Forgot your password?
typodupeerror

Comment: Keeping it cheap and low-energy (Score 1) 233

by hankwang (#47943343) Attached to: Slashdot Asks: What's In Your Home Datacenter?

I was using an old netbook (Atom N270) as a home media server, keeping it running Ubuntu Desktop 9.something. But it was too much of a pain to maintain: keeping it on mains power for a year seems to break the battery-charge-level monitoring, which makes the internal battery useless as a UPS. Too many processes insisted on writing to files every 5 minutes, which was spinning up the hard disk all the time. Also, it got uncomfortably hot with the lid closed all the time.

So I got a second-hand thin client (Via 1 GHz CPU, 1 GB internal flash drive, 1 GB RAM, gigabit ethernet) for 75 euros, installed Ubuntu Server 12.04 and a USB hard disk that has a auto spin-down feature. The 1 GiB "SSD" turned out to be too small for OS and log files, so I augmented it with a 4 GB USB thumb drive, while cursing that it is very hard to find one that performs well on lots of small writes (With 4 kB random block write tests, the throughput of most sticks is less than 0.01 MB/s). This thing takes about 13 W of power while idle (26 euros/year at our rates) and last time I looked, there are no alternatives on the market that have comparable horsepower for much less watts. The server hosts media files (MiniDLNA) and backups (snapshots) of the various computers/tablets/phones. Once or twice a year, I sync the disk with an external USB drive. If my house goes up in flames, all will be gone.

Since my media storage is mostly audio, a 1 TB drive is plenty. When I read here from people that have 40 TB of storage in NAS arrays, I wonder what they are hoarding. At 5 GB per hour of video, that's 8000 hours!

Ubuntu 12.04 server sucks for a headless server. I think it was waiting for a keypress after an unclean boot. (There was a comment on a forum of a guy who had to drive 100 miles to a datacenter to attach a keyboard and press Enter). That one was solved, but now it just hangs during filesystem checking at boot time - some bad boot-order dependency.

Network: ethernet in living room (A/V center) and work room (desktop/printer); wifi elsewhere. Visitors get the wifi password. Non-media file transfers and backups are always over ssh. (It t1urns out that the gigabit ethernet was overkill; the VIA CPU can barely saturate a 100 Mbps ethernet line, with the faster SSH cipher (arcfour).

Comment: Re:complete sensationalist bullshit (Score 1) 270

by hankwang (#47942969) Attached to: Study Finds Link Between Artificial Sweeteners and Glucose Intolerance

"your body isn't absorbing every calorie you put into it. Your body absorbs until it has what it needs"

Consider what happens if a lactose-intolerant person drinks a glass or two of milk: about 25 g of carbohydrates that their body can't absorb. It will lead to flatulence and diarrhea as a result of gut bacteria feasting on those unused calories and the inability of the body to extract water effectively from a sugar solution.

The fact that this is an abnormal response shows that the normal thing is to absorb every calorie.

Comment: Re:Why do people use internal TLDs? (Score 1) 101

by hankwang (#47693031) Attached to: ICANN Offers Fix For Domain Name Collisions

" I always just use split horizon DNS, and put everything under the corporate domain name, thus eliminating the problem."

I have something like that at home, a registered domain name example.com and a portion *.home.example.com that was only resolvable from my lan.

Then, a few months back, I upgraded to the new Linux Mint LTS, which did all queries simultaneously to my ISP (fallback DNS) and my LAN DNS, using the first response. Sometimes the ISP was faster, resulting in 'nonexistent host' errors.

It took me an hour to figure out what was wrong and how to repair it (networkmanager.conf, disable dnsmasq). Sigh. I wasn't the first to have this problem. The devs didn't really see the problem. https://bugs.launchpad.net/ubu...

Comment: Re:Meh (Score 4, Funny) 90

by hankwang (#47537411) Attached to: How Stanford Engineers Created a Fictitious Compression For HBO

"you could run it repeatedly on a data source until you were down to a single bit."

That's why you need two distinct compression algorithms. Sometimes one will work better, sometimes the other. While repeatedly compressing, don't forget to write down in which sequence you need to apply the decompression. I believe this can compress abitrary data down to zero bits, if you are patient enough.

