HNHacker News
TopNewBestAskShowJobs

charles-salvia

618 karma · joined December 29, 2016

I develop software that works sometimes.
submissionscomments
charles-salvia··on Why Facebook Keeps Beating Every Rival: It’s the Network
Yeah, in the 90s Microsoft was basically this unstoppable juggernaut that ruled the Desktop with an iron fist. The main cause of their decline was that a new market appeared which they failed to capture - namely, the mobile market (as well as the server market that became dominated by Linux). For Facebook to suffer a similar fate, I would think some new type of market would need to appear that Facebook fails to capitalize on - but it's hard to think of an analogous "new market" that could appear that would fundamentally displace a social network in the way that mobile displaced/disrupted Desktop. Maybe VR? (Although Facebook is trying to get into that as well...)

Still, from my perspective, Microsoft in the 90s seemed as unstoppable and ubiquitous as Facebook does today - actually, it seemed more unstoppable than Facebook, because back then we weren't yet used to the possibility of scrappy startups turning into billion dollar companies overnight.

charles-salvia··on How many jobs really require college?
I experienced pretty much the exact opposite of what you're saying. Most of the skills I have I learned by researching/studying by myself. College was mostly useless to me. And I'm much more productive when working remotely. So obviously, people are configured differently.
charles-salvia··on Saturn moon 'able to support life'
Based on the developments in exo-planet research along with discoveries like this over the past couple of years, I would conjecture that life is ridiculously common. But by life, I mean microbial life. I would imagine that multi-cellular life is much rarer, and life capable of reaching human-level intelligence and beyond is probably absurdly rare, on the level of something like one species per galaxy. It's all worthless conjecture, of course, but the point is I wouldn't be surprised if microbial life is very, very common.
charles-salvia··on [dead]
For any meeting that doesn't involve in-depth technical discussion - the kind of discussion where you might need to break out a whiteboard and start drawing nodes - I find that video conferencing is a perfectly fine substitute for face-to-face communication, as long as the connection is decent.
charles-salvia··on New research says starting university classes at 11am would improve learning
I usually just rely on drugs like... um, caffeine, rather than trying to religiously stick to a particular sleep schedule.
charles-salvia··on Mapping Strings in C++
The point is apparently to return "type_t::num_types" if the key doesn't exist in the map. Using at() throws an exception if the key doesn't exist.
charles-salvia··on ‘Negative Mass’ Created at Washington State University
The legendary "Alcubierre drive" concept for faster than light travel relies on the existence of negative mass to work. Up until now, I have always read that it was generally assumed that negative mass could not exist, making the Alcubierre drive little more than a fanciful equation. But... I guess not.

Edit: although, it's not clear to me how significant a breakthrough this really is, compared with earlier attempts to create negative mass in a laboratory setting.

charles-salvia··on The Website Obesity Crisis (2015)
Ads are definitely a huge problem, but it seems they are at least equally, perhaps even less of a contributor to bloat than the tons of off-the-shelf Javascript frameworks/libraries major websites throw in to perform analytics, or to implement stock features like user-comments, login, social-media stuff, etc.

I mean, just looking at the network resources for a fresh download of www.cnn.com, I see:

livefyre_base.js, 167KB - apparently some kind of user-comment library

cdn.optimizely.com/js/131788053.js, 93.2K - some library for user tracking/analytics or something,

controltag.js, 72K - something from a company called "Krux DMP" that seems to be yet more user analytics and tracking,

sdk.js, 61.9K - something from gigya.com, apparently some kind of login and user-preferences framework

jsmd.js, 87.5K - another analytics framework

... and various others.

I see a "pubads_impl_113.js" from doubleclick.net, which is obviously for ads, but if cnn.com is typical, then it seems that off-the-shelf Javascript frameworks for user analytics and other stock features like logins/social-media crap, etc., are just as big a factor in contributing to bloat as display ads.

charles-salvia··on High school journalists investigated a new principal’s credentials
It doesn't seem to matter as much in the technology world. When I look at people's resumes I barely even glance over the Education section. I mean, I'll take note if it's something extreme like a 4.0 at Stanford, but for the most part nobody cares anymore.
charles-salvia··on Unix Is Not an Acceptable Unix (2015)
What's funny is that some modern scripting languages like Python, with things like the "os" module, are almost passable as a general purpose shell scripting environment. I actually tend to write Python scripts instead of bash shell scripts for one-off file manipulation scripting, mostly because even after a decade-plus long career, I've never become much of a wizard at bash shell staples like sed or awk.

