etal

With a few words switched, I find this a pretty accurate description of the PhD and postdoc years:

http://en.wikipedia.org/wiki/Journeyman_years#Historic_roots

We might be agreeing here already, but my understanding was that the mindset Java fought against -- and finally won -- was: Interpreted, garbage-collected languages, or basically anything with a substantial runtime component, will be too slow and unpredictable to use in serious production environments.

So Java went all-out with (1) a static type system and basic language features that C++ coders could respect, (2) a level of hype and corporate sponsorship/support that PHBs could respect, (3) a portable, eventually fast VM, opening up some nice deployment options. There was no need to introduce really exciting language features, only to keep promoting the OOP concepts that C++ thought-leaders were so enthusiastic about at the time.

Not disagreeing with you. In fact, the first example is one reason why "print" was changed from a keyword to a function in Python 3:

  myfunc <- lambda x: (print x) or x*x  # SyntaxError

  myfunc <- lambda x: print(x) or x*x  # OK in Py3

The "magic methods" wrapped in in double-underscores are special because other parts of the language's syntax rely on them always meaning the same thing. (Some room for programmer misbehavior here.)

Example: If __len__ is defined on an object x, and __nonzero__ (Py2) or __bool__ (Py3) is not defined, then "not x" will test if x.__len__() == 0.

Similarly, __contains__ interacts with "foo in x", and __eq__ and __hash__ are expected to play well together for hashable objects (e.g. unique members of a set).

The double underscores are a useful code smell that indicates you're changing how an object interacts with Python's syntax.

Thanks!

http://en.wikipedia.org/wiki/Baumol%27s_cost_disease

Baumol's cost disease (also known as the Baumol Effect) [...] involves a rise of salaries in jobs that have experienced no increase of labor productivity in response to rising salaries in other jobs which did experience such labor productivity growth. [...]

Baumol's cost disease is often used to describe the lack of growth in productivity in public services such as public hospitals and state colleges. Since many public administration activities are heavily labor-intensive there is little growth in productivity over time because productivity gains come essentially from a better capital technology. As a result growth in the GDP will generate little more resources to be spent in public sector. Thus public sector production is more dependent on taxation level than growth in the GDP.

Those sectors are also much harder to automate, relative to the rest of the economy, since people want individual attention from doctors and teachers. So as the cost of everything else drops thanks to improved automation and technology, the proportion of our wealth that we spend on necessary but less-automated services increases.

(There's a term for this effect, but I'm having trouble finding it again.)

Did you just subtract the values in the original chart from 100%? If so, that's not the actual employment level. Unemployment means seeking work but not finding it; it excludes everyone who is disabled, discouraged, or simply not looking for work (e.g. full-time moms).

The overall unemployment level in the U.S. is a little under 10%, but the employment level is somewhere around 65-70% (iirc). So 20-25% of the population is still unaccounted for in this figure. A chart breaking that down by education might be interesting, too.

First, there are an awful lot of schools that offer semi-worthless MBA degrees as a profitable business. Much more so than in engineering or CS.

Then, it's because business expertise and talent is hard to verify. Programming or engineering ability is easy to prove, even at entry level. You can't prove you can manage a large group until you're already in charge of a large group. This means a company is taking a big risk on hiring MBAs versus CS grads, and needs to look for some clear, objective signals before going ahead.

Finally, executives like hiring managers who went to the same school they did. I don't know exactly why this is the case -- reciprocity, maybe, or some notion of a shared vision of how to run a business.

For kicks, here it is in R:

  PlusRandom <- function(initial, steps) {
    if (initial > 1) steps
    else PlusRandom(initial + urand(1), steps + 1)
  }

  mean(replicate(100000, PlusRandom(0,0))

  [1] 2.71872

Quick poll: When you were in school, especially college, were the best teachers young or old?

There are young teachers with talent, but in my opinion, the most outstanding lecturers were older and had a depth and variety of experience that the young ones couldn't match. Older teachers have the fabled 10,000 hours of deliberate practice to account for, as well as a diversity of experience and the time necessary to assemble some perspective on all of it.

So here's a big +1 for hiring adjunct professors from outside the usual pool of fresh young faces. Let the young people work, and the old people teach.

Stand: http://www.osha.gov/SLTC/etools/computerworkstations/positio...

It takes a little while to get used to it, but once you're there, standing all day is just fine.

