9,437 karma · joined November 23, 2008
Get in touch! jballanc@gmail.com https://github.com/jballanc
[ my public key: https://keybase.io/jballanc; my proof: https://keybase.io/jballanc/sigs/pxb6FPsVF9hoQMyb2ap0euWoNG3YxwN2g4VY94RRxX8 ]
Almost no one will attempt to parse SMILES for non-trivial molecules by hand, but almost every computational package can easily parse them.
In Julia, objects are data-only and methods are defined at a module level. So, whereas Scala's structural typing need only introspect the object being passed as an argument, Julia's behavioral typing requires introspection of the entire dispatch tree. The benefit to behavior typing and Julia's multi-dispatch is that if you are missing one or two methods for some type in order to be able to use it in some function, you can always define the missing methods locally.
If you're familiar with duck typing, behavioral typing is (in essence) the reification of duck typing in a concrete type system. In other words, instead of specifying the type of your argument as "Array", you could specify "some type that is indexable, iterable, and can be appended to".
In Julia (mind you this was just the idea I saw being considered), today you would do:
function foo(myarray::AbstractArray)
...
end
but in the future you might be able to do something like: function foo(myarray::ANY{getindex(), setindex(), iterate(), append()})
...
end
what's really neat, though, is combining this with type aliases, you could have: typealias Arraylike ANY{getindex(), setindex(), iterate(), append()}
function foo(myarray::Arraylike)
...
endOf course, in this regard Julia is not any different than Python, Ruby, JS, PHP, etc. Also, it's worth noting that Julia has cluster-computing as a concept baked into the language. So, whereas you might need to worry about distributing Go executables to multiple machines in a cluster, with Julia you need only have the Julia runtime installed, and from there any Julia program can distribute itself to any available nodes without further action required.
This is also why I'm excited by what's coming next in v2.0. Some of the early ideas being explored at JuliaCon with respect to traits and interfaces will begin to really shine a spotlight on its true power.
As for Julia 2.0, it was decided to put off implementation of traits and interfaces until then. That said, some of the ideas that were being bounced around during the post-con Hack day in these areas are really exciting. In short: Julia may be the first language to pull off behavioral typing in a practically usable way.
(Also, I think the comparison to CoffeeScript is rather unfortunate. In my mind, CoffeeScript was primarily addressing shortcomings in JS that JS has, since, done quite a bit to rectify. Julia is not trying to be a better "X", so there's no "X" to steal away its momentum.)
Awesome work! (And if you haven't already, check out some of Steven's other projects. Very fun/interesting code.)
* Declining dominance of the Microsoft/Oracle hegemony in secondary markets
During the '90s Microsoft and Oracle were particularly effective at entering emerging markets and pushing their solutions to such a degree that "computer programmer" became synonymous with "Microsoft/Oracle developer". Thus, if you wanted to become a software developer in one of these markets, it meant learning C#/Java, along with all the ecosystem, tooling, and inevitably...lock-in.
Fast-forward 20 or so years, and the Open Source movement is finally beginning to crack that shell of lock-in. Anecdotally, when I first moved to Turkey 4 years ago, technology job sites almost exclusively listed C#/Java positions. Today, I know many Python, Ruby, Scala, Go, etc. developers living and working in Turkey, and there's been an explosion of related conferences, trainings, events, etc.
The one last hurdle I see that needs to be overcome in these places is the notion that developer education is as much a concern for those that employ developers as for the developers themselves. Pretty much all of the conferences/trainings have to be held over weekends because it is still difficult for developers to get time off during the week to attend. Once that bit of cultural shift happens, I only foresee this trend accelerating.
Edit: Though it's worth noting that both those transitions were achieved in less time than since the original NYC pneumatic tubes and now.
The thing about hyperloop is that none of the technology is new...or even recent! The pneumatic mail system in NYC was in use between 1897 and 1953. What's new is that in the 1950s the energy economics made it cheaper to replace pneumatic tubes with automobiles. Fast forward 60 years and the energy economy has been turned on its head, thus hyperloop.
Then there's SL-1 (https://en.wikipedia.org/wiki/SL-1). That was a really stupidly simple failure, but of all the accidents I've read about I find it the most poignant reminder of why nuclear reactors are such complex systems in the first place.
