cossatot

No, many (if not most) academic fields have high-quality journals run by non-profit societies. In my field (geoscience) the best of these are trusted more than Nature and Science, which are higher impact even though it's common knowledge that there are more unreliable papers in the glamour journals, mostly because the editors optimize for impact instead of scientific solidity. The non-profit societies charge similar rates (a few hundred to a few thousand per article). The fees reflect outsourcing of web hosting, formatting (both LaTeX and Word are re-keyed to XML for many journals) and typesetting, as well as dead-tree publication costs which are actually common if rarely read.

Additionally, to get at your earlier question of evaluating a paper by its host journal: Because academics are strongly incentivized to publish in high-impact journals, if someone publishes in a crap journal it raises eyebrows, because the obvious question is 'why' and the obvious answer is 'because the authors didn't want rigorous review' or 'the work is of supreme unimportance'.

Furthermore, because of curation aside from peer review (editors weeding out solid but unimportant papers) the journal's brand makes a difference in how many people will actually see the paper. While many readers will come across papers either by specific searches in Google Scholar, or by going through the citations in a different paper, many of us still read the tables of contents of a few key journals every week or month to see what is being produced.

I'm a non-academic scientist (in the non-profit world currently) and though I will never face a promotion and tenure committee who really gets into impact factors etc., I still want my work to be as widely read as possible, so I push for the best journals that I can get in that still meet my criteria (non-profit societies that have open-access at least after a year embargo). Google Scholar also ranks papers by how highly cited they are, so there are still viral-type effects and getting eyeballs on a paper early greatly affects later discovery and engagement.

edit: I took 'paid-access' to mean 'pay to publish' which was the original thread's wording, and wrote based on that. Some of this reply still holds for 'pay to read' but not all.

Have you been specifically told by NSF that you can't? I have included OA funds as line items in NSF (Earth Science division) budgets before, and never received any flak. However, none of the grants ever got funded, so maybe they didn't get the fine-toothed comb treatment following peer review.

Sort of off topic but maybe not: I'm in Chengdu right now, my first trip to China since 2010. I had previously just been to Beijing among Chinese cities (lots of time camping in Xizang backcountry otherwise). Chengdu is so tranquil, orderly and clean compared to the Beijing of my memory. People don't yell or spit, traffic lights are obeyed and drivers slow for pedestrians, the city is quite nice. Is this just the difference between the megacities and the smaller cities, or has Chinese urban culture matured a bit?

Many of us do occupy the middle ground of thinking 'maybe' or 'I really don't care' or whatever. We also tend to avoid commenting because bitcoin discussions are among the most useless on HN.

No disagreement from me on the sorry state of things.

Is this your comic, perhaps?

I set up a Fedora/Cinnamon desktop at my last job and loved it.

Thanks. I clicked on the youtube link before reading the article and was surprised at first that the paper was about Warren G (as some geek off the street, I had never heard of Michael McDonald).

My grad school advice, as a PhD who is not good with rules:

- Understand what the rules are, why they are in place, and to whom they matter. Some are there to keep you from hurting yourself, some to keep you from hurting those on your level or below, some to keep your advisor from abusing you, and a lot to make sure you act in the best interest of the University via minimizing its liability, etc. You don't need to follow all of them all the time, but have some real justification when you do go off course, don't call attention to it, and don't screw anyone else over.

- For your research, you need a dual-prong strategy of making sure you check off the boxes for others, i.e. actually doing what your advisor got funded to do (she is on the hook for this for most grants), on time and under budget as well; then, pour all the rest of your efforts into researching what you are interested in, with an eye towards building your skill sets for the future and learning how to learn alone. Read widely enough to know the range of options for a solution space and deeply enough to know what has actually been tried in your field, and not just what has succeeded. Teach yourself stuff that your advisors don't know. The payoff from a broad familiarity with programming, statistics and discrete mathematics is large, especially if they're under-utilized in your field, and these skills are widely transferable. Furthermore experimenting/exploration through computing is often super cheap and the cost of failure is low, unlike destructive analysis of your physical samples.