The point is that it's not an iOS feature, it's an iPhone feature. The original question is about the operating system, not the hardware. This distinction doesn't matter much for the iPhone, but it certainly does for Android devices.

Does iOS provide some special support for high-resolution screens that Android doesn't? If not, then that's a hardware feature that any rival could easily pick up. Remember how Motorola Droid first upped the ante on screen resolution at the end of last year.

Java needs an environment that can generate, parse and rearrange Java code in a way that vimscript just isn't suited for. Even in Emacs it's a Herculean effort to make the editor understand Java at a deep enough level. Generally, if the language you're using co-evolved with a particular editor, use that editor, not vim. (That's not limited to Java.)

On the other hand, C as a language grew up alongside vi, as well as the Unix environment it leverages so well. Languages with a nice REPL, like Python and Ruby, also are concise and flexible enough that any text editor will do; the languages are meant to work well enough with minimal editing capabilities. In those cases, vim is still great, because it's optimized for rearranging text of any kind.

Yep, it's a bit of a rant, but an interesting one. Wozniak's an interesting person; I hope you've checked out Supermemo and its open-source counterpart, Mnemosyne.

I've been taking a 20-45 minute nap in the afternoon for close to a year now. I think it was Dustin Curtis's article on polyphasic sleep that prompted me, and I went with the "siesta" schedule because it seemed the most sustainable -- a lot of people do live that way, and getting work done in the mid-afternoon is hopeless for me anyway.

I was going to post Puredoxyk's rebuttal here too, but noprocrast kicked in. Thanks.

The main weakness in Wozniak's argument/rant is that he assumes all people work the same way. For most people, monophasic or biphasic sleep is most comfortable, and normal human biology seems to back that up.

But Puredoxyk says she has always had spectacular sleep disorders, and that prompted her to try Uberman originally. I doubt her brain chemistry follows the usual biphasic pattern, so it makes sense that frequent, regularly timed naps would work better for her than a forced monophasic schedule.

Steve Pavlina seems to have incredible self-control.

That's evolution. Start with something that works, add something else that's available, try it out, then see what you can drop.

Those examples do fit: Tyrosine kinase inhibitors turn an always-on enzyme into an always-off one, when the normal behavior is for it to act as an on/off switch according to the cell's immediate environment.

http://en.wikipedia.org/wiki/Erlotinib#Side_effects

http://en.wikipedia.org/wiki/Imatinib#Tolerability_and_adver...

Once the gene mutates in a cancer cell, it's never going to work properly again (unless all the cancer cells are killed, or the gene is somehow replaced). It's just less harmful if it's disabled than if it's permanently active.

Antibiotics tend to kill good bacteria in the body, too. Most people are willing to take a week or two of GI problems and some extra servings of yogurt in exchange for survival, but still, that's some collateral damage.

The common problem is, once one part of a system in the body begins malfunctioning, it's very difficult to fix that part without affecting the surrounding system, or other parts that work similarly but aren't malfunctioning.

Does numpy's record array do what you want?

http://www.scipy.org/RecordArrays

It doesn't quite have first-class status in the library the way data.frame does in R, but it does let you index an array using strings.

It's handy for the same reason FreeDOS is handy -- eventually Microsoft will EOL the platform, but there will still be programs that you need to run (for whatever reason) that require that platform.

You're right that there's no point in chasing the most recent Windows version.

Original post:

http://blog.diegovalle.net/2010/06/statistical-analysis-and-...

The author compares police records, death certificates, and the historical record of well-known massacres to identify inconsistencies regionally and over time. (Looks pretty rough in Juarez.)

Yes, charities and NGOs like the American Cancer Society fund grants, too. But basic research is usually way outside their core competency, so teaming up with universities is their best option.

Companies team up with academic labs occasionally, and I think this is great.

But the government does already run its own labs and fund semi-independent research-only institutions. I'd like to see an expansion of these (mainly the latter), as well as Master's programs in science/tech, to help this current glut of graduate students transition to non-academic careers.

Good point. I'm thinking of "dead labs", where a tenured professor runs out of grant money and can't secure any more, so students and lab staff gradually disappear on their own, since the only official long-term job is the PI's. One unproductive professor's salary isn't much compared to an entire office or building in the D.C. suburbs.

(Empirically, a professor doing X will continue doing X until retirement, regardless of the grant situation. But students can still opt out.)

