Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?

Comment Re:Math is a Chore (Score 2) 218

Math IS a chore. Learning IS a chore. People need to realize that not everything in life is "fun". You need to do the chores in order to get work done. Too many people don't want to put in the work.

While that might be partly true, it is also true that Math education is a chore because it was treated as a process of memorizing, not discovering - memorize process x,y,z so you can answer contrived questions a, b, and c. There is an excellent essay on this topic: A mathematician's lament.

Comment Re:Think? (Score 2) 522

You obviously don't have a teenage daughter. The formal definition of broadband is this: A man's wife and daughter can watch two different Netflix movies simultaneously, and he can still get work done.

We had a similar problem. My wife grounded our daughter and threatened to withhold sex from me, and now she is very happy with the internet speed.

Comment Get a projector (Score 1) 186

The only way I could get pair programming to work was to get in a room with a projector, and give the one developer the control of the keyboard. The other developer gives instructions and directs the flow (while also writing things out on the whiteboard to teach the other). If some changes/fixes take a lot of time, this lets the both developers work on their own laptops.

This worked irrespective of whether the junior/senior dev was projecting. It is more hands on for the driver, while giving more flexibility to the partner to do their thing. Traditionally the junior dev controls the keyboard, while the senior dev is like a mentor - they don't take the wheel from junior, but they don't need to be wasting time while code edits are happening.

Comment Re:Trump just says stuff (Score 1) 875

Actually, this is what happens when people can arbitrarily pick and choose their news. They can live in a reinforcing bubble of news and information - limiting their sources to those that will align with their views and make them more comfortable ("See - I can find 5 blogs that say exactly what I already know. Therefore, I must be right and smart").

Say what you will about the old print media, but you would at least be forced to glance at other articles while skipping to your favorite section.

Comment Re:Google shits out another fail (Score 1) 91

Another cigar feverishly brainwashes an unruffled bodice ripper. He called her the Interloper (or was it Timosha?). A somnambulist daydreams, but the bride beyond a ballerina eagerly steals pencils from a toothpick. Jacques, although somewhat soothed by a labyrinth near a snow and some seldom lowly gypsy, still can be kind to her from the sprightly gypsy, can be kind to her a trombone related to a maestro with a likeable onlooker, and slyly boogies the dark side of her alchemist.

She prepared to enter The Scary Door.

Comment Re:I don't understand. (Score 3, Informative) 129

Consider minimizing x^2 when x can take values in -10 to 10 (we know the answer is 0, since we only consider real valued numbers). If we wanted to solve this problem, there are several approaches; some example approaches are: randomly try a lot of different values, set the derivative to zero, or try a cutting plane algorithm. In general, we might not know the analytical expression for the function we are trying to minimize (or it might be too complex) so we can't really find the derivative efficiently. Derivatives can also be computationally expensive to compute, so let's ignore that approach.

What we can do is to say let me find a solution for which the function is less than some threshold t, and then keep reducing t till I can't find any more solutions; this is what the article meant by finding a smaller circle inside a larger one (for each value of t, I try to find solutions that are smaller than t).

What cutting planes do is chop up the original function in to pieces - in some pieces, I know the value will be much larger than my threshold (so I don't have to search in those pieces), and in others it might be smaller - I focus on searching these pieces to find a value that is smaller than my threshold (after which I can reduce the threshold and try again). This is what (in a simplistic sense) cutting plane algorithms do; they chop up my search space.

How we select the points for chopping is crucial - bad choices (say choosing one point after another at random, or trying points very close to each other), I spend a lot of time chopping unnecessarily (or not benefiting from chopping by much). We also want to make sure our cuts really do divide the problem in to pieces which I can discard searching, and those pieces I discard should (ideally) be quite large. Until this work, the time taken to decide where to chop was n^3.373; they brought down the time to n^3 (where n is the number of variables that the function I am trying to minimize takes as inputs).

They also said that for special classes of functions, they can really improve the total computation time significantly (from n^6 to n^2).

I'm glossing over (and am certain I've got some details wrong) many issues to give a taste of the big picture idea of cutting plane approaches in general; there has been decades of work on these problems that you can read (I recommend anything written by Prof. Stephen Boyd as an introduction to some of this research).

Comment The two key contributions (Score 3, Interesting) 129

Some very interesting results, but the two key contributions are (almost verbatim from the article):
1) With the best general-purpose cutting-plane method, the time required to select each new point to test was proportional to the number of elements raised to the power 3.373. Sidford, Lee, and Wong get that down to 3.
2) And in many of those cases (submodular minimization, submodular flow, matroid intersection, and semidefinite programming), they report dramatic improvements in efficiency, from running times that scale with the fifth or sixth power of the number of variables down to the second or third power.

So this seems to be a better cutting plane approach that improves the cutting process by reducing the time to find the next test point (in general), and for certain structured problems (like SDPs) this approach reduces the computation time significantly.

This does raise some interesting questions, such as: is there a limit to how fast you can (in a general problem, not using a specific problem's structure) find the next test point? Even if we don't know what algorithm gets us there, it would be useful to have a bound for future research.

Comment Re:Separate H1Bs (Score 1) 112

There is jackass. There are different Visa programs: O-1 and EB-1. How the fuck did the US government spend $500,000 on your education? Unless you are talking about research grants, in which case they would be available to other researchers, not just you. Move along troll.