- Keep your datasets small enough that you can master them. In the natural sciences, many scientists just want more and more data but getting and maintaining data is labor-intensive and has a huge opportunity cost. You don't always learn a lot of transferable skills that way unless you want to be a lab tech forever. The major exception is when obtaining data is primarily observation and measurement, which are really important to building intuition and just noticing subtleties.

Is this a vision from a world without open source?

They care to the extent that the alumni are willing and able to make fat donations.

Always a good time to re-read Cosma Shalizi's classic "So You Think You Have a Power Law — Well Isn't That Special?" (http://bactra.org/weblog/491.html)

It's not just Americans. When I was working in Italy, some of my Italian colleagues would put an entire sugar packet into a single shot of espresso.

I must be alone in thinking that times are pretty good and the US is not fracturing, and today is nothing like the late 1960s or 70s.

Crime/violence:

Both violent and property crime rates are way down [1].

My home region of the Ozarks (SW Missouri, NW Arkansas, NE Oklahoma) is much safer than when I was in jr. high and high school (1995-2001) when there was a huge amount of rural violence associated mostly with the meth trade. I had friends and neighbors killed both purposefully and incidentally with this, none of whom were users or involved in the trade in any way--just children or landlords or neighbors or whatever. We used to get out of school because the cops would be raiding meth labs within shooting distance of the school. This has largely subsided. There is definitely violence associated with the opioid crisis, but the deaths have been overdoses and suicides at least among the people I know.

War:

There are definitely some nasty and protracted conflicts abroad, but, unlike in 1968, the US is not mired in them and losing thousands of soldiers a year--Wikipedia gives 4,491 deaths for US soldiers for the entire Iraq war, for example. Not that it wasn't a horrible and unnecessary loss, and several orders of magnitude worse for Iraqis than the US, but this was also the case for Vietnam (and Cambodia, Laos, etc). While there continues to be strife in the Middle East and a few other places, and I worry about Africa being a major site of east vs. west proxy war in the coming decades, globally the situation seems more peaceful and democratic than in the Cold War or the colonial era.

Societal issues in the US:

There is a lot of sound and fury, but protests are largely peaceful. The protests involved with Ferguson, Charlottesville, etc. were much more mild than in the 1960s, or the race riots of Tulsa (1921, 39-300 deaths, 10,000+ people left homeless [2]), LA (1992, 63 deaths [3]), and so forth.

Gay rights were granted largely peacefully as well--lots of progress here, though it wasn't frictionless.

Similarly, the recent wave of defenestrations of powerful men due to sexual abuse is to me a net positive even if there are inevitable witch hunts associated.

Domestic terrorism seems to be less of a concern. I don't know about rates but from what I understand there were a lot of bombings from both far left and far right sources in the 1960s and 1970s. We haven't had any high-profile political assassinations in a while.

The rise in single-parent households is a concern. I have no idea how much this reduces the incidence of domestic violence as moms don't live with abusive dads. Maybe a little?

Homelessness is also a major concern but is slightly decreasing over the past decade, not even accounting for population growth [4]. I have no idea how it compares to the 1960s or so. I think a lot of mentally ill people were institutionalized then, which isn't the case now AFAIK.

The economy:

Parts of the economy are good (stocks, general employment levels), and though there are still problems with stagnant wages for many, debt levels (particularly student loans), and housing prices in many places are terrifying (I'm living in SV on a pretty meager nonprofit salary, because my wife has a great job that pays OK). But it's a lot better than it was a decade ago, or in the 70s. There are still huge issues with medical care that need to be resolved, but at least lots of people are employed as paper pushers...

Inequality is a major concern, I concede. And I worry about an upcoming stock market crash.

I think it's possible that trade work will pay better and better, and as more tradespeople rejoin the middle class communities (neighborhoods, school districts etc.) some of the inequalities of social status between the higher-income blue-collar work and lower-income white-collar work will dissipate.

