> It seems like engineering guidelines, if composed carefully, would help solve at least some of these problems.
That's now how this works. It's hard to know how to design programs, it comes from experience, you are paying your engineers to apply that expertise. There isn't a formula for all problems.
Every large software company I have worked at simply accepts this, and allows engineers to do solve the problem the way they want, while requiring interfaces between teams/software to work properly.
Besides that, there are a few approaches you can take to share knowledge:
1. Limit the number of technologies used in your stack. You just listed 4 languages that all fill a similar role. If that was 1 that would reduce a lot of the cognitive load.
2. Cross pollinate engineers so that they get to work on all parts of the code.
I don't think "moral reassurance" has anything to do with it, at least in the hard sciences. The universe will smack you hard with reality if you're grifting, there are limits to lying in this space (at least). For example, you can't lie about creating full self-driving because it simply won't work in all the conditions/scenarios under which we expect human-like performance.
> a medical researcher at the University of Texas, found that only 11% of 53 pre-clinical cancer studies had replications that could confirm conclusions from the original studies
> A survey of cancer researchers found that half of them had been unable to reproduce a published result
> A 2016 survey by Nature on 1,576 researchers who took a brief online questionnaire on reproducibility found that more than 70% of researchers have tried and failed to reproduce another scientist's experiment results (including 87% of chemists, 77% of biologists, 69% of physicists and engineers, 67% of medical researchers, 64% of earth and environmental scientists, and 62% of all others
How many of those papers were retracted, you think?
The vast majority of "failed replication" has nothing to do with fraud or false findings, and everything to do with "You have to write what you did to get this result that other people can follow, but you aren't really incentivized to be comprehensive and clear and note down every possible contributor so nobody can recreate the reaction you had in your paper because you didn't write down that you did it in an especially cold lab"
I am so confused why HN is so tolerant of startup grifting, but dismisses cryptocurrency as an absolute evil. Are concepts like proof of work not even interesting?
I would rather have fast and cheap online payments than a chat bot I have to twist around until I can talk to customer service.
> Are concepts like proof of work not even interesting?
Sure, but we had that with SETI@home, and then Folding@home, and they were doing something useful with their work rather than just finding which random numbers hashed a certain way in order to get a token whose primary economic value was the assumption that "remove trust" is a useful thing in economics.
We also had a blockchain of sorts before bitcoin, and still regularly use it: it's called "git".
And we already have fast and cheap online payments. In my case this was initially in the form of paypal, then my banks got good at it more directly; in Kenya since 2005, and now more countries in the region, there's a thing called M-PESA which does much the same but works with pre-feature-phones: https://commons.wikimedia.org/wiki/File:M-Pesa_on_Nokia_1100...
but proof of work / crypto isn't fast, and nothing makes it particularly cheap.
If we are talking fast and cheap online payments, micropayments instead of subscriptions, let's go. Too much of crypto is a solution looking for a problem (and finding ones that don't fit).
> Are concepts like proof of work not even interesting?
PoW only addresses a single problematic aspect of this stuff (and only reduces it, doesn't actually resolve it). It does nothing about all the other problematic aspects.
> deregulation lead to the consolidation of airlines
Explain. I see a handful of identical mega corps with a government protected monopoly (regulations + access to airports). Hasn't regulation increased consolidation to share the cost of compliance?
Like the pork barrel shops in the airport, why is this a private business at all?
I presume they are referring to the airline deregulation act of 1978. There used to be dozens of regional airlines whereas now we notably have 3-4 giant corporations after decades of aggressive mergers and acquisitions.
You’d be surprised but there are still regional airlines. This is because a company like delta franchises some routes basically. You might go on a delta flight and ride on a delta plane, but the operating company is some almost unheard of regional one.
I’d guess also you wouldn't have to handle the maintenance end of these smaller planes eg. your bombardiers and what have you if you’ve spun off all those flights
>I see a handful of identical mega corps with a government protected monopoly (regulations + access to airports)
I see the opposite: new, brightly-colored airlines seem to pop up every year, each offering substantially the same thing: sub-$100 direct tickets to Florida (and probably other) destinations from low- and mid-tier airports. And they're all catering to the people who these rewards programs are shedding.