O-1 is an extremely hard category to get a visa under. EB-1 is not a visa, it is a permanent residence stream (which is what my employer can file for my green card under, IF they choose). Again, EB-1 is extremely hard to get a green card under - it is for exceptional researchers who are presented as such valuable resources that the country would suffer a significant loss if they weren't allowed to remain (hint, most Ph.D. holders are not, despite what their lawyers might argue).

As for who was eligible for the money - yes, the money was available to my advisor to spend on anyone he chose. He chose to spend it on me because he felt my contributions were valuable. The fact that it could have potentially gone to someone else: why is that relevant?

And how did they spent nearly half a million on me? I got a stipend, a tuition waiver, health insurance, travel reimbursement, the occasional computers/software, and full pay in summers for 6 years. So that is nearly $50,000 per year (I was in an expensive state, my initial stipend was over $2k a month) that is shown to the funding agency (NSF). On top of that, there is a university overhead: For every $1 my advisor could spend, the funding agency would have to give $1.5-$1.75, while the university skimmed off the top. That adds up fast, and I rounded up.

Comment Separate H1Bs (Score 5, Interesting) 112

As someone who has benefited from the STEM extension, it is strange that they are targeting this, instead of fixing the H1b issue.

I got my doctoral degree in STEM, and did not get my H1b in the lottery system the first time. If I was forced to leave, the US would have spent nearly half a million dollars on my education, and got one year of tax (not counting my research work, which is freely available to anyone) in return.

Like most people making use of the STEM extension, I am being paid as much or more than my US co-workers. This isn't a "consulting" gig where I am forced to work for my company at sub-standard wages under pain of getting kicked out of the US - STEM graduates have been educated in renowned US universities, and I had four job offers by the time I graduated.

I think there should be a different H1b tracks for people who are hired "internally" i.e. the person is already in the US, and was educated here (people who currently benefit from the 17-month STEM extension), and the other type of H1b that I hear exists (where a company brings in people from overseas purely to do a job).

Comment Re:Translation (Score 3, Interesting) 85

We want employees that work for free!

No, what he wants is to allow paying users the ability to have a lot of the benefits of open source, while not reducing the company's downstream cashflow. It might come as a surprise to you, but not every project can be open sourced and still make money. In fact, I always wondered about the pay-for-support model - why should I make my software easy to use and maintain (eve at large-scales) if my main source of income requires users to come and pay me to help with the product?

Also, this is a scientific computing software - the target market is small as it is. Very likely the core is something that is highly mathematical, and not something the average programmer can work on. I wouldn't be surprised if there were several statisticians and researchers on the payroll. People like these don't necessarily write high quality code - see the R library for example (yes, there are many excellent packages, but a huge number are written sloppily by academics). The customers who use it might need highly specific tweaks for their application - something they can only do with the source. I myself would have loved to "fix" some of Matlab's functions for my specific use case. But if the customers don't have the source, there isn't much they can do.

This type of license should actually become the norm - but it is unlikely to happen. There is always the risk that a customer who bought software might give a copy of the source to their friend, and before long I lost all control of the product. While binaries are still copied today, companies are trying to crack down on it through registration and DRM. If the source is out, there is even less they can do to stop it.

Comment Re: Harvard is the right place (Score 4, Insightful) 348

Also, see "I haven't called the police in 45 years, I don't know why my taxes go to pay them."

While you might not have directly gone to a clinic in 45 years, you definitely have benefited from good healthcare services. Notice how the US doesn't have polio, ebola, or any other epidemic? You're welcome. Now you can also get some meds for the delusions of tyranny.

Comment Assets valuation? (Score 1) 335

I wonder how they value assets? Clearly, sharing a few songs is theft of several million dollars - which makes my MP3 player a treasure trove. Also, patents and other intellectual property cannot be cheap at any price.

I'm curious as to the exact valuation methodology. The links take me to the federal reserve site, which links to balance sheets. While it does list intellectual property, how do you accurately value something that isn't sold/bought on a market? If a company holds a patent that they never try to sell, how can it be valued accurately?

Comment Ask the users. (Score 3, Insightful) 244

The problem with a lot of documentation is that they aren't written from a user's perspective - they are written by people who wrote the software, and know what to do. Letting go of your design assumptions is almost impossible.

I have long felt that the first draft of documentation must not be written by the person who wrote it; you have to allow your users to "send" you the first draft (either through email questions/screenshots/etc.). Then you realize how many assumptions you made that are non-obvious to your users.

Obviously, this isn't really practical for OSS - you might not be able to pay for usability testing and feedback. Which is why I prefer to include screen shots in documentation as much as possible. Also, I try to follow this basic formula for documentation: What (what is the user trying to do - make it clear what this section of the documentation address), How (how can the user achiever her goals), Why (this is where you might, if you choose, try to explain a design/implementation philosophy - it comes third, so that someone who doesn't care doesn't skip the entire section. Clarity and brevity are important.).

This is the principle I follow for user stories as well - create an end-to-end user workflow (which is basically just many small What/How/Why sections tied together).

Slashdot Top Deals

You have a tendency to feel you are superior to most computers.