Some people seem ready to hate on particular features but it seems more of a fit and tooling problem to me. If things don't fit your brain then it can be hard to appreciate them. If the tools you are comfortable using can't take advantage of particular features then those features become worthless to you. To take a stab at some of the common complaints:

Dynamic typing and the way variable scope works in Python means that a compiler can't catch as many errors. To people used to using a compiler to catch errors this can be very aggrivating, their workflow and productivity will suffer because of this language aspect. On the other hand the people that use a unit test heavy workflow will benefit from this aspect because it reduces compile/start up times and makes writing unit tests much easier. Also if you use an IDE with context aware autocomplete then static typing really helps it help you. OTOH If you like to do most of your debugging within a running program from a gdb/pdb prompt then things being more dynamic can be very helpful.

Syntatically significant whitespace can be great for enforcing style guides for checked in code. It promotes writing smaller units of code (by making long code blocks horrible to work with) and that plays well with TDD and code readability for some people. It can also get in the way of development and stops tools from automattically formatting code for you. It is also a pain when working at interactive prompts and makes that style of interactive debugging harder. Just a tip: if you are forced to work with significant whitespace and hate that you can't even see it then try getting your editor to highlight it differently. I find having spaces, tabs and trailing whitespace all use different background colours most helpful.

Run time speed is often an issue for at least some of the functionality in any given project. The tooling that languages give to handle this can be too specific so people can end up hating the languages because the things they try just don't work. For instance if you are used to using threads for performance then Python is going to be a bit miserable for you until you stop trying that. Some people prefer sticking to just one language for a code base, this favours using something that can go as low-level as you need for the performance critical parts. Some people use something more comfortable for most of the development and switch to a different language just for the performance critical parts. Python with Cython when you need it can be a great combo. I think the ML landscape is a good example of combining high level conveniece (Python) with low level compute (CUDA/OpenCL) to get things done both quickly (developer time) and efficiently (hardware time).

Scripting vs real language... I personally don't get this and haven't heard any salient arguments about it, just bigotry I guess. I've worked with interpreted C and compiled Python so I don't even know where people think the boundaries are. Maybe LISP changed me too much to comprehend :-)

Wishing people happiness with their languages and tools, and the understanding that the two go hand in hand.

I don't think anyone is saying that the kids don't matter. The compromise is between the power of the state to fight crime, and the ability to protect people from the state should it start acting malicously. Sexual abuse hurts people and is an extreme often used to justify more state power. Holocaust hurts people and is an extreme that can be used to justify less state power. Deciding how much you would like to enable the state vs individuals is complicated precisely because people do matter and there are threats all over the place.

Just considering those two extremes I wouldn't like to guess at which has caused the most harm over the last 100 years. I am more concerned about harm being done en masse though, which is why I favour private communication and feel it should be protected as a human right.

Having a working Python implementation will also give you a better understanding of which parts are performance/memory sensitive. This may help guide you while rewriting in a different language, or you may find that you can achieve your goals just by hot spot optimizing your Python code using some of these fine tools:

http://www.numpy.org/ can give you compact arrays of unboxed types and fast operations over them.
http://cython.org/ is an amazing and versatile tool which allows you to compile your Python code, optionally add type information, optionally manage memory yourself, optionally interact with C/C++ code very easily.
OpenCL/CUDA if your work can take advantage of them.


If the standard Python runtime isn't simply too big for your project I can't recommend Cython highly enough. Only tackling the parts that need it is one of the keys to successful optimization and Cython lets you do just that even if it is just one loop in an otherwise pure Python file.

It isn't just the end result that is important, it is also the means.

Would we celebrate bank robbers, drug dealers and fraudsters if they were to spend a little of their money on charity? No we wouldn't because that money was gained as the result of illegal and immoral actions.

While Gates was running Microsoft it was found guilty of abusing its desktop monopoly to take out competition in other markets. Arguably their dominant and stagnant position has set back the computing industry many years. Significant amounts of money were made as the result of illegal and immoral actions taken while Mr. Gates was running the show, and this has cost society. Giving a little back now doesn't make it alright.

Yes, it requires active cooling to shutdown cleanly, as in be fit for starting up again in the future.

