The sales of computers are going down last years, and there are more other devices in the age of "Internet of Things" that are harmful for the freedom of the users. Even simple climate control is not your device, but is designed to spy on your family habits, "phone home" - all in the name of optimizing your utility bills. In the US the practical disadvantage of this unfreedom can likely be just unsolicited junk mail, in other countries with higher corruption levels this data can be sold to burglars who will visit your home when the heating/AC is set on "vacation" level. Freedom of the users of the devices (including all 4 classic components) is much broader then just that of the computer operators, and the choice if free/non-free is not only about your philosophy, but be a matter of survival.

As for Richard Stallman - I just can not see, where is that his one mistake. For years there are continuing attempts by others to create "hardware GPL" but there is no universal solution (we use exactly the combination mentioned in the article: GPLv3 for software/firmware/FPGA and CERN OHL for the hardware) caused by fundamental differences of the software and hardware. His last year article http://www.wired.com/2015/03/r... provides a lot of practical instructions how to build free hardware in a not-yet-so-free hardware world, there are "levels of design".

We are using Xilinx FPGA for 13+years, when first choosing between Xilinx and Altera the decisive factor was the license terms - while both were proprietary, Xilinx zero-cost software had no expiration, while Altera's one had only 90 days. Our products are based of Free Software/Open Hardware (licensed under GPLv3 and CERN OHL) so it is critical for us to avoid expensive tools as our users would have to use them too just to be able to rebuild the executable image (bitstream in the case of the FPGA) from the source code provided with the products.

Unfortunately Xilinx is gradually migrating away from openness and freedom, working with Zynq I noticed quiet a few undocumented hardware modules, even some primitives now can only be simulated using encrypted code (and so not compatible with free software simulators) - you may find my opinion on this issues in "FPGA is for Freedom" blog post.

Intel is known to be more free software friendly. Do you think that such attitude will get to their new Altera FPGA department? Should we at Elphel already start migration to Altera/Intel?

Things did not change much (to the better, at least) - http://blog.elphel.com/2013/10/fpga-is-for-freedom/

Yes, you are correct. And the NSA acronym also has the same source. Maybe the correct formula is
DHS+NSA == KGB
But honestly - I'm afraid that NSA has much superior capabilities against the citizens of the country than the KGB had. With the brand new facility some 15 miles from my home in Utah.

For the first 36 years of my life I lived in the USSR. It was un-free country, but many of us valued freedom and learned to love it more than the state religion - "Communism" that the government preached. When Bush started his war in Iraq as a retaliation to the 9/11, I noticed that with Patriot Act, "security" in the airports this country started to grow more and more similar to the failed state - Soviet Union. Some features that I believed to be unique trademarks of the totalitarian states sneaked in the everyday life of Americans. There was a government slogan in the USSR - "to catch up with and pass the USA", but now it seems that the USA is trying to catch up with the KGB-ish nature of the USSR. It is sad for me to see that many born-Americans believe that Freedom is given to them by God, by their brave predecessors or just by the Land they are born on. Freedom has to be fought for by every generation all over again, citizens have to prove they deserve it.
I would never vote for Republicans - for me they share much more with Soviets than just the red color, but when Obama (whom I voted for) calls Snowden a "traitor" (instead of a hero), I'm thinking that Putin in his place would do exactly the same. Putin, who's main enemies are Russian citizens.

https://encrypted.google.com/search?q=fahrenheit+451+hound

When Apple will completely move production from China to US - will they employ more Americans or less than now?

Utah as a state votes 2:1 for Republican candidate, but in Salt Lake City alone it is about 50/50. And Rocky - when Bush's war was going on and he personally was visiting SLC, Rocky organized anti-war event. I was trying to imagine mayor of Moscow (city where I lived so far most part of my life) greeted Mr. Putin in a similar way.

I'm Xmission customer for 18 years and they are the best. They always notified subscribers of any interruptions of the service even if it happened for 5 minutes in the middle of the night, decribing what went wrong and what have they done to prevent similar problems in the future.
And I still drive with Pete Ashdown sticker on the back of my car since he ran for the US Senate - but it is not easy do win for a Democrat in one of the most Republican states.

