Slashdot videos: Now with more Slashdot!
The $28000/ drone apprehension figure did include "overhead".
I don't mean to argue that the drone program isn't inexpensive or inefficient - just that without comparing it with other methods (such as active patrols, etc.), the $28000/apprehension figure generates outrage only because $28000 seems like a VERY BIG NUMBER. And if you're going to get outraged by a VERY BIG NUMBER solely because of the metric of VERY BIG NUMBER/apprehension, it might be helpful to have a reference number of sorts rather than the costs people are used to from their daily lives.
The math for calculating this cost is deceptively simple-minded - and the article doesn't offer any way to compare it with other methods.
A (very) brief search for the US Border Patrol budget and apprehensions found these:
or FY2014 budget of $13.6 Billion and 486,651 apprehensions.
That gives an average cost of $27946/apprehension for the entire organization. My (very, very) simple minded calculation is remarkably similar to the Office of Inspector General's figures for just the drone program. If anything, it shows that just introducing drones doesn't change the cost-per-apprehension of the Border Patrol. A more important question would be whether cost-per-apprehension is even a valid metric for the Border Patrol.
3 years ago, my son went off to college with a Republic Wireless WiFi phone and subscription (early adopter.) His mother got a Republic Wireless subscription too. The lack of a multi-year contract and the low price were the appealing features to them, and there wasn't much in the way of competition back then for something that included a data plan.
My observations are as follows:
- Young people don't use a phone, they use a portable communications and entertainment device. Voice calls are one of multiple communications methods, and not a very important one. Therefore, voice quality doesn't matter much.
- WiFi-phones don't work well in environments where the WiFi is locked down tightly (such as hospitals) and BYO Devices aren't supported.
- My ex doesn't like accepting my calls.
For now, I'm still old school enough to want a phone that'll work in an emergency, in some rural environments where the minor carriers haven't built out yet. But that might change in a few years as I find myself hardly traveling anymore. I don't use a phone as a multi-media communications and entertainment device, but that might change when I replace my old iPhone-4.
A WiFi based phone service with cell-network backup would work for me 99% of the time, but I'm old enough and conservative enough to want a few more '9's.
Another thing I believe is that if WiFi based phone service expands and becomes common for large venues (such as sports stadiums, hotels, etc.), it could result in cell-network based phone service not growing in depth. The combination would reduce the service level expectations of voice service, displacing to more communications via. instant messaging/social networking.
Burn the witch!
She turned me into a newt!
Another convert to FreeBSD?
Let's ignore (for now) how (US) laws make the major automobile manufacturers dependent on their dealer networks to sell cars (etc.) - and the dealers are dependent on their service bays to stay in business. Consider only the operational aspects of how software updates are applied to cars - which is a VERY manual process with technicians and experts trained in ways to communicate with each other, and with (typically) a several day window in which the update(s) can be applied while the owner finds alternate transportation.
The existing process is (relatively) forgiving, since a technician has documentation, experience, and additional technical support to call for help. The customer is already inconvenienced, so adding a few hours (or even days) to the update process while problems are worked out is (barely) tolerable. Moreover, two cars of the same model (and trim) but manufactured a few weeks or months apart may have different controllers - something that the technician could verify, but the owner might not.
I suspect that software updates for most major automobile manufacturers is more like the state of firmware, driver, and OS updates was for Windows back in the 1990s.
Changing this will take time.
Heck, just create your own Board to issue Certificates of Authenticity - just like Rand Paul tried to do for ophthalmology (see http://www.washingtonpost.com/... )
Also, "Herbal" remedies can be marketed and sold without FDA involvement. I'm not sure whether the alternative medicine culture regrets or revels in the lack of scientific testing of most remedies, but it sure creates a wide open market where shysters can sell snake oil. And if it's homeopathic snake-oil, they don't need to squeeze the snake.
There are also certain herbal remedies where the evidence is inconclusive - but where it has been a popular folk-remedy for a long time (i.e.: echinacea
The article is about scheduling access to thread prioritization lists implemented as linked lists, not scheduling individual instructions. The issue involves multiple CPUs contending for the root (start) of a linked list, which causes cache-invalidation in the CPUs, thereby causing systemic inefficiencies.
A very dumbed-down description of their solution is to have multiple "root" lists in order to reduce the frequency of cache-contention/invalidation.
Thanks. That makes it more steerable (provided that the edges of the disk remain in a plane too.)
A disk 1/5 mile with a sensor 10 to 100 miles away (precisely aligned on the axis of the disk) isn't going to be very steerable, especially if the distances from the EDGES of the disk to the sensor all have to match within a half-wavelength in order for the interferometry to work right.
And wouldn't the changing relative positions of earth, moon, and sun cause disturbances in the disk? Is the solar wind sufficiently uniform over distances of 1/2 mile at earth orbit to not be a concern for causing non-uniform disturbances to the disk?
"geostationary" MUST be a mistake in the article. I don't see how the sensor can maintain a 1/2 wavelength position from the disk at a range of 10 to 100 miles unless the sensor is powered (ion drive?) somehow.