GP is referring to the deregulation of the late 1970s. Before that, there were a large number of smaller regional airlines in the US, that have since mostly disappeared.
A 15 minute human review process that checks if it:
- is copy pasted from another source or the internet
- respects copyright
- is written in the language it says it is
In graduate school nobody can understand each other's papers, so the colleges set up a massive list of formatting guidelines. The idea being the focus on formatting means you at least had to spend a minimum amount of effort.
A 15-minute human review is not going to be able to check all three of those things, or really do much of a job on even either of the first two in isolation.
I wouldn't say people in the west are "better people in their hearts" but they absolutely follow more strict norms regarding honesty and theft. This is one of the main reasons why companies pay a premium for workers in the west when they could hire from other places.
One example: accounting scandals in the US are rare. I don't think anyone trusts accounting figures for public companies from India or china.
> in which case I'd like to see some evidence.
The alternative to trust is enforcement. The evidence you are looking for is the prevalence of the latter principle in systems that deal with things of value.
>One example: accounting scandals in the US are rare. I don't think anyone trusts accounting figures for public companies from India or china.
I work in accounting. You’re right, I trust the US more than India or China, but you’d be surprised at the liberties US companies make and how many individuals from India are auditing their work. Auditors (excluding partners and some senior managers) are just not equipped to deal with the technical accounting concepts and to challenge management. Remember, the company employs the auditors. You certainly don’t want to ruin a $1MM contract for your firm but pressing too hard.
> You certainly don’t want to ruin a $1MM contract for your firm but pressing too hard.
Only in the west would you get any push back at all. The auditor would feel some duty to bring up an issue even if it reflected poorly on the company and even their own managers.
I agree they likely wouldn't push an issue beyond its welcome, but this particular value is unheard of in many parts of the world.
> This is one of the main reasons why companies pay a premium for workers in the west when they could hire from other places.
By this token "other places" companies would also want to pay a premium to hire US workers outside of sheer competency. Yet we're not seeing Samsung massively moving institutional operations to US centers for instance.
> accounting scandals
Whait, what ? You're telling me that while the crypto bubble is bursting and they're going to prison for egregious fraud ?
Also, scandals being few in number would probably be a sign of overcorruption and systematic rot or the controlling structure. I'm not sure that's what we want.
> Yet we're not seeing Samsung massively moving institutional operations to US centers for instance.
Culture matters a lot. Korea is a high context society, and relationship building is extremely important. Its very difficult to migrate functions to other locations when the way to get things done is through building trust and relationships over long periods of time. That said, Samsung does have significant offices all over the world.
Counterpoint: scihub, libgen links are routinely shared on HN as a matter of course (likewise in academia). I've seen HN threads in which people unapologetically reminisce about torrenting movies/music.
Perhaps your will object: "But the publishers/Disney/etc are evil, greedy entities and I don't owe them anything." But I'm sure anyone who's stolen corporate secret elsewhere can come up with a similar justification in their head! After all, that may well be why they left the company in the first place.
Instead of `m4` or `sed` find and replace, the author should try `envsubst`. It's a program that replaces bash style variable references (for example `$TITLE`) with their value in the environment.
I agree that `envsubst` is a good choice for this. Unfortunately, it is not part of posix, so you can't rely on it being present everywhere. But as part of gettext, it is still very common.
That's now how this works. It's hard to know how to design programs, it comes from experience, you are paying your engineers to apply that expertise. There isn't a formula for all problems.
Every large software company I have worked at simply accepts this, and allows engineers to do solve the problem the way they want, while requiring interfaces between teams/software to work properly.
Besides that, there are a few approaches you can take to share knowledge: 1. Limit the number of technologies used in your stack. You just listed 4 languages that all fill a similar role. If that was 1 that would reduce a lot of the cognitive load. 2. Cross pollinate engineers so that they get to work on all parts of the code.