The environment:

Way better than in the mid-20th century. We have a lot more protected lands than in the 1960s, and air and water quality are much improved (thanks, Nixon!). From Wikipedia:

"In the United States between 1970 and 2006, citizens enjoyed the following reductions in annual pollution emissions:

- carbon monoxide emissions fell from 197 million tons to 89 million tons

- nitrogen oxide emissions fell from 27 million tons to 19 million tons

- sulfur dioxide emissions fell from 31 million tons to 15 million tons

- particulate emissions fell by 80%

- lead emissions fell by more than 98%"

Without looking stuff up, I think that the broad population shifts from rural and semi-rural to urban and suburban since the 1950s are decreasing the pressure on the environment; national forests in most places are recovering from the ~1850s-1950s logging period.

Climate change is bad, forest fires are bad, the decrease in insect populations are probably quite bad and may indicate a very rotten ecological foundation. However, the effects of climate change aren't directly tearing our country apart at present; the debate over it (and the underlying struggle for power and authority between science and government vs. church and business) is or did contribute. Once agriculture begins to fail in the Great Plains and CA central valley, it'll get real.

---

Honestly, I don't get it. Sometimes I think things have been decent for long enough that we've forgotten what it was like when things were really bad--the Civil War, WWI, the Great Depression, WWII, Vietnam, etc.

I also think that a lot more of the seeming chaos is that more voices are included than previously, particularly from marginalized communities.

This aren't perfect, but are they awful? Or are we just bored and riled up about small matters such as the rhetoric of our politicians, or their seeming inability to get stuff done?

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

[2]: https://en.wikipedia.org/wiki/Tulsa_race_riot

[3]: https://en.wikipedia.org/wiki/1992_Los_Angeles_riots

[4]: https://journalistsresource.org/studies/government/health-ca...

From TFA:

Unlike money-losing Snap, Dropbox will come to the table with annualized sales of more than $1 billion, Chief Executive Officer Drew Houston said in an interview last year. It’s also been profitable, excluding interest, taxes, depreciation and amortization. Those benchmarks are the product of more than two years of focusing the company, expanding its product suite for businesses and reining in expenses, Houston said at the time.

They are related phenomena. Basically, sand and mud have pore space in between the grains that is filled with fluids (water, oil or gas). When those fluids are extracted, the rock contracts because the pore pressure decreases, causing an increase in density (compaction) and a decrease in volume, which leads to subsidence. Some of this is recoverable, and some of it isn't, because the grains will find a more efficient packing arrangement.

You are completely correct in that there are many different factors involved that are all causing problems; unfortunately they are all interrelated.

Geologist here: while overly managing of the Mississippi does make a difference, that only controls where new sediment gets deposited. Even if new sediment was deposited in sinking urban areas, it wouldn’t ameliorate the problem of subsidence because it would be deposited on existing infrastructure, burying it.

The real reason for the subsidence is that the sediment that makes up river deltas slowly compacts under its own weight, expelling pore water and becoming more dense. This is natural and unavoidable, although it might be marginally enhanced by oil and gas extraction (it is a bigger problem with groundwater withdrawal in other areas).

I think an additional limit is data. Particularly for sciences with strong path dependence, such as ecology, cosmology, evolutionary biology and (my field) geosciences, time has simply erased nearly all of the record of previous events. This is both a big problem (it is quite frustrating) and an opportunity, as we progress through both discovering new records of the past, and developing new techniques for analyzing existing records through new or improved proxies, instrumentation, or theory. I think it’s obvious to the practitioners that we will never know the vast majority of what we want to know, but we can learn enough of what we need to know to keep moving foreword.

They have your phone number if anyone with an android phone has you in their contacts.

Ok thanks to @temporal here is the diagram. It's a cross-section of a subduction zone, i.e. it's in a vertical plane (not map view).