Not at all. When the cooling fails you get a meltdown, which the containment shell is designed to contain. There is nothing wrong with a contained meltdown as far as safety is concerned. The problem with Chernobyl is that there wasn't a containment shell, and uncontained meltdowns give meltdowns a bad name. The other problem is just the cost of cleaning up the sealed ball you are left with, anything less than a meltdown is vastly easier to deal with using existing processing facilities.

If it wasn't for the containment shell, then fair enough. I'm shocked that Chernobyl didn't have one, they only have to withstand temperatures of 500C, so a good strong steel is more than adequate.

These reactors are designed to be passively safe. When the control rods slide into place the main reaction stops and it is safe to walk away. With them in place and a cooling failure all you will get is a harmless meltdown, which is a low temperature affair where the entire core gets wrecked but is easily contained by the containment shell.

One reason that people are working to maintain adequate cooling is because a meltdown puts the containment shell to the test. This is a test that it should handle with ease as meltdown temperatures are around the 500C mark and the shell can withstand over 2000C, but it would still be better not to have to test it at all.

Another reason is that a wrecked core is a significant financial loss - in terms of the core, all the useful material it still contains, and the difficulty of cleaning it up. If cooling is maintained perfectly then everything can be saved. Even by resorting to sea water for cooling (which makes the core permanently unusable) the system can be cleaned up easily, and the fuel can be reclaimed.

Of course people are still evacuating and for a very good reason. An unparalleled disaster has just wiped out all the backup facilities necessary for a clean shutdown. There is now the potential for a meltdown, which means that the last line of defense (the containment shell) would be tested. As reliable as containment shells have proved to be there is simply no point in taking risks.

Thankfully the work done to maintain cooling means that the meltdown may still be averted, and it is also giving people time to evacuate just in case there is a meltdown and a (very unlikely) containment failure.

EA also produce Battle Field Heroes which is free to play. They originally stated that the premium content would be limited to perks, and not give a significant game play advantage. They later decided to change that, so although you can still play for free you are very much limited to being cannon fodder for those that have bought the extremely powerful abilities and weapons.

Why should we expect anything to be different with Lord of Ultima? In this sort of game you can spend months building everything up, and someone else can destroy it in a week, so people are going to feel compelled to buy competitive advantages if they are available.

Free to start but expensive once you are addicted, this is becoming such a common business model it needs its own word.

If you are a content creator and you can't make a living, then do something else. There are many jobs I would like to do for a living, but I wouldn't get paid enough (or at all) so I do something else, it is quite straight forward. In fact if this bill goes through I will probably be put out of business (the whole hosting industry in this country will probably be collateral damage) so I will in fact do something else, thanks.

Of course it is a shame that people copy content without paying for it, but it isn't really a problem compared to the problems of other businesses, and certainly not worth breaking something as important as the internet over.

Shops for instance rely on physical goods. It is illegal to steal them, and yet this still happens. Shop owners spend a lot of money on physical security, and insurance in case things get stolen, and yet still don't campaign for the right to have people stopped from going near their shops (i.e. some form of imprisonment) without any proof of actual theft. When comparing "theft of intellectual property" to actual theft it is clear that the content industry has it easy, and this bill is only about alleged infringement, not actual infringement, to make matters far worse.

But now we have an industry that has relatively few problems demanding that the internet is damaged for no good reason, and the people using the internet should pay for this expensive whim. I say no good reason because (1) no one has proved file sharing is actually bad for said industry (2) the industry seems to be doing well despite the global recession, and (3) the only point that really matters: no one has proven that the content industry is more important to more people than a free and open internet.

I can live without TV, without films, without music, and it really wouldn't bother me. If all those industries disappeared overnight then fine, so long, and thanks for all the fish. The internet, however, improves my quality of life so much (as well as providing my income) that I would gladly keep it over the entertainment industry.

Liberty is more important than entertainment.

To give them a reasonable chance of making a good decision make sure they understand that when development time is limited C is rarely going to produce the fastest results.

Given the obvious (and hopefully obviously wrong to anyone with experience) observation that C is harder to program but runs faster, and Python is easy to program but runs slower, I think a lot of people would go for C because it seems to be better placed for winning the competition.

