That one, it turns out, is even less of an island, and mostly not even visible: the "great Pacific garbage patch" is not really a macroscopic phenomenon, but rather an area of the Pacific Ocean with higher than normal concentrations of plastic particles, mostly suspended beneath the surface. The larger pieces are broken down by wave action fairly quickly, so it's not a giant mass of floating milk jugs or anything like that.
If it's like previous pumice rafts, it's more like a large area of debris than a big island. Here's a random photo showing a boat plowing a path through one made up of smaller pieces. Not really the kind of thing you can walk around on, though the description of this one having an edge like an ice-shelf makes it sound like it may have larger rocks in it. Here is a NASA satellite photo of a 2006 occurrence with a more obvious origin (it's adjacent to an erupting volcano).
This rover actually does have color cameras on it. The primary navigation cameras are B&W, but there are two color 2-mpixel cameras used for sending back photos. NASA has some information on that. Not sure why the linked image here is in B&W. Perhaps they sometimes take B&W photos to save bandwidth? The MarsEarth link is 29 kbit, basically '90s modem speed.
I would say most small businesses I know actually don't use any specific financial software, but do everything in a spreadsheet package. Excel rules the small-business world in a lot more ways than you might expect. You can probably do most similar things in LibreOffice. Now whether this is a good idea varies. The con is that you can end up with a sprawling spreadsheet-and-macros mess, but the pro is some flexibility in doing complex things, and simplicity in doing easy things.
GnuCash is not a bad option either, but it works best if your processes map on cleanly to one of its default processes. It does standard double-entry bookkeeping just fine. Its documentation is pretty good, also. But if you want to be doing significant scripting or customized report-generation, I find spreadsheets easier than dealing with GnuCash scripting+reports.
Depends on what kind of business to some extent. For example, if you need to interface with shopping-cart software or something of that sort, you may have more specific requirements.
If it weren't for their disqualifying views on abortion, I think Romney might well have gone for someone like Lingle. She's popular, and would probably put a certain number of independents in play (even if not Hawaii itself). I don't see what Ryan adds really. Romney has more credibility on the economy than Paul "Medicare Part D" Ryan does.
I think part of the reason she seems more harmless and amusing now is that the likelihood of her ever occupying the Presidency has declined. Palin as political pundit vs. Palin as a person one McCain health problem away from the Oval Office are pretty different scenarios.
There was speculation that Romney would pick a woman as running-mate, but the problem is that too many high-profile Republican women are pro-choice, which is a disqualifying factor with the GOP base. That removes five of the best possible options of highly qualified, popular Republican women: Susan Collins or Olympia Snow (Maine Senators), Linda Lingle (former Hawaii Governor), Lisa Murkowski (Alaska Senator), or Jodi Rell (former Connecticut governor).
What's left? Well there are some who are way too far on the other side to be appealing to independents. Romney could've energized his base by picking Michelle Bachmann, for example, but it would probably not win him the election. Same with Jan Brewer (Arizona governor).
Once you remove the two categories of "too moderate" and "too firebrand", the remaining GOP women tend not to have much name recognition. Who else would he pick?
I don't see a real tech angle on this, and it's not even particularly surprising or that interesting from a political angle. GOP candidate makes reasonably "safe" GOP choice, perhaps leaning right, but given his current party, it's not like there was much elsewhere he could've leant. I was vaguely intrigued by a suggestion that Romney could try a gambit of picking a pro-choice moderate to win over independents, at the risk of really pissing off the Republican base. But I don't think anyone seriously expected that to happen.
Overall I think the pick is a NOOP; few people who previously disliked Romney are now going to be won over, and vice versa.
I wouldn't be surprised if they already employ some version of that, at least in some contexts. They include all sorts of non-public "content quality" signals nowadays; it's no longer primarily based on the link graph like it was in the PageRank days.
I wonder if it's actually possible to commit suicide by swallowing placebos? Or is there some limit to the nocebo effect's severity that'd prevent that?
Responding to self after tracking down a copy of the paper:
One interesting thing they suggest is that, since in this study the "truthiness" effect happened in both directions, or even with unrelated images, previous studies showing that images produce a bias might need to be re-run with control images that are unrelated, i.e. placebo images.
For example, the paper mentions a 2008 paper that found public trust in neuroscience findings was higher if accompanied by an image of a brain scan. That article speculated that "part of the fascination, and the credibility, of brain imaging research lies in the persuasive power of the actual brain images themselves". But the authors of this paper point out that perhaps it was just the presence of any image at all: what would happen if you re-represented the same articles, not with brain scans, but with just photos of the neuroscientists, or of the MRI machine? The authors hypothesize that you might get more people believing in the results in those cases, too, in which case it wouldn't actually be that the brain-scan images are serving any persuasive or evidentiary role in and of themselves.