Comment: Re:OR (Score 1) 579

by hankwang (#47369417) Attached to: Unintended Consequences For Traffic Safety Feature

"Advanced driving courses teach vehicle dynamics, skid control, proper reactionary techniques to road hazards, proactive hazard evaluation, and so on; they cost $300 here, and you can go all the way to $1500 for driving/racing combined classes"

That's cheap. Here in Netherlands, a regular driving license will cost you around 30 hours(*) of instruction, plus 10 or so hours to study the traffic rules in all kinds of edge cases, and about 1500 euros for instruction, theory exam, and driving exam. It doesn't include skid control.

Traffic fatalities (per capita) are a factor 3 lower in Netherlands and Germany, compared to the US.

(*) it took me more like 75 hours of instruction and considerably more money... started at later age and generally bad body coordination/multitasking....

Comment: Re:How much reduced sleep is tied to long commutes (Score 1) 710

by hankwang (#47313049) Attached to: Workaholism In America Is Hurting the Economy

"I know people who are losing two hours of their life a day commuting each way, "

I commute well over 2 hours, 4 days per week. I don't see it as lost time. I'm reading slashdot and other sites in the train like now (plenty of space since I travel after the peak hours). In addition, 15 km of cycling per day, which is my only exercise. Fortunately the climate over here allows cycling.

But the idea of driving a car for 2 h/day horrifies me...

Comment: Re:Oblig Prior Art Question (Score 1) 56

by hankwang (#47082943) Attached to: Questionable Patents From MakerBot

When a patent is filed, I believe the USPTO keeps it confidential for a long time (a year?) until it is well along in process, to avoid revealing its secrets long before the patent is decided.

In the standard procedure, the application is kept secret for 18 months; then the application is published; the USPTO will then wait another year or so (depending on the back log it could be much more) before deciding whether or not to grant the patent. In this time slot between publication and decision, competitors could point out relevant prior art to USPTO, which would affect the decision.

In the US system, one can also file a provisional patent application and wait 12 months before filing the final application, which will essentially stretch the confidential period from 18 to 30 months. This was the case here. The final application can differ from the provisional application (errors corrected, more examples provided, reworded claims, etc.). In case of relevant prior art that was published between the provisional and final application, the provisional application will count.

Comment: Re:Has this ever happened to you? (Score 1) 216

by hankwang (#47000833) Attached to: Who controls the HVAC at work?

"a big portion of what you experience with regards to a temperatures bearability is based on humidity. In a very dry climate, 40C won't feel bad. In a very humid climate, you can start sweating at 16C."

In addition to that, the temperature is only sampled at one point. There can be quite a difference between temperature of supply and exhaust. I think a normal office should be ventilated at 100 m3/h

A human, with computer and lighting (250 W) can generate 8 C temperature increase. If you're unlucky with the airflow patterns, you could have people sitting in zones on both ends of that range. Moreover, between summer and winter, the walls may have different temperatures, which will affect both the local air temperatures and the energy balance of thermal radiation between warm bodies and walls.

Comment: Re:BMI is a lie! (Score 1) 329

by hankwang (#46966543) Attached to: Gaining On the US: Most Europeans To Be Overweight By 2030

Bah, 99.9% of the people who complain that their BMI is high because of muscles don't have that much muscles. This is Olaf Tufte, former olympic champion

Well, that's an example of a guy who, as you say, are almost pure muscle. Go and Google for "strongest man competition". Most of those guys have quite a bit more fat, but I doubt that their overal body fat percentage is that high.

Comment: Re: Our patent system is totally broken (Score 1) 152

by hankwang (#46956921) Attached to: USPTO Approves Amazon Patent For Taking Pictures

"he's hard pressed for time. The patent office is over worked, understaffed, and runs on quotas"

I'm actually amazed about what these examiners can achieve. Depending on what you assume for the hourly rate of an examiner, including all organizational overhead, they have 4 to 8 hours to read and understand the application, search prior art, and write their response.

I sometimes have to proofread draft patent applications of my own inventions, and it takes me typically 4 hours to review those (check that what the attorney wrote is a correct description of what I think the invention is). They turn my 2-page description of an idea into 25 pages of dense legalese, but at least I believe that I should have some advantage in understanding the idea, compared to the examiner.

At the source of every error which is blamed on the computer you will find at least two human errors, including the error of blaming it on the computer.

Working...