Graduate assistantship stipends are probably an OK measure. At my school, funded psychology students get around $9000 or $10000 during the academic year, and have to work as a TA each semester (grading or teaching lab sessions). Genetics students earn $20000 and simply work in their PI's lab. This is for students funded through the graduate school -- PIs can also hire grad students and postdocs with money from their own grants, in addition to those already-funded students who join their labs. In either case, the graduate school waives tuition for anyone on an assistantship.

Plus, science students have relatively good options in industry after graduation, and as far as I can tell, more postdoc positions available.

There are also programs like this for science students:

http://www.sciencemasters.com/

Oh my god, I hate journalists.

1. This is not an article about career opportunities for STEM (science, tech, engineering, math) majors -- of which there are plenty. This is about the very specific problem of there being a massive glut of Ph.D. students aiming for tenure-track academic jobs. The article makes no distinction here.

2. The problem of academia's imploding pyramid scheme is not at all limited to science. It's worse in the soft sciences and humanities -- picture the same situation of tenured professors overseeing teams of cringing graduate students, but without the NSF/NIH/DOD grant money to hire postdocs. Now, where do you go when you graduate? (Hint: you don't, unless you already have a job lined up.)

The article mentions the rules of supply and demand, but not in relation to tenure. Academia's cultural problem of deterring students from any career other than academic research gets a couple paragraphs near the bottom. But that's the problem! Regardless of the field, too many students are pushed into academic careers, but the world just doesn't need that many professors.

Lots of people love their job. It takes the arrogance of a professor to state that everyone should want yours.

3. Visa issues -- smart, talented students come from outside the U.S. to attend grad school on a student visa that prohibits non-academic work. They stay on as postdocs after graduation because every other kind of job is prohibited, unless some benevolent employer wants to lend a hand with the work visa application process, which probably won't happen. If they leave academia, they have to leave the U.S. Market distortion, anyone? Anyone with U.S. citizenship is clearly better off with a Master's or other professional degree; industry misses out on talent while academia suffers a labor glut.

4. Why does all scientific research have to be done by universities? If labs are funded with government money anyway, it's just as efficient (or more) to directly hire lab techs and researchers at government labs... assuming the research really is important. Academic labs are optimized for publishing journal articles, in absence of a better metric. Could we measure progress better by bringing the research work closer to the customer?

If there's a killer app on their platform, then Apple can't change their platform in a way that breaks the app. Consider the hoops Microsoft has had to go through with Windows and IE to avoid breaking important third-party products. Since the whole point of Apple products is high-end, uncompromising design, letting killer apps get the upper hand would quickly make iOS stagnate and lose relevance. Once that's happened, the only way to get ahead again is something like the MacOS-to-OS X upheaval, which wasn't particularly pretty.

Lee Kuan Yew, modern Singapore's daddy, has spoken and written a lot of interesting things about his philosophy in making the country the way it is (politically and culturally) and how it should relate to the rest of the world. Here's a great interview:

http://www.lee-kuan-yew.com/lee-kuan-yew-charlie-rose.html

Not nearly as many Singaporeans drive, though, and a pretty substantial proportion of the cars on the street are taxis. So, given that taxi drivers are generally good drivers but prone to the occasional stunt to get a passenger somewhere on time, this could just be pragmatic, where the government isn't interested in punishing the taxi drivers that keep the city running smoothly.

Also, Singaporeans are better trained than Californians, so less of a deterrent is needed.

And, U.S. traffic laws and fines seem decidedly optimized for raising money for the local city government, rather than the safety or convenience of drivers or pedestrians, whereas Singapore has a more monolithic and paternalistic government which can raise the money it needs through other means.

There's one more nice side effect of having ~1 open space on every block -- drivers don't need to circle around the neighborhood multiple times looking for spaces, so the streets clear up a bit.

I was going to say that Python fixes in programming languages whatever Markdown fixes in other plain-text markup syntaxes. But thinking about it, you're right -- it fixes the tendency of other languages to invent special syntax to represent standard, frequently used concepts. Python follows the conventions you might use when sketching out an idea on a napkin, and takes each convention to its logical, parser-friendly conclusion.

There are few language features that would be hard to succinctly express in a line of pseudocode. ("If the implementation is hard to explain, it's a bad idea.")

Result: straightforward, algorithmic code looks like pseudocode, and any code that uses special language features really pops out -- if you see an at-sign or double-underscores, you know something Python-specific is happening.