suspect Darwin, Perth, and Hobart are similarly cheaper. Generalising and saying all our capitals are expensive is incorrect, though I can understand why perceptions are driven by what happens in Melbourne and Sydney since nearly 50% of the population lives in one or the other.

roganartu··on Slack outage: Connectivity issues affecting all workspaces
The reality is this is how you design highly available systems, and it is also imo one of the reasons microservices have gained so much popularity.

Driving features with microservices makes it easier to isolate their failure and just fall back to not having that feature. The trade off is that monoliths are generally easier to work with when the product and team are small, and failure scenarios with distributed systems are often much more complex.

An analogy to your Google failure examples for slack might be something like the "somebody is typing" feature failing for some reason. In an SoA you would expect it to just stop working without breaking anything else, but one could easily imagine a monolith where it causes a cascading failure and takes the whole app down. Most services have countless dependencies like this.

roganartu··on Losing 100 pounds in 276 days
No calories at all, just drink water when we're thirsty.

I've found it has worked wonders for resetting my relationship with hunger; I feel I'm far less susceptible to boredom hunger in the few days immediately after the fast, something I'm hoping will progressively improve further.

Honestly, we haven't found it that hard at all. The first two weeks were a bit of a struggle just because you aren't used to going that long without food, but now it's just the day we don't eat and we don't really pay it much attention past that.

roganartu··on Losing 100 pounds in 276 days
Another approach that some people might prefer is a 24h once a week fast, which I've found much easier than 12-16h daily.

I don't eat breakfast per se, just a protein shake but I found it difficult to cut out for a number of reasons.

A couple of months ago my wife and I started doing a weekly 24h fast on Sundays from breakfast Sunday to breakfast Monday. The first couple of times it was a bit of struggle but now we don't really notice anymore.

Whatever fits your lifestyle is the way to go, otherwise it's always going to be a struggle to maintain.

roganartu··on America’s Millennials Are Waking Up to a Grim Financial Future
Until you realise that those who care the most often hold extreme views and are the most outspoken even when representing a relative minority. Deeply conservative fundamentalist Christians, anti-vaxxers, and climate change deniers are some examples that come to mind.
roganartu··on America’s Millennials Are Waking Up to a Grim Financial Future
> So, yes, one party has made it a mission to suppress turnout; yes, Election Day is a work day, which makes it harder for working people to vote; etc. -- we can change all that, but first we have to work in the system as it is today.

This is why non-compulsory voting hasn't ever made sense to me; it effectively biases the system towards those who care the most and enables those same people to influence the system further by making it difficult enough for others that they won't bother.

When it's compulsory to vote, things like weekday voting stop happening because laws will be introduced making it illegal to prevent your employees from taking time off work to go vote, so naturally both employers and employees alike want weekend voting.

Additionally, it incentivises the government to make voting accessible, because it is unreasonable for them to enforce voting via fines etc if it's not really possible for people to vote.

When I moved to the US from Australia I heard stories from my new colleagues about people in poor neighbourhoods lining up literally all day just to maybe get to vote before the booth closed, while those in richer neighbourhoods had practically no line at their voting booths. This is not how democracy should work. Your "freedom to choose not to vote" only serves to hurt any other freedoms you may have had.

roganartu··on Stopping the RSI pain that almost destroyed my programming career (2016)
Starting Strength and Stronglifts 5x5 are probably the two most popular. These programs are based on what's called beginner gains. You wills start with what will feel like a very light weight, but you increase it every workout until you can't lift it anymore. You'll be amazed at what your body can do after 6 months of this.

When the parent said compound lifts they were referring to lifts that (when you are doing them heavy and correctly) are whole body exercises: Deadlifts, squats, and bench are the three main lifts. Programs will typically mix in some other lifts to round out the program a little like power cleans, overhead press, and bent over rows.

Most beginner strength programs are three days a week, I've found most people do Monday, Wednesday, Friday, but it's flexible as long as you have at least one rest day in between workout days. If you don't mess around and just go to the gym to work you can easily be done in 30-45 mins so under 3 hours a week is very realistic.