But it's just that a general purpose programming language like Python doesn't treat executing subprocesses in enough of a first-class manner, in the way the bash shell does, to make it feel like a seamless shell-scripting experience. You still have to go through the motions of "import os", etc. just to get started, and you can't directly execute other programs, instead having to use "system" or the more cumbersome subprocess module. But still, it's.... tantalizingly close to something like a "better" shell that includes support for lists and other data structures instead of just "everything is a string".

charles-salvia··on Interesting data structures: the BK-tree
Presumably, the idea is to count sufficiently similar images as the same image when doing the deduplication, so traditional hashing doesn't help here.
charles-salvia··on 'Hitler Tamed by Prison' (1924)
GWB had the choice of (1) start war in Iraq for vague reasons, (2) don't start war in Iraq.

He chose (1).

charles-salvia··on Interesting data structures: the BK-tree
There is ... a crude spell checker could be implemented by dumping a dictionary of words into a metric tree and using levenshtein distance (or damerau-levenshtein distance) as the distance function. The spell check would then consist of doing a K nearest neighbor lookup.

But it doesn't work as well as you expect. A metric tree requires a true metric to be used as the distance function, which means the distance function must satisfy both symmetry and triangle inequality. Levenshtein distance satisfies this - but unfortunately this usually isn't enough for a robust spell-checker that just "does the right thing". For that, you usually need something like a weighted levenshtein distance (e.g., an incorrect first character should weigh more than an incorrect Nth character, or an incorrect consonant should weigh more than an incorrect vowel). But it is much harder to guarantee that such a distance function is a true metric.

charles-salvia··on Interesting data structures: the BK-tree
There are many metric-based data structures, such as M-tree, Vantage Point Tree, etc. They all suffer from the same major problem: they all rely on triangle inequality to speed up search by pruning impossible paths. But this doesn't ever result in an easily measurable complexity guarantee (like say, the guaranteed log(N) of a balanced binary search tree), so it's very common for certain datasets to result in severe performance degradation, particularly when the data is high-dimensional.

Metric-based data structures face the peculiar problem that the identity of each data point itself is not an actual "key", in the way that a data point in a hash-table or binary tree is a key. Rather, each data point in a metric tree is only "indexed" with relation to it's distance from all other keys. This makes it hard to get easily measured performance guarantees.

charles-salvia··on Python coding interview challenges
To me this sounds like you just develop dependency-laden bloatware, i.e. you'll throw in some 5 gigabyte Javascript library just to do one thing.
charles-salvia··on After All These Years, the World Is Still Powered by C Programming
Still, the thing is, I really don't buy into any arguments in favor of starting a new project, in C, today. The problem is that C suffers from some really serious disadvantages as a systems programming language, and it has no significant advantages over C++ or Rust, apart from subjective ideas like "it's simpler" or circumstantial issues like "my team doesn't know C++".

Firstly, apart from a simple call stack, C has no automatic resource management. There are no automatic destructors. Thus it is too easy to forget to free some memory or unlock a mutex or decrement a ref count or whatever, especially when many code paths are possible after a resource has been allocated. The other options are to use the "goto cleanup" or "goto err" idiom, which is still way more error prone than automatic destructors.

And secondly, despite being touted as so "efficient", C offers no way to write code that is both generic and free of runtime overhead, apart from preprocessor macros (which are filled with their own issues), or automatic code generation scripts. Both C++ and Rust have metaprogramming facilities which provide the ability to write generic code without sacrificing performance or maintainability.

