Cool man, looking forward to any updates. Do you have a repo or gab or something that we can follow? Gitlab seems less of SJW sellouts than github, and I really like their all-in-one CI, if you're looking for a place to host a repo.

I think you're right about the unsexy bit in general, but the place I'm looking at is quite far on the unsexy end of the spectrum, I'd say. The average employee age is like 45+. I don't want to say too much but basically it is government, or close to it. There's no stock options or sexy silicon valley campus or famous product or website behind it. Most of the younger guys are family men, basically. So I don't know what it is. I can't picture how it happens other than some crazy SJW showing up one day and basically demanding a job, "or else" they would cry racism to the press. If I do end up working there I'll try to find out the origin of it and how powerful they are or if it's strictly a token / theater thing.

No degree? Or just no CS degree? If the former I wish I had advice but I'd be guessing. One of those hackerrank type sites might help prove your ability to get you in the door. The "old" issue is kind of there even if you do have a degree. There are just so many places that would rather have young guys who are easy to manipulate into overwork. I've thought about going into some trade instead, where the technology isn't constantly changing. Or at least getting closer to hardware, where it's not like with web development where the tools are changing every damn day.

No mods sounds nice for having a non-echo-chamber space. Would you run into problems with people posting illegal content if it were unmoderated?

I actually think that a decent number of women programmers at certain companies these days might be doing their more substantive work from home, with "help" from their programmer boyfriends or husbands. Not that they are necessarily unqualified themselves, but it's just a logical strategy for a couple working in tech trying to maximize their combined income, since some companies will grant young women big salaries over-easily just to avoid discrimination stuff, and since, in many cases the dude is going to be the stronger programmer, and probably willing to work overtime (and it provides an arrangement where he she is willing to accept him playing video games all evening, because well at least he's also giving her round-the-clock office-hours!)

The hardware interaction jobs though-- with actual devices like airplane components, medical devices, etc... yeah you can do some work from home but it's not the same, and it's not generic IT skills like a lot of these web services companies are.

I'll watch for your AMA's. I've been fairly early in following the various censorship driven exoduses in social media and am always willing to try something new.

I'll have to see if I can find my own way of bringing it up.

So in your experience then there have been some places that don't respond negatively? I live on the west coast and so far everyone I've talked to has been really guarded and seemingly scared of wrongthink.

Oh cool. Yeah I wouldn't have thought of picking up a machine at home. Not sure what I could make... maybe some parts to refurbish a piano? Or maybe some custom eyeglass (plastic) frames?

Yeah I agree with that sentiment to some extent. There are academics though that are geniuses (and it's especially clear when this is the case in technical fields) and drive the field forward immensely with their productivity. So it's not really fair to generalize always. But they are the 0.01%.

Research is a little different than teaching. I've encountered a lot of untalented and/or weak-willed people in research unfortunately, because it's a system that can be gamed pretty easily. But at the same time there are a minority who do really exceptional work. They too can drive the field forward for example by creating free software and things like that.

It's a real dilemma because on the one hand, governments funding research can produce a lot of value in the economy, but on the other hand, governments throwing at money at things tends to create giant inefficient parasitic bureaucracies that won't go away, and that harbor diversity candidates and the like.

Ask to be seen by the whole group (you would work with) for 15 or 20 minutes to find out about everyone

Are you talking about asking to have the whole group in the room at the same time to get a sense of the dynamics? That sounds like that would indeed be a wealth of information seeing the group dynamics. I'm just so used to going through the normal interview "loop" and not asking for anything special like that.

Ha, nice. Yeah I actually thought about CNC programming at one time (that's at 93% on your list). Knew a guy who worked CNC.

That's a fair suggestion. I'm really an academic at heart though so it would be a tough shift for me. I've worked mostly in research. I still believe in research to some extent but I think that research organizations are pretty vulnerable, because there's not a whole lot of testosterone in them, to put it bluntly.

Yeah, it's ironic that we insist on using multiline comment syntax, but then usually subvert the convenience of it by adding "* " on each line to make it look nice.