You will absolutely feel uncomfortable when you start. Everyone does. Read a lot about form and pay attention to yours (gyms have lots of mirrors, lift near some especially when you're starting out). It's really easy to hurt yourself once the weight increases if you have bad form. Consider getting a good personal trainer to teach you the fundamentals of the lifts.

roganartu··on If Fortnite Were a Website, It Would Rival Reddit and Amazon
Anecdotally, the process to get to Facebook for many members of my family is to Google "f" or "facebook". They know how to go directly there but for whatever reason prefer Googling it. I have heard the same from friends about their (typically older) family too.
roganartu··on Patent trolls are shaking us down for selling “rubies” in our game
Insurance doesn't mitigate damage, it amortizes it (typically across a group of entities). The expected average annual claims on an insurer's policies need to be lower than the sum of the premiums or the insurer loses money.

This is why people in high risk categories pay more for insurance (eg young people and people with high performance cars have higher car insurance premiums because statistically they crash more often and/or have higher rates of total loss)

roganartu··on FCC report finds almost no broadband competition at 100Mbps speeds
The flakiness is a byproduct of constantly degrading last-mile infrastructure that is prohibitively expensive to maintain.

With the exception of FTTP installations that were completed before the NBN[0] became a political football, the majority of consumer internet connections in Australia have a copper last-mile serving ADSL2+. Flakiness with this copper is caused by general network neglect (due to unsustainable maintenance costs) resulting in things like contractors wrapping tangled messes of cables in plastic shopping bags to keep them "dry"[1].

Anecdotally, several ADSL2+ connections I had over multiple years and residences in Australia would noticeably degrade for a week or two after heavy rains. I know others who experienced similar behaviour.

[0] https://en.wikipedia.org/wiki/National_Broadband_Network

[1] https://www.google.com/search?q=telstra+australia+plastic+ba...

roganartu··on The growing body of evidence that digital distraction is damaging our minds
K prisms You're at a restaurant with your kid, and they're bored because you're having adult conversations with your friends.

Call me old fashioned, but my parents just used to make us sit there and be quiet. Bored? That's life, learn to deal with it.

Both the parent and the child in your example are lacking discipline. The parent unable to tolerate a bored child and the child unable to tolerate boredom. This is a pattern I have personally noticed as early as babies being pushed around in prams with a smartphone or tablet in hand. A two year old definitely doesn't need a phone to relieve boredom but it's easy.

roganartu··on What I learned in 2017 writing Go
There is a detailed explanation of how package names are used by the compiler here: https://golang.org/cmd/go/#hdr-Description_of_package_lists

Specifically this bit is relevant here: > Every package in a program must have a unique import path. By convention, this is arranged by starting each path with a unique prefix that belongs to you. For example, paths used internally at Google all begin with 'google', and paths denoting remote repositories begin with the path to the code, such as 'github.com/user/repo'.

Most (all?) Go tooling assumes this convention. Conceptually, I don't see it as that different from how other languages determine input paths, except in Go they are guaranteed to be unique which I consider a good thing. In Java for example you still have file namespaces they're just inside the project itself (looking at you `src/java/com/example/package`) or Python which looks for the root module name in `PYTHONPATH` and then recurses down for submodule names.

Most if not all languages use file structures to namespace imports in one way or another, Go just applies it globally instead of at a project level. I agree it raises the barrier to entry for someone to just come in and run `go run` on a project but I find it works very well once you've got it setup. Convention over configuration is a big thing in Go tooling.

roganartu··on Your Mother’s Maiden Name Is Not a Secret
> (nope, Transferwise doesn't have it)

fwiw Transferwise has had this for around 3 months now: https://transferwise.com/borderless

roganartu··on The fake Facebook profile industry
That is the porn-equivalent of suggesting that abstinence is the solution to teen pregnancy. It may be technically correct insofar as abstinence obviously works for those that rigorously follow it, but fails to account for the simple fact that teenagers have sex regardless of whether they are told not to.

Shaming porn usage does nothing to reduce its prevalence and just encourages users to hide and deny, hence your suggestion is not productive.

roganartu··on U.S. regulators accuse Palantir of bias against Asians
> As for "freezing up during a stressful interview": is that not some indication of what normal workplace stresses will do to your ability to complete technical work? I certainly wouldn't want to work with somebody who is knowledgeable but will wimp out the moment you ask them a tough question on the spot.

I would argue that a workplace that is continuously stressful is not healthy, and optimising your interview process to filter for this is probably not a great idea.

Additionally, I would expect that when asked a "tough question" at work that you would have both the time and resources available to find or determine the answer, things that one likely doesn't have in an interview (and certainly not to the same extent).

← PreviousPage 4 of 5Next →