I don't think it requires assuming the poll was biased or that "internet sites" are posting un-vetted charges. A simpler explanation is that, even if the TSA does suck, most Americans either don't know or don't care. In particular, a significant percentage of Americans don't fly regularly, and they tend to support whatever air-security measures some official claims are necessary. To them, something that sounds like security is good, and who cares if someone's inconvenienced, because it's not them anyway. For example, a 2010 poll found that x-ray scanners and new pat-down procedures were more popular among non-fliers:
Among Americans who fly at least once a year, 58 percent support the new x-ray scanners, versus 70 percent of Americans who fly less often than that. Support for the new pat-down procedures is at 44 percent among fliers, meanwhile, versus 52 percent among those who do not fly regularly.
I suppose this one is more miniaturized, but "virtual reality" and "haptics" people have been trying this sort of thing for a while, without having yet really come up with a compelling win. Here is a 1995 Popular Science article about a device that takes basically the same approach to characterizing what constitutes texture perceived by fingertips (by simulating sliding/bumpiness/resistance/stickiness). Available for only $20,000! And a 1996 textbook devoted large sections to the topic as well.
I dunno, I tend to find Craigslist more appealing than most of its replacements. I suspect a lot of other Craigslist users have similar views: I just want one simple site for my area, not a bunch of fancy zoomable web 2.0 maps like Yelp or something.
The court decision that forced this was actually written by a conservative Reagan appointee. The 3-judge panel overall had 2 Republican and 1 Democratic appointees.
That was actually a fairly interesting interview. I wouldn't have guessed Rob would be into frictionless sharing (albeit opt-in). "You post to Slashdot but you would never Tweet".
Although I wonder why "Slashdot TV" didn't videotape the interview? It's more like a radio interview with a gratuitous slideshow on it.
It does say he ordered the TV from a third-party merchant, via Amazon's marketplace. It's possible the third-party merchant sells both TVs and firearms, and mis-shipped his order.
I believe they're testing each font's performance against the average performance, rather than looking for pairwise differences between a specific pair of fonts.
They don't seem to represent the sampling uncertainty graphically as error bars, but if you scroll down to the paragraph that starts with "Are the results the product of chance?", they do a basic statistical analysis, and find that Baskerville performs better than average with p < 0.01 (and still p < 0.05, if you do a Bonferroni correction).
I believe that's the case, yes, but it might be harder than that to fool it. It's likely that it ignores things that don't look like human faces when building its model, to weed out the "joke" tags, like when someone tags an inanimate object as one of their friends. So if you kept tagging a dog, it might discount those data points. You'd probably have to consistently mistag another human face as you, and ideally the same person (if you tag a bunch of different people as you it might discount those as spurious tags also).
Slashdot Mirror
That one, it turns out, is even less of an island, and mostly not even visible: the "great Pacific garbage patch" is not really a macroscopic phenomenon, but rather an area of the Pacific Ocean with higher than normal concentrations of plastic particles, mostly suspended beneath the surface. The larger pieces are broken down by wave action fairly quickly, so it's not a giant mass of floating milk jugs or anything like that.
If it's like previous pumice rafts, it's more like a large area of debris than a big island. Here's a random photo showing a boat plowing a path through one made up of smaller pieces. Not really the kind of thing you can walk around on, though the description of this one having an edge like an ice-shelf makes it sound like it may have larger rocks in it. Here is a NASA satellite photo of a 2006 occurrence with a more obvious origin (it's adjacent to an erupting volcano).
This rover actually does have color cameras on it. The primary navigation cameras are B&W, but there are two color 2-mpixel cameras used for sending back photos. NASA has some information on that. Not sure why the linked image here is in B&W. Perhaps they sometimes take B&W photos to save bandwidth? The MarsEarth link is 29 kbit, basically '90s modem speed.
I would say most small businesses I know actually don't use any specific financial software, but do everything in a spreadsheet package. Excel rules the small-business world in a lot more ways than you might expect. You can probably do most similar things in LibreOffice. Now whether this is a good idea varies. The con is that you can end up with a sprawling spreadsheet-and-macros mess, but the pro is some flexibility in doing complex things, and simplicity in doing easy things.
GnuCash is not a bad option either, but it works best if your processes map on cleanly to one of its default processes. It does standard double-entry bookkeeping just fine. Its documentation is pretty good, also. But if you want to be doing significant scripting or customized report-generation, I find spreadsheets easier than dealing with GnuCash scripting+reports.
Depends on what kind of business to some extent. For example, if you need to interface with shopping-cart software or something of that sort, you may have more specific requirements.
If it weren't for their disqualifying views on abortion, I think Romney might well have gone for someone like Lingle. She's popular, and would probably put a certain number of independents in play (even if not Hawaii itself). I don't see what Ryan adds really. Romney has more credibility on the economy than Paul "Medicare Part D" Ryan does.
Oh, and Ryan is Roman Catholic, not Jewish.
I think part of the reason she seems more harmless and amusing now is that the likelihood of her ever occupying the Presidency has declined. Palin as political pundit vs. Palin as a person one McCain health problem away from the Oval Office are pretty different scenarios.
Perhaps Romney is simultaneously announcing his choice of running mate and his plan to commit suicide if elected.