I mean, with C++, the automatic deterministic allocation/destruction alone is worth using it over C. And all of the downsides (larger binaries, etc.) can be overcome without much effort. I concede that the extra language complexity of C++ over C could be a disadvantage, but in my mind that doesn't outweigh the benefits.

charles-salvia··on Earth is a ‘dwarf planet’ because it has not cleared its orbit
This is kind of silly... this is really a problem with the human obsession with creating taxonomies and classifications of natural phenomena, when natural phenomena are basically collections of atoms that have very fuzzy boundaries. Clustering algorithms are, in general, very subjective, and it's no surprise that our scientific efforts to "cluster" everything often produce counter-intuitive or undesirable results at certain boundary cases.

It's the same problem with biological taxonomies - is a lungfish a fish or an amphibian? The reality is it doesn't matter apart from the artificial taxonomy we impose on it so that our brains can better cope with the complexity of the Universe.

Perhaps a better (but more annoying) way of classifying everything would be to use hierarchically clustered taxonomies for everything, so that we could make statements like "Earth is considered a planet if we cut the dendrogram at level 5, but a dwarf planet if we cut at level 4, etc".

charles-salvia··on Why is it so hard to study anti-matter?
Your bullshit detector is hopelessly broken. At a certain point, these "mathematical simulations" (i.e. effects observed in a laboratory) end up having real world applications. Anti-matter has already crossed the threshold of being "effects observed in a laboratory" to actual applied science. Hospital PET scans use anti-matter regularly for brain scans. [1]

[1] https://en.wikipedia.org/wiki/Positron_emission_tomography

charles-salvia··on Why is it so hard to study anti-matter?
CERN claims to have created anti-hydrogen atoms in 1995 [1], and Fermilab confirmed the claim shortly after by doing the same thing.

You're perfectly free to doubt these claims (for whatever strange reason), but doing so would be pretty arbitrary, unless you consistently distrust the entire scientific establishment, in which case, your "bullshit detector" is badly broken because it keeps reporting too many false positives and needs to be retrained.

[1] https://home.cern/topics/antimatter/antimatter-cern

charles-salvia··on Ex-Trump adviser Flynn seeks immunity for testimony in Russia probe: WSJ
No power on Earth is that competent. And the Tea Party was very clearly a reaction to Obama getting elected, and egged on by the Republican establishment.
charles-salvia··on New York Has a Great Subway, If You’re Not in a Wheelchair
It runs 24 hours and extensively covers 4 of the 5 boroughs. Yes, it's dirty, but the coverage + availability is what makes it a great system - it's certainly unparalleled in the Western hemisphere. Everytime I go to San Francisco I realize how awesome the NYC subway really is.
charles-salvia··on Software Engineer Starts Unlikely Business: A Weekly Newspaper
I agree. A lot of the complaints about "fake news" and everything aren't so much directed at actual false information being reported - but rather the selection or emphasis of certain facts over others.

An obvious example was Fox News vs. CNN during the weeks leading up to the election. Neither company outright lied about anything (for the most part), but both published headlines emphasizing an entirely different set of facts. Fox News was constantly blasting headlines about the FBI investigating Hillary's email, while CNN was constantly blasting headlines about Trump's sexual misconduct allegations, etc. Neither were really lying in any sense - but they both painted completely different pictures designed to elicit differing opinions/reactions from their readers.

Of course, if you want real fake news - that exists as well and always has before it became a buzz word. The National Enquirer has been publishing fake news for decades, and nobody cares.

charles-salvia··on The AI Misinformation Epidemic
Yeah, terms like "machine learning" and "AI" have basically become buzz words which, to most laymen, probably encourage the idea that we're on the cusp of creating Data from Star Trek.

Unfortunately, the reality is that.... sorry... it's mostly just statistical algorithms based around regression and intermediate calculus. State-of-the-art "deep" neural networks are not really anything like the absurdly parallel, asynchronous biological networks that power our neo-cortex - rather, they're basically an application of matrix multiplication designed to "learn" a function by iteratively minimizing an error value using gradient descent. It's still very unclear what, if anything, this algorithm has in common with how a human brain actually operates. It turns out that these kind of statistical algorithms can work pretty well when you have petabytes of data to learn from. But we're still not anywhere near the unsupervised learning capabilities demonstrated by a human infant.

charles-salvia··on An Indian Ocean island with a community that has been there for 60K years (2015)
What do you want us to do? These guys keep throwing spears and shooting arrows at anyone that approaches their island. Should we... should we ask the Indian government to send in an armed infantry unit to force the Sentinelese to sit down and watch a video explaining the benefits of 21st century civilization?

