Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

We Could Fix Everything, We Just Don't

erikmcclure.com

Post reply on HN
Print view View original

141–150 of 421 posts

Re: We Could Fix Everything, We Just Don't

#141

Post by tempestn »

smsm42 wrote: post #79

It is adorable when a person who knows a lot about some things decides that he knows everything about all things, all problems are simple to solve, just do this and that, and the reason all problems in the world aren't solved yet is because all other people just stupid and for some stupid reason refuse to do obvious things that are so clear to this person. I guess this is where the different between a lot of knowledg…

I think you have some valid criticisms of the article, but they'd probably be better received without the condescending tone. (For example, calling someone "adorable" sarcastically is guaranteed to elicit defensiveness IMO.)

Re: We Could Fix Everything, We Just Don't

#142

Post by nolongerthere »

smsm42 wrote: post #59
grecy wrote: post #38

Earlier quoted context omitted.

> the question is at what cost? The world will be a wonderful place when schools have all the funding they need and the Air Force must hold a bake sale to pay for new fighter jets. The US spent $8,000,000,000,000 on the "war on terror" [1], while 21% of adults are illiterate [2] [1] https://www.brown.edu/news/2021-09-01/costsofwar [2] https://www.crossrivertherapy.com/research/literacy-statisti...

You are implying the cause of illiteracy is the lack of funds. It is not so - in fact, the very worst schools often spend per pupil the same as the very best ones, or at least very close to it. Throwing money at the problem is not the solution - it would cause the worst schools spend even more and keep churning out ignoramuses. The bad schools are bad not because they can't find money to become good but because the w…

The literacy issue and lack of proper education can almost always be traced back to the local family life, specifically single parent households. Households that don’t have two present parents are unlikely to be able to afford the time to enforce school attendance and homework. If you want to fix the education problems, start by fixing the breakdown of the family unit.

Put another way, the reason some schools spend twice as much money but have worse outcomes is because the children in those schools are starting from much farther back. There isn’t a reasonable amount of money you could throw at the problem school to solve the issue.

Re: We Could Fix Everything, We Just Don't

#143

Post by maxbond »

maxbond wrote: post #81
mikewarot wrote: post #36

I'd like to point out that computer security was solved in the 1970s after lessons were learned by the Military during the Viet Nam conflict. We've blamed everything except our operating systems, which are fundamentally flawed because they don't incorporate Capability Based Security[1]. Things that are not capability based security , but people seem to think they are, include: App Permission flags on your smartphone…

Every major operating system has a capability model (eg SELinux or Group Policies). Mobile operating systems have had capabilities from the jump, but still get hacked. (That being said, capabilities are great, we should use them and they do help. If you're not familiar with this pattern, look it up, it's worth knowing.)

Re your edit, what disqualifies eg smartphone app permissions from being capabilities? As far as I can see - they literally are.

Re: We Could Fix Everything, We Just Don't

#144

Post by speedgoose »

xboxnolifes wrote: post #39
croes wrote: post #33

Earlier quoted context omitted.

>COVID was a great example of this. Was it? How many fought against social distancing, masks and vaccines? Same with climate change, how many fight against necessary measures.

A lot of the people fighting against these things are not feeling the pain. Many people in my family are some degree of anti-vaccine / masking mandates. I can guarantee you that if a single close family member ended up hospitalized due to covid, their tune would do a fast 180. All of their beliefs are based on whether someone they care about had something happen to them.

A family member died from Covid-19, and another family member stayed against vaccine and social distancing. We aren’t all very bright on every topic.

Re: We Could Fix Everything, We Just Don't

#145

Post by Towaway69 »

RangerScience wrote: post #75

Increasingly, I suspect the real value of type systems is that they force slower, "more correct" development, allowing you to "win the argument" with Management over taking the time to Do Things Right. They put a floor on code quality (albeit, IMHO, a pretty low one - you can also royally fuck up a type system) - preventing the worst excesses of rushed development.

If your neighbour is using a type less language and is making more profit than you, then your boss will question the choice of language.

Development pace is set by the competition.

Edit: and if there is no competition, the longer it takes, the longer it will take to be profitable.

Re: We Could Fix Everything, We Just Don't

#147

Post by rileymat2 »

Towaway69 wrote: post #137
rileymat2 wrote: post #62

I have some problems with the way modern software is changing, but I don’t think it is QA related, more a race to the bottom with perpetual beginner focused UX. But as far as the author is referencing QA is there a good measure of quality over time? My recollection of consumer software started 30 years ago, and it was incredibly unstable. BSOD frequently, my uptime is far far better now. Save frequently, because you…

Have ransomware attacks increased in that time? Isn't the increase in attack surfaces for ransomware actors an indication of decline in software quality? Software has become incredibly complex and interdepended.

I was trying to come up with a metric and threw that out because there is also more effort in creating ransomware and software ubiquity now.

Thought about CVE counts, but that has methodological problems too.

Re: We Could Fix Everything, We Just Don't

#148

Post by nolongerthere »

maxbond wrote: post #135
throwaway290 wrote: post #129

Earlier quoted context omitted.

Sure, but those rich elites themselves like to live well and somehow it turns out they when it comes to their own they really don't like to die or see their big prosperous families die. You can't rule radicalized impoverished populations if you yourself are dead. Letting them see that their target can retaliate in this way is not cheap.

Given current events, I don't understand how you can hold this view? We've recently seen a truly horrifying terrorist attack, and the response was as brutal as you would expect. Surely the people who planned and executed it knew that better than we do, since presumably it's their business to know such things. Surely that's a constructive proof that this notion is mistaken?

This is a fascinating take: what do you suggest the proper response would be to the current event you referenced in light of your opinion that any response will only radicalize the local population?

Re: We Could Fix Everything, We Just Don't

#149

Post by ryanjshaw »

maxbond wrote: post #81
mikewarot wrote: post #36

I'd like to point out that computer security was solved in the 1970s after lessons were learned by the Military during the Viet Nam conflict. We've blamed everything except our operating systems, which are fundamentally flawed because they don't incorporate Capability Based Security[1]. Things that are not capability based security , but people seem to think they are, include: App Permission flags on your smartphone…

Every major operating system has a capability model (eg SELinux or Group Policies). Mobile operating systems have had capabilities from the jump, but still get hacked. (That being said, capabilities are great, we should use them and they do help. If you're not familiar with this pattern, look it up, it's worth knowing.)

I don't believe that Linux capabilities are correctly named. They describe an ambient privilege, whereas capabilities represent explicit authorisation i.e. one invokes a capability to use it, and this is why entire classes of security vulnerabilities aren't possible in true capability-based systems.

Unfortunately capabilities have long been misunderstood as equivalent to access control lists, and therefore ignored or misnamed due to the difficulty of building working capability-based systems Vs ACL-based systems. (ACL-based systems should be built on top of a capability-based core, and only apply to a smallest possible domain of objects - a single global namespace with an ACL-based access control mechanism is where the common security vulnerabilities of modern operating systems begin.)

Smartphone permissions are also ambient privileges, but that's not the reason they get hacked: they get hacked because the operating systems are not designed around capabilities from boot and up, and have shared critical global namespaces with an ambient authority access control mechanism. Capabilities provide defense in depth, they don't make a system automatically unhackable but they do make it so that a critical attack path is highly unlikely to exist.

Post reply on HN