               0
              0       depth (km)
  o->____====Δ== <-c  -0
         \            
  (fast)  * <-e       -50 
           \          
        d-> \         -100
  (slow)     \          
                      

The 'o' is the oceanic plate (thin line), the 'c' is the continental plate (double lines), 'd' points to the downgoing slab, 'e' points to where the earthquake occurred. O and C are the surface of the earth. 'fast' and 'slow' indicate that the slab is sinking/subducting fast at shallow depths and slower at deep depths. The Δ is an arc volcano that is smoking (like Mt. St. Helens or Fuji) that are always present above subduction zones because at ~100 km depths, the subducting slab loses some water which rises and then melts the upper mantle and lower crust of the continental crust above (these rocks are very hot and adding water lowers their melting temperature without changing the actual temperature much).

Continents are basically rafts of granite. Like any raft, if you put more stuff on it, it sinks a bit.

Quick explainer about intraslab/intraplate subduction zone events (a slab is a subducting plate):

There are two major classifications of crust on earth: oceanic and continental. Oceanic crust (mostly basalt/gabro) is quite a bit more dense and thin than continental crust (mostly granite). Both float on the more dense mantle like big rafts, but the much thinner and denser oceanic crust sits ~3 km lower on average, creating ocean basins. In fact, the cold oceanic crust (and its attached subcrustal lithosphere) is in some places more dense than warmer mantle at depth; this density contrast is part of the convection cells in the mantle that drive plate tectonics.

When an oceanic tectonic plate converges with a continental plate, the denser oceanic plate slides underneath, at a subduction zone. (edit: I tried some ascii art to illustrate but HN deleted the spaces and ruined it; I fixed it in a comment below.)

There are three types of earthquakes that happen in oceanic plates in subduction zone environments. The first, most familiar and strongest are subduction zone or 'megathrust' earthquakes, like Tohoku (Japan 2011). This is an earthquake caused by the release of stored energy due to frictional locking between the downgoing oceanic plate and the upper continental plate.

The second type is an 'outer rise' event. These are almost always small (M~6 max) and not very damaging. This is caused by the flexing of the oceanic plate as it gets ready to dive down (@stephengillie, this is the stress you're asking about).

These types of earthquakes are easily observed because they're shallow and relatively well understood.

The third type of earthquake is what the Mexican event was; it's an earthquake that happens within the subducting (downgoing) slab/plate at depth (see diagram in comment below). It's not clear exactly why these happen but the pattern of energy release is consistent with vertical(ish) contraction and horizontal(ish) extension, i.e. normal faulting. The best explanation for why a downgoing slab would be contracting vertically and extending horizontally is that it is slowing down as it descends, probably due to a decreasing density contrast between the cold (but warming) slab and the mantle around it. So basically the downgoing slab is buckling like a train would if the front train cars slammed on the brakes but not the back train cars. Actually the best way to visualize this is to think about dropping a flexible rubber sheet (that is more dense than water) into a pool. It will continue to sink in the pool but more slowly than in the air above, and will buckle at the interface due to the buoyancy contrast.

But (unlike at the surface) we can't directly test this hypothesis because we don't really have any data on the velocity field to see if there is a deceleration of the downgoing slab (plate). Many of these deep intraslab events, particularly at 100-700 km depth, are thought to be associated with density changes in the downgoing slab as it undergoes mineralogical transformations to thermodynamically restabilize in vastly different P-T conditions.

moonka is describing a theory called the 'elastic rebound theory' that has been the dominant theory of earthquake mechanics since it was proposed by Harry Reid in 1910 after studying the 1906 San Francisco earthquake. It basically states that the earth's upper crust is an elastic material, and (as they said) stress builds as the different sides of the fault move with respect to one another, but the fault remains 'locked', i.e. it's not slipping, usually from the surface to 10-20 km depth.

This produces both shear stress (i.e. stress that promotes sliding on the fault) and normal stress (i.e. stress that resist sliding on the fault due to friction).

One common manifestation of the elastic rebound theory says that there is some relatively fixed stress ratio at which the shear stress overcomes the frictional resistance to slip and the fault slides rapidly in an earthquake. The basic assumption (which is supported by data for some well-studied faults such as the San Andreas) is that the rates of strain accumulation are generally constant through time, and therefore earthquakes on a fault are quasiperiodic in time and have more or less similar magnitudes. These are called 'characteristic earthquakes'.