Meanwhile, way more people actually living in 21st century civilization in India are currently experiencing extreme poverty and suffering,[1] despite having instant access to Vivaldi.

[1] https://en.wikipedia.org/wiki/Dharavi

charles-salvia··on SHA-1 Collision Detection on GitHub.com
And in the latter case, most human programmers will be unemployed, and thus unconcerned with hash collisions.
charles-salvia··on Suicides in Rural America Increased More Than 40% in 16 Years
You're venting at the wrong leaf node
charles-salvia··on Suicides in Rural America Increased More Than 40% in 16 Years
I really sympathize with this, honestly. At a certain point, as I'm working really hard to debug a segmentation fault late at night, and then catching the subway early in the morning, the idea of having to sympathize with a bunch of welfare-receiving meth addicts* who keep voting themselves into oblivion really strains my patience.

But think of it this way: forget about "rural vs urban", and think of it as a problem with education levels. Rural areas tend to be, on average, much less educated and far more religious. Lack of education has a corrosive effect on democracy, because it garbles the decision making process of entire voting blocs.

So why should Blue State people take the initiative to care? Precisely because we are the more educated (on average) and the more cosmopolitan. When engaging with any population that is sufficiently less educated and more provincial, a more educated person is very likely to become frustrated with all the self-destructive behavior observed. Think of a first-world humanitarian organization trying to build schools or hospitals in an impoverished third world country. They may encounter all kinds of resistance based around irrational fears, superstitions, shamanism, and other self-destructive world views. It's really the same thing that happens when Obama tries to make healthcare more widely affordable and available, only to run up against widespread resistance based around conspiracy theories, tribalism and close-mindedness. (It also doesn't help that some of the highly educated people in power regularly exploit the fears of the lesser educated for their own political gains.)

But if you think about the situation as a problem of education, rather than "urban vs. rural", it becomes easier to avoid losing your sympathy when you encounter this sort of self-destructive behavior among the less-educated.

* it's a joke, relax.

charles-salvia··on Suicides in Rural America Increased More Than 40% in 16 Years
What the hell are you talking about?
charles-salvia··on Suicides in Rural America Increased More Than 40% in 16 Years
I can understand your sentiment - but unfortunately we can't just ignore rural America.

The problem is basically that, on average, rural areas tend to be less educated and far more religious. They vote against social programs that would benefit them for religious or tribalistic reasons. They vote against Planned Parenthood because abortion is an abomination to them, they vote for "repeal and replace" because they vaguely associate anything coming out of the Obama administration with a sinister Muslim/Communist/Homosexual agenda to expand governmental powers and take their guns or whatever. (These are generalizations and stereotypes, obviously - but as an explanation for average voting patterns they are valid generalizations.)

charles-salvia··on How the baby boomers destroyed everything
I'm not a baby boomer, so I have no interest in defending them, but there's something inherently tautological about this whole line of reasoning. The argument is basically "everything sucks right now, so blame the largest voting bloc/age-group". I mean, yeah, that line of reasoning is more or less valid in any moderately-functioning democracy. But it doesn't really actually explain anything useful. I mean, it's not like people born in the period after WW2 to the mid 60s are somehow genetically predisposed to be complete sociopaths.

I realize the article attempts to touch a bit on why the baby boomers may have been psychologically predisposed to be so irresponsible (raised on TV or something), but that seems pretty hand-wavy. The reality is it's simply human nature to take the path of least resistance and put off solving hard problems in favor of easy short-term gains. And after WW2, America experienced unprecedented economic growth as a new emergent industrial superpower. But that kind of growth can't last forever. The baby-boomers were just around at the right time to benefit from it, and their natural human short-sightedness pretty much guaranteed they wouldn't prioritize making decisions that benefited future generations.

← PreviousPage 2 of 5Next →