Perhaps in domains where verifying that an answer is correct is easy, but finding the answers is hard. Perhaps at the output nodes you verify that a result is correct, and if not, feed the data back in to the input layer.

My confusion with this idea is whether it is easy or even possible to express interesting problems, and whether it can solve problems efficiently. This example strikes me as an over-engineered bubble-sort. I can't tell if he is just trying to provide inspiration with fun toy examples, or if this is a sketch of a more concrete idea that would make more sense but would be too technical for the type of video he's trying to create.

The other thing is, if this is talking about distributed computing, it seems like it might be leaving a lot out. He doesn't talk about how the grid maps to actual compute resources. Or how nodes discover their neighbors and so on.

I guess I could take a look at the paper ...

Why would she (knowingly) troll though? It seems like that would undermine the image of the project as a serious program for girls. Whoever is organizing the thing wouldn't want that, and no model is going to honestly get some pleasure out of sneaking in some clever troll posts to trick her own fanbase.

My guess is that the image you link to could have been someone trolling with a girl-coder spoof account, but then it got picked up by karlie's fans and retweeted by karlie herself, in earnest.

The image in this post on the other hand is probably made by karlie. She is probably only doing this project for image or maybe to help girls but obviously doesn't actually have the time or motivation to get very far in learning to program herself. So whenever it is time for a photo op she sits herself down and pulls up the first lesson until she has some shit on her screen that looks like code, snaps a picture, and then runs off to her next thing.

I don't see how managers could do anything differently to keep devs who are on the spectrum happier, when they also have to answer to the bottom line. There isn't enough budget to let the devs run their own show or work on interesting internal projects. We saw this with how 20% time fell apart at google. The best you can do is have hot young women working in the building and as recruiters, give the devs perks that help make up for their lack of self-care (free food, massages, organized activities etc), and stroke their egos by telling them they are working for the hottest company and that they are rockstars. And these methods have already been pushed to their limit, and they don't work that well. Their autistic urges will never be satisfied by the type of work they are given in a for-profit world, so they either work on curing their autistic side, or they get married right away so they can be autistic at home. Or if you're Amazon you pay the devs enough for quality Asian hookers and call it good. Which is the best strategy in my opinion. Because it won't be long before the devs figure out that they aren't getting laid by any of the girls around the office. And that when your employer is telling thousands of other dweebs that they are rockstars, paying them just as much as you, and every time you go out to the bars you and your type are dime-a-dozen ... You get the idea. It only works for a short time til they figure it out.

Would it be accurate to say though that *nix sysadmin is gradually moving away from shell scripts for more complex tasks? Not that you could necessarily get away without the ability, but that its importance is waning. I've heard that a lot of the scripts in modern distros are python.

The second round was a 2-on-1 interview with a whiteboard problem. I was asked to write a function findSum(array1, array2, sum) that returns true if two numbers from two sorted arrays add up to a given sum. I was able to solve it with my eyes closed using a double for loop — O(n²) time complexity. For the remaining time I was able to reduce the time complexity by converting arrays to hash maps, then finding the difference by subtracting sum from the first list element and checking if that difference is a “key” of the second hash map.

This is illustrative of this guy's lack of interest in problem-solving. Any interview question that I struggle with on the spot, I will spend time puzzling over after the interview until I get it. It's natural CS instinct and also good for future interviewing. (Although I agree that CS interviewing is pretty fucked up, and hate having to do it) I would also probably make sure to have it solved before writing about it in a blog post.

Given x1...xn in array1 and y1...ym in array2. Let i=1 and j=m. While i<=n and j>=1: if xi+yj == sum, return true. Else if xi+yj < sum, i++. Else, j--. End-while. Return false.

Pretty sure this is correct and would write out a proof if I had more time or wasn't on my phone.

This solution seems intuitive to me, and is even hinted at by the arrays being pre-sorted, which he apparently didn't try to make use of.