However, code normally has bottlenecks, so 95% of run time will be spent in 5% of the code. Optimizing the parts that aren't bottlenecks is a severe waste of time, and will surely affect your competition results. Anyone who knew what they were doing and wanted to win would (1) write the bulk in a high level language, (2) use a profiler and identify the bottle necks (3) spend their time trying to find better algorithms at the bottle necks, (4) finally optimize the code at the bottlenecks, pushing it to a lower level language if skill and time allowed.

Going with the OP's original inclination would work well here. Python can be used for stages 1-3 and will give students time to try more algorithms on the all important stage 3. For stage 4 you can use c to create a dll/so and use the ctypes library to drive it directly from Python. If anyone stumbles at this stage then it isn't too much of a problem, they already have a viable entry they can submit and be proud of from the end of stage 3, this extra bit is just for the really talented to go for 1st place.

Not only would this approach give the students an advantage in the competition, but also a really good idea of how to tackle projects in the real world too.

We set up an authorized SSH key on every VPS at provisioning time, and tell our customers about it. However, as you state, having access is one thing and having permission to use that access is another entirely. We will only access a customer's VPS with their explicit consent, to do anything else would be immoral and illegal.

This would be quite different if it was a managed VPS, as that implies permission, otherwise how would the hosting provider be able to manage it? However, it does not seem to be the case here.

If the host found that the VPS was doing anything against their TOS (they suspected malware, and could have confirmed that from the network traffic) then the proper cause of action would be to notify the customer, and if they deemed it necessary they could have cut it off from the network until the customer responds. Attempting to gain unauthorized access is not the right approach, it is a quick way to lose customers and it is also illegal just about anywhere with computer crime laws.

How is that? They both support full virtualization, and they both support paravirtualization for things like networking and disks (they even use the same protocols for this - virtio). Most of the debate I have seen regarding the two is whether it is better to have a specialized hypervisor (Xen) or use a general purpose, but heavily developed and very much optimized, hypervisor (Linux). The performance benchmarks are six of one and half a dozen of the other so this debate doesn't seem overly relevant.

I'm not sure why you think this is the case, I consider them to be pretty much equivalent. The reason we use KVM is because Xen wasn't up to date enough to run on our brand new hardware. KVM does a fine job as a Xen replacement, and we have been using it as such for years.

But we were all told about the 10 ARM netbooks that would appear on the market by Q3 2009. It is now Q4 so they must exist, and you must be wrong.

I'm pretty sure this is a Microsoft stunt to make their market share look better. If you can't make geeks buy Windows, then make sure they don't buy anything at all because of all the sweet smelling vapourware on the perpetual horizon. Then again I'll blame them for most things, including a sock I lost.

Or listen to swedish death metal, there might be lyrics at some imaginary level, but nothing I can discern. Literally something to scream along to.

One place I worked had a guitar we could use when we wanted. Again, no lyrics, just people quietly playing away to the best of their ability, easily drowned out by headphones if necessary.

There is something about the rhythm, pace and harmony of playing on a guitar for half an hour that seemed to help me make the right changes to the code when I went back to the computer. This was a place where the importance of making the right changes (and preferably only the right changes) was well understood.

You can achieve live migration with iSCSI and AoE too, and if you use a SAN you will probably continue to use one of these network block device protocols.

What DRBD does it make it relatively simple to set up a redundant SAN, using commodity hardware, from which you can export iSCSI devices etc.

Of course if you are going to use local storage for your VPSs it is just as easy to set DRBD up on those hosts and forgo any network block device layer on top of it. Dual primary mode makes live migration in this scenario particularly convenient.

I'm not about to dismiss your experience, but things have changed over the last 15 years so it might not be as relevant as it once was.

In that time processors have become much faster, memory has become much cheaper, commodity servers have also become much cheaper and a lot of software has become free. While that has happened hard disks have become only a little faster. As a result many people consider custom hardware for driving those disks to be unnecessary - generic hardware is more than fast enough and is significantly cheaper.

There might still be some compelling reasons to go with expensive redundant SAN equipment, but for many situations a couple of generic servers full of disks and running Linux and DRBD will do an admirable job. The bottleneck will most likely be the disks or the network, both of which can be addressed by spending some of the vast amount of money saved by not going with typical enterprise solutions.