Wikipedia:Reference desk/Archives/Computing/2015 April 5

Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

April 5

set, export, environment variables and finding pythonpath

Where are the paths that make up the PYTHONPATH stored? When I run the set command (exactly set | grep PYTHON or set | grep python), I can not find any reference to them. I am aware that environment variables can be set with the export command, but where does the result of, for example, export PYTHONPATH=$PYTHONPATH:foo/bar go to?

I tried finding them in /etc/environment, but there were only PATH="/usr/local/sbin:/usr/local/bin: ... there. In .bashrc or .profile there is nothing related to Python either. There is no .bash_profile file.

So, when I go to the Python console and type sys.path, where do all these path ('/usr/lib/python3.4', '/usr/lib/python3.4/) come from? --Senteni (talk) 01:54, 5 April 2015 (UTC)

sys.path consists of the path of the current script, the contents of the environment variable PYTHONPATH, some platform-specific library directories, and whatever is added to it by Lib/site.py. PYTHONPATH, even if it exists, isn't the complete search path; it's just a list of non-standard directories to add to it. -- BenRG (talk) 20:26, 5 April 2015 (UTC)

Quantum multithreading of classical programs?

Would an n-qubit quantum Turing machine, in general, be able to simulate O(2ⁿ) classical computers at once? If not, what's the best it could do if replacing a large set of independent or loosely-coupled classical computers? NeonMerlin 04:54, 5 April 2015 (UTC)

Did you throw as many tech terms and buzzwords into that as possible? Makibromo (talk) 18:18, 5 April 2015 (UTC)

The best it can do is one. Even if you run the same algorithm on a superposition of inputs, you can only measure the result for one input. -- BenRG (talk) 20:06, 5 April 2015 (UTC)

Could I at least choose one of the inputs based on the result, and know which input had been chosen that way (e.g. if one of a suite of independent unit tests were failing, find out which one)? NeonMerlin 00:13, 6 April 2015 (UTC)

No. That would be a linear-time SAT solver. You can use Grover's algorithm to find the failing test in O(√n) iterations (vs O(n) classically). -- BenRG (talk) 03:01, 6 April 2015 (UTC)

Computer station

I spend $4,000 on my computer and I posted a picture of it online and someone commented that it was a "tryhard ricer station". What does that mean? JamieAllysonApril (talk) 18:15, 5 April 2015 (UTC)

Something very uncomplimentary, I'm afraid. "Tryhard" means you've tried hard to build a good system but not succeeded, and see "riced out" for the meaning of "ricer". Tevildo (talk) 18:29, 5 April 2015 (UTC)

DVD readers in 100 years?

Assume that you write on archival DVDs now and that the disk hasn't deteriorated in 100 years? Will there be hardware that can read them in 100 years? (Many, many formats have gotten to where there is no way to read them.) Bubba73 ^{You talkin' to me?} 22:29, 5 April 2015 (UTC)

Who knows what electronic devices will be made in 100 years. Experience with recent storage technologies (e.g. VHS) suggests you might see 20 years of backwards compatibility on ordinary consumer devices (CD is a rare exception with a longer window, most things are shorter). Beyond that you'd depend either on new machines (made for librarians) which can read DVDs, and on trying to use antique drives. The latter isn't easy either - consumer electronic devices like DVD drives aren't made for long shelf life. Plastics offgas, insulators cure and become brittle, and capacitors leak. People trying to maintain 1980s era microcomputers usually find they have to replace at least the capacitors in the power supplies. Many of the subtler failure modes of modern electronics, including capacitor plague, tin whiskers, and electromigration have been observed over timescales of a year or two, never mind a century. Environmentally controlled storage with the devices turned off will probably prolong their lifespan, but I wouldn't bet on turning on a modern laptop after even 30 years of storage. Additionally, a significant problem with trying to use a 100 year old DVD drive may be the flash memory parts that store the firmware (and the firmware in magnetic HDDs, the boot UEFI/BIOS in PCs, and all the storage in SSDs, tablets, smartphones, etc.). If the firmware flash loses or corrupts the firmware over your 100 year period, the DVD drive won't start, and the drive is effectively useless. This paper speculates that flash memory might last 100 years, but it's hard to know with any certainty (the manufacturers don't say - they probably don't know, and definitely don't care), and that paper is based on a 1995 era technology, so who knows how more modern (much smaller element-size) parts will perform. -- Finlay McWalterᚠTalk 23:06, 5 April 2015 (UTC)

Back to flash: encouragingly, this 2005 Freescale paper says they promise 10 to 20 years, but project >100. -- Finlay McWalterᚠTalk 23:16, 5 April 2015 (UTC)

In 100 years, people might be able to restore things like today, like replacing capcitors, but the plastic disk will not survive due harden out. The older it get, the slower it can go.[1][2] Lenses of players are made from plastc getting milky over time. The die is made for 20 years, CDRs and DVD±Rs will be not longer readable. --Hans Haase (有问题吗) 14:16, 6 April 2015 (UTC)

The relevant article here is Digital preservation. But definitely no one expects DVD technology to last 100 years, even if a concrete disk could last 100 years, some companies claim to produce an archival DVD that can last 1,000 years, but in real life you would always need to follow a preservation strategy. The linked article cites ways to store information long-term: refreshing, migrating, and replication. --Senteni (talk) 14:59, 6 April 2015 (UTC)