(Note: Just so you know what to expect, Turkish dramas are sort of a cross between their US counterparts and Spanish Telenovelas. There is some action, but much more emphasis on interpersonal conflicts...and more than a few very long, overly dramatic shots backed with suspenseful music.)
The insistence of US media companies to continue geographically restricting content in an effort to make a few extra bucks has done more to kill US cultural imperialism than decades of "home grown" efforts. As a concrete example, in the past few years there has been an explosion of GoT-style dramas in Turkey (which they've even begun exporting to neighbors in the region).
Perhaps not surprisingly, this season GoT was available same-day on Turkish cable, with either dubbed or original voice tracks.
There are efforts underway that may eventually change this (Open Data, Open Publishing, pre-print servers, etc.), but it's a big ship to turn.
Consequently, this is why you'll hear many physicists claim that string theory is not science. It certainly looks like science, and some day it might become science, but currently as the bulk of string theory is not falsifiable, it is not, strictly speaking, science.
The short answer is: peer pressure.
The longer version requires a look back at the history of scientific publishing and how it's been perverted in more recent times.
Originally, scientists (or natural philosophers as they were known then) would carry out experiments and record their findings in their personal papers. The problem with this is the question of priority: who discovered what first. If you leave this question up to individuals recording their own results, then there's nothing to stop those individuals from backdating their own writings in a false attempt to claim priority.
So a tradition grew of communicating results via letters. When you discover something, don't just write it down, send it in a letter to a fellow natural philosopher. That gives you a (presumably) independent means of verifying priority. Of course, this does nothing to prevent two friends from colluding to falsely claim the mantel of priority.
Eventually the tradition morphed. Instead of just writing a letter to a colleague, you would write the letter and the colleague would read it to a gathering of natural philosophers (or, if you had the means of travel, you could read your findings, or in some cases demonstrate them, to the gathering in person). This gives you a much more robust means of claiming priority. Eventually, these societies organized themselves, one of the earliest being the Royal Society.
It didn't take long before these groups decided that all the letters being read should be recorded in some format. Thus, you have the "Philosophical Transactions of the Royal Society", the first scientific journal...
...in 1665. Since then the enterprise of scientific publishing has changed very little. Why is this bit of history important? Because the point of the original journals was not to be a full record of the experiments being performed, but rather just of discoveries so that their priority could be established. The scientists of the time still communicated their efforts, both successful and not, with each other in myriad ways. Perhaps more importantly, positions at universities were often based on these personal communications (and recommendations from peers).
Fast forward 400 or so years and scientific output (and indeed the number of scientists and universities producing this output) undergoes exponential growth. Suddenly, hiring professors based purely on personal communications becomes impractical. Ah, but what's this? There's also an exponential growth in scientific journals, and some candidates for these positions have their name on more by-lines than other candidates. Surely they must excel at science, no?
In other words, scientific journals, which were originally just one part of a scientists work, were perverted to become a stand-in for all of a scientists work. Yet, owing to their history as a means of establishing priority for discovery, the papers published continued to be propelled by tradition to emphasize positive results.
If a practice that has remained unchanged for 4 and a half centuries isn't ripe for disruption, I don't know what is!
* make an observation
* form a hypothesis that explains the observation
* design a test to disprove the hypothesis
* carry out the test and collect data
* analyze the results and either reject the hypothesis or formulate a new test
It is the necessity of a hypothesis to be falsifiable, and the intent of the scientist to disprove (not prove) their hypothesis that sets science apart.
> "lamentable archiving and data-sharing practices" that prevent most of the discipline's body of work being re-analysed.
Keeping data private before publication is (at this point in time) understandable. Once results are published, however, there is no excuse for not depositing the raw data in an open repository for later re-evaluation.
In short: as automation is better able to handle easy-to-predict scenarios, that increasingly leaves only rare, complex scenarios for the humans to deal with which, by definition, they are far worse at handling.
The "cautionary tale" is that releasing GMO crops before their full impact has been studied and assessed only adds fuel to the GMO/anti-GMO fire. When it was found that bt-Corn could pose a risk to Monarchs, those in the anti-GMO camp had all the ammunition they needed to start shouting about GMOs being unsafe. When it was later found that the risk was negligible in practice, the GMO camp had all the reason they needed to continue turning a deaf ear to the anti-GMO camp.