But the problem is, a bunch of play-the-game candidates will memorize tree-traversal (and other stock problem) solutions, so in comparison anybody who is trying to solve them from first principles on the spot will look unprepared or subpar.

Man, this hits the nail on the head. In fact, I think this is the antidote to impostor syndrome: If you have proven to yourself that you can write programs that you consider to be beautiful. Then, whenever there is some performance review or criticism or anything that would make you feel self-doubt, your brain instantly recalls your ability to write beautiful programs and then is able to divert the self-doubt to something external, or something internal that has nothing to do with programming ability. E.g. you will think "I'm a fantastic programmer, BUT it would have been impossible to perform well given:

• the shitty codebase I had to work with
• the messy and contradictory business logic in the specification
• the wrong tools for the job
• insufficient time or resources
• how little I am paid

OR (internal reasons not related to innate programming ability):

• my laziness
• my lack of interest in the assignment
• my lack of communication abilities
• my inexperience working on shitty codebases with shitty tools in shitty environments

Even with the latter set of internal reasons it's not likely to make you feel like an impostor. You're a good programmer, but you just hate your job. And who doesn't?

Probably you will have to write your beautiful programs in your spare time, but (rarely) there are little opportunities to squeeze in a fragment of beautiful code at work. It won't help to program in your spare time if you are trying to commit to open source projects or something, and not allowing yourself the freedom to be an artist about it.

Looking back, I never realized it, but if I did poorly on something programming related I always instantly had a rebuttal like one of the above which could place blame on something not related to by innate ability. Especially in college. Once I hit the working world and had long stretches of working on CRUD pieces-of-shit, my confidence started to slip a little bit. But as long as I periodically find some little opportunity to feel that "magic" again of being able to write insanely cool and beautiful little bits of code, I will feel rejuvinated. Sometimes I can squeeze it in at work or sometimes I have to do something outside of work. But it really works.

As a side note-- this is how many working artists stay strong in the face of criticism and market pressure. They retain some little sanctuary where they get to enjoy their own ability and believe what they are doing is beautiful and worthwhile for its own sake. Examples might be: practicing material that they aren't currently working on for their job, creating goofy side-pieces that aren't intended for any practical purpose, performing or creating material for friends or in a volunteer or philanthropic capacity, etc.

This is something that school and jobs won't teach you, and sometimes will actively mock. I remember trying to do something the "right" way on the job once and a project manager said to me "we aren't trying to create the Mona Lisa, here." You have to learn this on your own and be willing to maintain your belief that you, as a programmer, are a working artist, and deserve to be called an artist, even in the face of harshest criticism and embarrassing failure.

Most S/W Eng jobs are a grind. If you're really good and manage to go a more academic or research route, you can actually get a job that involves some creativity and the suspense of research (your work will occasionally answer a theoretical question whose answer wasn't readily apparent). I find the latter pretty motivating, and the creativity part helps pass the time and keep it engaging. But it is all just tinkering with machines or abstract concepts, so it will still never be sexy or cool and you will always yearn to be in nature or exercising or socializing (unless you're autistic and have never understood the appreciation for such things). The fact that you spend SO much of your time dealing with machines / systems leaves you craving these things, and it can feel a bit like every day after work is a process of rehabilitating yourself and re-discovering your humanness.

I don't know why anyone would consciously choose programming or CS just from looking at it from the outside. If you are going to be good enough that you are going to land a job that is actually sorta fun and engaging, you probably already are the type of person who will fall into CS without any practical consideration. You probably were a bit of a whiz kid, tinkered with machines or abstract puzzles a lot, and are naturally keen on being in your head and don't mind the solitude.

If you do decide to go into it as an adult because of practical reasons, consider that you are getting exactly what you asked for. If you wanted it to be cool and fun you probably already would have found enjoyment in it, and surrounded yourself with friends who think it's cool. If you're getting into it because someone told you it was something, well enjoy grinding away with a bunch of other followers.

Keeping the full title, "How to Safely Store Your Users' Passwords in 2016" probably would have avoided this misunderstanding.