There was speculation that Romney would pick a woman as running-mate, but the problem is that too many high-profile Republican women are pro-choice, which is a disqualifying factor with the GOP base. That removes five of the best possible options of highly qualified, popular Republican women: Susan Collins or Olympia Snow (Maine Senators), Linda Lingle (former Hawaii Governor), Lisa Murkowski (Alaska Senator), or Jodi Rell (former Connecticut governor).
What's left? Well there are some who are way too far on the other side to be appealing to independents. Romney could've energized his base by picking Michelle Bachmann, for example, but it would probably not win him the election. Same with Jan Brewer (Arizona governor).
Once you remove the two categories of "too moderate" and "too firebrand", the remaining GOP women tend not to have much name recognition. Who else would he pick?
I don't see a real tech angle on this, and it's not even particularly surprising or that interesting from a political angle. GOP candidate makes reasonably "safe" GOP choice, perhaps leaning right, but given his current party, it's not like there was much elsewhere he could've leant. I was vaguely intrigued by a suggestion that Romney could try a gambit of picking a pro-choice moderate to win over independents, at the risk of really pissing off the Republican base. But I don't think anyone seriously expected that to happen.
Overall I think the pick is a NOOP; few people who previously disliked Romney are now going to be won over, and vice versa.
In a related move, they could stop turning up YouTube search results that won't play in my location...
I wouldn't be surprised if they already employ some version of that, at least in some contexts. They include all sorts of non-public "content quality" signals nowadays; it's no longer primarily based on the link graph like it was in the PageRank days.
I wonder if it's actually possible to commit suicide by swallowing placebos? Or is there some limit to the nocebo effect's severity that'd prevent that?
Responding to self after tracking down a copy of the paper:
One interesting thing they suggest is that, since in this study the "truthiness" effect happened in both directions, or even with unrelated images, previous studies showing that images produce a bias might need to be re-run with control images that are unrelated, i.e. placebo images.
For example, the paper mentions a 2008 paper that found public trust in neuroscience findings was higher if accompanied by an image of a brain scan. That article speculated that "part of the fascination, and the credibility, of brain imaging research lies in the persuasive power of the actual brain images themselves". But the authors of this paper point out that perhaps it was just the presence of any image at all: what would happen if you re-represented the same articles, not with brain scans, but with just photos of the neuroscientists, or of the MRI machine? The authors hypothesize that you might get more people believing in the results in those cases, too, in which case it wouldn't actually be that the brain-scan images are serving any persuasive or evidentiary role in and of themselves.
Filmmakers noticed early on that juxtaposing images had significant effects on perception, with the Kuleshov Effect being one famous demonstration.
I don't think it requires assuming the poll was biased or that "internet sites" are posting un-vetted charges. A simpler explanation is that, even if the TSA does suck, most Americans either don't know or don't care. In particular, a significant percentage of Americans don't fly regularly, and they tend to support whatever air-security measures some official claims are necessary. To them, something that sounds like security is good, and who cares if someone's inconvenienced, because it's not them anyway. For example, a 2010 poll found that x-ray scanners and new pat-down procedures were more popular among non-fliers:
Neoliberal software development?
I suppose this one is more miniaturized, but "virtual reality" and "haptics" people have been trying this sort of thing for a while, without having yet really come up with a compelling win. Here is a 1995 Popular Science article about a device that takes basically the same approach to characterizing what constitutes texture perceived by fingertips (by simulating sliding/bumpiness/resistance/stickiness). Available for only $20,000! And a 1996 textbook devoted large sections to the topic as well.
I dunno, I tend to find Craigslist more appealing than most of its replacements. I suspect a lot of other Craigslist users have similar views: I just want one simple site for my area, not a bunch of fancy zoomable web 2.0 maps like Yelp or something.
The court decision that forced this was actually written by a conservative Reagan appointee. The 3-judge panel overall had 2 Republican and 1 Democratic appointees.
That was actually a fairly interesting interview. I wouldn't have guessed Rob would be into frictionless sharing (albeit opt-in). "You post to Slashdot but you would never Tweet".
Although I wonder why "Slashdot TV" didn't videotape the interview? It's more like a radio interview with a gratuitous slideshow on it.
It does say he ordered the TV from a third-party merchant, via Amazon's marketplace. It's possible the third-party merchant sells both TVs and firearms, and mis-shipped his order.
SIG seems to market it as a "patrol rifle". I don't think they're positioning it as a hunting rifle, although I suppose you could hunt with it anyway.
I believe they're testing each font's performance against the average performance, rather than looking for pairwise differences between a specific pair of fonts.
They don't seem to represent the sampling uncertainty graphically as error bars, but if you scroll down to the paragraph that starts with "Are the results the product of chance?", they do a basic statistical analysis, and find that Baskerville performs better than average with p < 0.01 (and still p < 0.05, if you do a Bonferroni correction).
I believe that's the case, yes, but it might be harder than that to fool it. It's likely that it ignores things that don't look like human faces when building its model, to weed out the "joke" tags, like when someone tags an inanimate object as one of their friends. So if you kept tagging a dog, it might discount those data points. You'd probably have to consistently mistag another human face as you, and ideally the same person (if you tag a bunch of different people as you it might discount those as spurious tags also).