Yes, you are absolutely right - QE is already high even in "boring" cameras. It is also true that the readout noise of the modern boring sensors used in cellphone cameras is also of the order of a single electron (each photoelectron counts), so you can not significantly increase low-light sensitivity for the same pixel size and same exposure time. You can make large pixel (or use "binning" of the smaller ones) - and it is well known part of the existent technology. Other parameter important for the low-light imaging is dark current - this is why many astronomical CCDs are cooled and exposed for hours - just to be able to increase exposure. Exposure is not "exponential", in is not less linear than QE, so if you increase exposure 1000 times you'll be able to detect 1000 dimmer light source, But for video that is irrelevant - boring sensors already have negligible dark current for video, so this (the only way to increase "sensitivity" for the same pixel size) is only applicable to long exposure scientific applications.

There are some sentences in the article that indicate that they mean exactly that - using longer exposure:

1. 1. "which "trap" light-generated electron particles for a longer period of time, translating into a stronger electrical signal ". You see? "longer times"
2. 2. "graphene sensor, being more sensitive, will negate the need to increase ISO in low light settings". Increasing exposure does not need to increase ISO. People made nice low-light photos with ancient low-ISO photographic plates using long exposures. Isn't it the same?

And of course, talking about "5 times less expensive" without having a viable technology of integrating graphene with CMOS - what the crap is it? You know why silicon-based CCD technology is dying now? Just because you can not combine CCD (light capturing array) and CMOS (phase drivers, ADC, memory, CPU if you font SoC) on the same silicon chip - I do not see how carbon graphene is more compatible with CMOS.

I believe that the decreasing demand for STEM educated people and the desire to get only "the brightest" is just another indication that the robots are taking over the engineering/IT jobs too. In the past each leading engineer needed a bunch of average ones who would do routine work (and learn in the process) on the design of the Master. Now "the brightest" can just use more of the CAD tools, so it is not just the uneducated low-paying jobs that are eliminated by the "robots", they are coming for our jobs also.

Lithium deuteride ( http://en.wikipedia.org/wiki/Lithium_hydride#Lithium_deuteride ) can look like a grain of salt literally - it is just not as clear as NaCl.

We (Elphel) are an Open Hardware company for more than 11 years, all our products are distributed exclusively under Free licenses (GNU GPL, GNU FDL, CERN OHL). And yes - customers are willing to pay extra for the freedom they get. Of course it is not an easy business, production volume rarely can go high as most application require either single are just a few units, but our products are used in most US National Labs, NASA, many universities and research centers around the world. And it is fun to develop such stuff that can be used in some innovative ways we would never think of ourselves, so we try to combine high performance with "hackability" - this is the minimal combination needed for most scientific applications. So we do not consider "Open Hardware" as some DIYish and simple stuff only (it is very important, of course, we love Arduino). And would never use "openness" as an excuse for inferior performance, would not develop "poor man's" replacements of the real proprietary stuff.

As Big money+robots combination is more efficient in more and more jobs, the communist-socialist outcome is not imminent. I see new emerging way of living - we'll all become pets of few rich ones, who'll be able to mine on asteroids without any need for us.

You know - Bill Gates is supporting Africans in Africa, financing their cure from AIDS and probably more. And it is not because they are just all lazy and do not want to work - but they are already not needed in the world economy. So when robots will take all our jobs - and I mean "all", we can not hide behind our high education, PhD - it is just a matter of time - when robots will get our creative jobs too. So, like a nice guy - Bill Gates, the owners of the New Virtual World will just feed us and give some money to live and buy their Internet-advertised goods.

But we will not have any rights, We will b e just pets.

Building "open source" cameras for more than 10 years I would say that the codecs designed fro video distribution may be not the best for the cameras, where you have to preserve as much as possible of the original sensor data while having reasonable compression. It does not need to be completely lossless (as for editing) - the sensor (and just the physical world itself) has some noises, the the compression errors should be just below that. For the video distribution the task is different - reduce bandwidth while preserving _perceived_ video.