Emotes isn't bad, I hadn't heard that before. Didn't we used to just call them smilies (even though they're not all smiling)? I am sometimes willing to say "emoticon", but definitely not "emoji." I don't know how the term came about but it sure sounds like it cam out of a marketing department, as opposed to users.

Great little demonstration. I am extremely fond of array-based languages and believe that one day their features will become standard, due mainly to the fact that parallelism is implicit in everything written in an array style, meaning that sophisticated interpreters will be able to automatically portion out parallel work to multiple cores or multiple machines in a cluster. And once you learn an array-based language, you get the feeling that "this is how computers think."

The way this implicit parallelism looks is that iteration/loops are implicit. As the "J for C programmers" book quips (J is a modernized version of APL):

I promise, if you code in J for 6 months, you will no longer think in loops, and if you stay with it for 2 years, you will see that looping code was an artifact of early programming languages, ready to be displayed in museums along with vacuum tubes, delay lines, and punched cards. Remember, in the 1960s programmers laughed at the idea of programming without goto's!

APL has its roots in mainframe computing where vector (SIMD) processors were the norm. One way to familiarize yourself in the modern day is to find an opportunity to program in MATLAB (or GNU Octave) or Python+numpy. MATLAB's array paradigms were inspired by APL, and numpy was inspired by MATLAB.

Here is an example function from numpy where implicit looping over multidimensional arrays is offered as a feature. The result is that the code will run faster because the iteration is pushed into low-level routines in the interpreter, instead of for-loops in the high-level language itself, and in theory could be parallelized:

numpy.matmul(a, b, out=None) Matrix product of two arrays.

The behavior depends on the arguments in the following way.

• If both arguments are 2-D they are multiplied like conventional matrices.
• If either argument is N-D, N > 2, it is treated as a stack of matrices residing in the last two indexes and broadcast accordingly.

[https://docs.scipy.org/doc/numpy-1.10.1/reference/generated/numpy.matmul.html]

The part about a "stack of matrices" means it will automatically iterate across all the dimensions other than the last two, so you don't have to write any for-loops.

What is really amazing about languages like J or APL is that this paradigm is directly codified in the language, and so it gives you this extensibility for free. While MATLAB and numpy have borrowed some of these paradigms, they are limited as languages and so these implicit-looping features are implemented manually as special cases-- only some of the time and often with limited support (often only working for 2-d arrays, for example).

The numpy.matmul function is a perfect case in point: they only introduced the "stack of matrices" functionality in a recent release. But in J it would have been there from the start, with the built-in notion of "verb rank." This is what the matmul function would look like in J:

matmul =: +/ . * "2

I haven't tested it, but it is in all likelihood just as fast as the numpy version. The J interpreter will recognize +/ . * as matrix multiplication and use BLAS just like numpy does.

The "2 is all that is needed to add the "stack of matrices" functionality, and in a J implementation it would have been there from the start because it is so easy, it is good practice, and it is obvious that matrix multiplication should be a "rank 2" operation (operate on 2-d arrays).

The J for C Programmers book puts it this way (in the parlance of J, which is a bit foreign to the uninitiated):

We have been focusing so closely on shapes, cells, and frames that we haven't paid attention to one of the great benefits of assigning a rank to a verb: extensibility. When you give a verb the rank r, you only have to write the verb to perform correctly on cells of rank ≤r. Your verb will also automatically work on operands with rank >r: J will apply the verb on all the r-cells and collect the results using the frame.

[http://www.jsoftware.com/help/jforc/loopless_code_i_verbs_have_r.htm#_Toc191734340]

J has its problems, and while I would say absolutely learn it to become a better programmer, don't expect to use it on the job. But I think this language is "ahead of its time" and this type of thinking will become the norm as computing refocuses on parallelism.

In my experience, it's not necessarily an intelligence problem but rather a passion one. I have never met a woman who had had spent their teen years alone at home coding. But plenty of men who have.