In this situation, then the likelihood of an earthquake on a given part of a fault does increase with time since the last earthquake. I wrote a blog post quantifying this for the southern San Andreas Fault [1], using data from the geologic record and only a few assumptions (basically just that the geologic record contains all of the earthquakes).

However, though very few researchers disagree with the fundamental principles of elastic rebound theory, there is an ongoing scientific debate over whether earthquakes are periodic and characteristic, or more or less random. Many statistical seismologists believe that they are the common form of 'random', i.e. that the likelihood of an earthquake doesn't change with time since the last earthquake. Some faults (and especially larger fault systems) seem to behave in this fashion, but for an individual fault this implies very different earthquake physics. There are huge fights in places like California between senior scientists over this topic when doing seismic hazard modeling. Additionally there are other statistical distributions than either characteristic/quasiperiodic and random (Poissonian) that produce different behavior [2,3]--with some models, the longer it's been since the last earthquake, the longer it will be until the next, and with others after some time for stress to renew on the fault, it becomes basically random.

For storms, floods, etc. the likelihood of a big event happening in a given year is more or less equal every year and therefore 'random' with the exception of some multiyear cycles such as El Niño. However, over shorter timescales, they aren't. Big storms are more likely to occur during rainy seasons rather than dry seasons, and big floods are more likely soon after previous big floods (i.e. days to weeks) because the meterological conditions that generate them may be persistent and the groundwater is still high, which means that it takes less rain to get the same river levels.

[1]: http://rocksandwater.net/blog/2016/07/wrightwood-recurrence/ [2]: https://pdfs.semanticscholar.org/3071/de51c5abc4f770b0b11ede... [3]: http://windofweef.jp/library/o_m/disaster_prevention/img/10....

Census tracts are good because they're high resolution and still spatial polygons but are more or less sized by population, having ~2-10k people in each (so population density is factored out).

Matlab's documentation is really good, except when you need to look at the algorithm and can't because it's closed-source...

Julia is really fast but the DataFrames/DataTables package was kind of slow and feature-incomplete. Because Pandas is so optimized, Julia wasn't winning out over Python on the speed side of things, and was definitely losing on the ergonimics, so for data science it wasn't the best choice (unlike say numerical simulation).

I'm sure this will change if it hasn't already, especially with the better parallelization that Julia is capable of.

REI is a coop as well, with a somewhat similar dividend structure. However I doubt that all of the profits are fully distributed at the end of the year to members. I would bet that there are 'members' who are investors/owners, not typical consumers, who actually take some of the profit home (but that is a guess, I am happy to be wrong). And of course any business should have savings as well, for multi-year capital expenditures, incidents and accidents, etc.

Marge: "I brought you a tuna sandwich. They say it's brain food. I guess because there's so much dolphin in it, and you know how smart they are."

It seems that a lot of what is discussed in the paper has to do with longer-term effects, i.e. that nutrients can affect neural configuration, e.g. synaptic plasticity.

I have suspected that there are also shorter term effects (i.e. hourly to weekly) caused by food and alcohol intake that probably work through a variety of channels. First, they do impact mood (whatever that is) which certainly impacts my cognitive ability (both problem solving and focus). There are surely blood sugar effects in many people (not myself throughout the day, but people I know well). I have also always wondered why exactly hangovers so drastically impair cognitive performance and I have noticed myself to be more mentally sharp if I go a couple weeks without drinking.

It'll be nice to see how this plays out longer term with more specific nutrients both macro and micro, to find out how we can adjust our diets accordingly (although I would guess that a veggie- and fish-heavy diet is probably the way to go).

What is an example of an on-demand, app-mediated service that is actually profitable, regardless of whether labor is contract, etc.? Is any one in this space actually making money off customers?

Fortunately the Trump administration (in partial continuation of Obama policies) is attempting to reduce FEMA subsidy for people living in floodplains[1]. It's a messy process to end an unhealthy subsidy that many depend on but with this particular problem it may be worth buying people out--lives are at risk as well as just dollars.

[1]: https://www.bloomberg.com/news/articles/2017-08-23/trump-fem...