Yeah this is what insults me about the whole thing. It is part of the male condition that some of us spend our coming of age years doing geeky things instead of working on how to get laid.

When I first heard that any woman thought that there were social structures that discouraged women from programming, I thought back to how my parents let me have the old family computer in my bedroom, while my sister got nothing. I thought for a second "Huh, maybe it was unfair ... maybe she would have benefited from having the computer." And then I quickly realized, -- "Oh-- she would have just gotten herself into all kinds of trouble."

My sister is incredibly smart, but of course she couldn't resist the life of a normal young woman. She was always a bit jealous of my skills, of my talent when it came to playing guitar and writing music, and being funny. But I was always a bit jealous of the fact that she dated way more successfully and so on. It is just the condition of being the genders that we are.

I wonder if they are controlling for pull-requests on documentation files or README's.

There is no impartial automatic way to measure the quality of a code commit. It could be an entirely cosmetic change, or it could appear large because of some light refactoring that was easy to perform. Commit size and pull-request acceptance rate won't even begin to tell you what you're seeing.

From ars' resident snowflake SJW contributor-- http://arstechnica.com/author/annalee/

Who is dressed like Paula Poundstone while in a non-comedic role.

So are they not aware of outsourcing in the tech industry?

I think with tech / STEM the butthurt is mostly about the lack of women. Also despite being diverse in terms of lots of foreigners, I think SJW's still might not be content until there is good representation of all backgrounds. There are very few black programmers and engineers for example.

They should start with taking away graduation ceremonies for kindergartners and no kid left behind. Is this the result of it?

Not sure I follow. Do you mean that these practices are lowering the quality of our education vs other countries? With tech, outsourcing is going to be a given because there aren't enough people in the US graduating with those credentials to meet the demand.

it is likely they are trying to build a base to support a woman in the white house

Yeah I totally see this as a factor.

not that Hillary is incompetent, you can't be where you are without doing something right. Moral, ethical etc..are all a different subject

Yeah, it ain't a job just anyone could do, making it to her level. And that probably includes some valid and useful skills along with some unscrupulousness or sociopathy.

It's because it is a white-male-dominated field, and SJW theory dictates that whenever demographics are not in line with the general population that it must be caused by systematic oppression from privileged classes. That's why tech is their favorite target because despite all their efforts it has been impossible to influence the demographics the way they want, since none of the "oppressed" groups actually have any real technical ability or interest, and it can't be faked or fudged like it can in nontechnical fields. Furthermore, the tech sector is well paid and relatively respected and powerful, so that just magnifies their interest in conquering it even more.

The more puzzling question is why their narrative is so favored in the media. Which to try to explain that I think you end up going down a /pol/esque rabbit-hole. Not that they aren't right with their explanations, but it is just hard to make convincing sociopolitical arguments because it is so damned complicated.

Good on you for documenting and giving concrete examples of the limitations of the software on the homepage. Looks like a valuable tool for some use-cases.

Yeah you are probably right. It might depend on the position / industry / resume a lot. I think probably once you get away from entry level programmer for high profile companies, you'd get less BS candidates.

That's the thing though- if you have someone who comes in claiming to be an experienced coder, you don't need to be so condescending as to make them write a code snippet. It should be apparent just from talking to them. Ask them for a story about a time when they were challenged by a low-level programming situation on a past job. If they start talking about how a concurrency bug was introduced by the fact that an assignment to a 32-bit variable marked as 'volatile' was not properly ensured to be an atomic operation by the proprietary compiler for the 16-bit architecture they were working on, you'll have a pretty good idea they can code. Then you can ask them more questions about what the software did and what the implementation looked like and so on. How hard is it for a programmer to recognize another programmer in conversation? Trivial.

The only reason coding questions are needed is at a shithole like amazon where they have such high turnover that they don't even know what percentage of their current devs are capable of programming. They have to institute such policies because nobody plans on working there longer than a couple years and so they don't give two fucks about the capabilities of the candidate.

If you can't be transparent to your boss about your copy pasting, then that is a red flag.

Also, if all you know how to do is wire existing shit together, you won't be able to succeed in a programming job.

It's still a skill though and makes you more valuable than your average person who can't do anything with a computer outside of MS office. But this isn't usually its own position. Usually you would still need domain credentials in some field, get an "analyst" position or something like that, and then just use the computer skills to push your productivity to the next level. There are lots of non-programmerd who have the skills you do and it sounds like you'd fall into that crop. It's not a bad place to be but it'd be hard to get a legit programming gig.

Coming in python 3.6.

Thanks for sharing; I've always wondered why python didn't have the perl/ruby style interpolation.

Perfect time to make the switch to

truly random

It depends if you're talking about randomness in the physical world, or randomness in information (in the abstract world).

In physics, this would be the "hidden variables" that other commenters mentioned.

https://en.wikipedia.org/wiki/Hidden_variable_theory

In information, probably the "hardest" or most faithful definition of randomness would be Kolmogorov Complexity:

https://en.wikipedia.org/wiki/Kolmogorov_complexity#Kolmogorov_randomness

Basically, it is the question of whether the data can be compressed by any means.

It is known that strings that cannot be compressed do exist:

https://en.wikipedia.org/wiki/Incompressible_string

But finding those strings, or answering whether a given string is incompressible, turns out to be incomputable.

https://en.wikipedia.org/wiki/Computability

So you can't even assess whether something is "truly" random in this sense.

Even cryptographically strong PRNG's don't come close to having the property of Kolmogorov Randomness, which is actually by design, since the goal is to get a large output that "looks" random from a small program. And when we say "looks random" that means that the problem of determining the predicting the next numbers in the sequence is one that belongs to a particularly "hard" group of problems. The "hardness" of problems is a field of its own:

https://en.wikipedia.org/wiki/Complexity_theory

As for evaluating the "randomness" of some data or a RNG or PRNG in practice, you would probably rely on this (quite remarkable) test suite developed by NIST:

http://csrc.nist.gov/groups/ST/toolkit/rng/documentation_software.html

The tool performs a myriad of statistical tests. If the data fails one of these tests, you can guarantee that it is not random. But if a string passes all of the tests, there is still no guarantee it is random (if there was this would contradict the proven incomputability of Kolmogorov complexity that I mentioned earlier). But it is considered a good benchmark for randomness. If your hardware-based RNG passes, you are in good shape. If your PRNG passes and its design is theoretically sound from the basis of the "hardness" aspect I mentioned earlier, then you are in good shape. An example of a PRNG that wouldn't qualify would be one that outputs the digits of Pi; while the digits of Pi do indeed "look" random and successfully pass all the tests in the suite, what's missing is the computational "hardness" in its design. The matter of predicting the next number in the sequence is as easy as finding where you are in the expansion of Pi and then following along.

The test suite includes a companion paper that will make your head spin; it is really in depth. The first chapter is an overview though and is pretty readable.

It does feel like there should be something such as "truly random." But when examined deeply it is a very murky issue, and the best you can do to understand it is get a relevant PhD (or two) and stare at it from different directions with a sense of appreciation for how tangible it feels while simultaneously being totally contradictory and unexplainable. It may ultimately be an unanswerable question, and further, may even be unanswerable whether it is answerable (if you can wrap your head around that).

To me, it seems like there are abstract reasoning capabilities, visual thought processes, etc, that make one naturally talented at mathematics, programming, engineering, and certain aspects of music and arts as well. If you have that type of brain, you'll have good potential to become competent in those fields. But it may be the case that some folks just don't have this capability, or that you have to have some exposure when you are young in order to develop it. These people are still drawn to programming out of fear of being unemployable, and for status or trendiness, good wages, etc. Or they are feminists trying to prove that they can do anything men can do. These folks in general are going to end up calling themselves programmers even though they are just copying and pasting things together. Not that this isn't a valid set of expertise or an art in and of itself. But they aren't programmers and it's virtually guaranteed that the media won't make any attempt to distinguish between these and actual programmers.