As a public health issue, even a 1% increase in brain cancers might be significant because they are expensive and hard to treat.
As a public health issue, I'd suggest driving while talking on a cell phone poses a much more substantial risk to the user, as well as to people around him or her, that any effects of ionizing radiation.
Did you actually read the article? That quote isn't from McCullagh:
Jim Harper, a policy analyst at the free-market Cato Institute, says that he welcomes the idea of a police-only Web interface as long as it's designed carefully. "A system like this should have strong logins, should require that the request be documented fully, and should produce statistical information so there can be strong oversight," he says. "I think that's a good thing to have."
I guess these days the Cato Institute favors freedom in the marketplace, but not quite so much freedom when it comes to mere citizens. Of course, the Cato Institute is full of "policy analysts" with deep understanding of the technologies and difficulties involved in determining the authenticity of a request and the potential for information leakage.
How, precisely, do we determine if a request is "fully documented" when it's coming from someone claiming to be a police officer using a web browser? Wouldn't full documentation require at a minimum that the search warrant somehow be uploaded to the site and reviewed for legitimacy before the information is released?
This follows from my finding that, on the evidence and on a proper interpretation of the law, a person makes each film available online only once through the BitTorrent system and electronically transmits each film only once through that system.
This is a very strange argument. If I torrent a movie and let it seed indefinitely, I will almost certainly have distributed more than one copy of the film. Did the justice really believe that torrenting is a one-for-one kind of activity where a downloaded work is uploaded once and only once? I haven't read the decision, but I wonder how much of it concerned downloading versus uploading.
These comments don't really alter the basic thrust of his decision, but they do give one pause to wonder how much the justice really understood about the mechanisms of the "BitTorrent system."
The problem I have with cap and trade is not that the method being used is inefficient, but that the value of the carbon credits is being set based on political motives.
Isn't the solution to valuation of the credits to allocate them by auction? That's where the EU method failed in my mind. You can't just distribute the credits to existing polluters initially; you need to auction them off. Market mechanisms in situations like these probably come pretty close to the theoretical model, particularly when it comes to "full information." I'd presume a utility has enough knowledge to calculate what they should pay to emit a ton of carbon dioxide. At some price it becomes more economical to invest in mitigation technologies than to buy the permits.
If you're a network executive trying to evaluate the long-term profitability for a given show, then these Nielsen ratings are, at best, an incomplete picture. You have to look at possible syndication deals, DVD sales, iTunes sales, Netflix licensing, Hulu views, merchandising, and probably some other stuff.
My understanding is that all this stuff constitutes "gravy." Producers look to cover their costs, or perhaps make a small profit, during the original network run. I'd bet that the other sources of revenue you list are pretty highly correlated with ratings during the initial broadcast run. Shows that don't succeed in building an audience on network television aren't very likely to sell lots of DVDs or do well in syndication. There are, of course, exceptions to this, but I'd bet the relationship is pretty strong for most normal types of television programming.
I wondered about this myself, but I think it ultimately comes down to trust. Why should advertisers trust Hulu to provide accurate numbers on viewing when they have a clear incentive to bump up their figures? The value of Nielsen is its impartiality.
Now if there were a way for a third-party to manage the logging of web visits, that might pose an alternative to Nielsen, if the advertisers and agencies would take the service provider seriously. So far, web measurement seems to rely on recruited panels like Alexa uses. Measuring web traffic at the servers themselves doesn't seem to be very common as of yet. Of course the web is incredibly decentralized which makes server-based measurements very difficult to manage. Still, a large number of consumers probably visit a fairly small number of sites (a few hundred sites probably constitute a large fraction of all visits in the US), so installing a server-based measurement technique at these sites could be quite effective.
There's a lot more to Nielsen than television ratings, too. They also measure sales data and correlate exposures with product sales. From the linked article:
Nielsen measures product sales, market share, distribution, price and merchandising conditions in tens of thousands of retail outlets such as grocery stores, drug stores, mass merchandisers and convenience stores. Reporting periods can be as short as a single day for selected electronic point-of-sale (POS) information or up to bimonthly for manual field audits. Data-collection methods will vary by country and type of outlet being reported.
Nielsen also measures the purchasing behavior of more than 250,000 households in 27 countries through our industry-leading consumer panels. Our US panel is the largest and most representative static sample in the country.
And you'd be wrong. They measure DVR usage the same way they measure every other medium connected to a TV including video games. They only count playback, of course, not recording, since advertisers only care about viewers.
Evidence continues to surface about American and other Western firms cooperating with repressive governments in their efforts to censor and eavesdrop on their citizens. Why didn't Mrs. Clinton mention them in her speech?
No, only IE 5.01 SP4 and IE 8 are not vulnerable without enabling "data execution prevention." The attackers apparently targeted IE 6, but nearly all other versions can be compromised.
From TFA:
"A security feature known as data execution prevention, which prevents data loaded into memory from being executed, will block the particular exploits McAfee has observed. But Kurtz warned the vulnerability exists in all versions of IE except for IE 5.01, service pack 4, and that it would be possible for attackers to work around the protection.
In an advisory, Microsoft recommended people use DEP, which by default is enabled in IE 8 but must be turned on in prior versions. The statement also advised users on Vista and later versions of Windows to run IE in protected mode. The advisory didn't say when an update would be released that patches the vulnerability."
Here's another vote in favor of Sphinx. I recently was presented with an online shopping site whose search functions were pathetically slow and inaccurate. I replaced these with Sphinx and now get incredibly fast results which are nearly always on target. You'll want to play with the weights assigned to fields and other features to optimize the searches, but if your content is already stored in a MySQL or PostgreSQL database, Sphinx should be one of your top contenders.
As the parent says, the indexing isn't real-time, but Sphinx has features to enable you to keep live indexes active while you reindex. The frequency of re-indexing will obviously depend on how important recency is for your users.
If your content is just text files, I'd consider htdig as well. While it's no longer being actively maintained, I've used it for years to index web archives of listserver postings with great success.
I don't know what parallel universe most of the commenters are coming from -- whether most of them are childless or just get their version of reality from FOX News, I don't know -- but the environment in which my teenager finds herself is highly competitive, not remotely cocooning or coddling, and in many ways significantly more stressful than the one I grew up in. And I don't have her on any medication.
I gave up my mod points to make another comment in this thread, but I have to endorse the parent comment as spot on. My daughter, a senior in high school, works harder than I ever did and carries a much larger load of homework than I ever did, despite the fact that I graduated from what was then a competitive private high school and attended an Ivy League college. We're just wrapping up the college admissions grind, and it's far more competitive and stressful out there than it was when I graduated in 1967. In her suburban Boston public high school, kids are told they must apply to at least eight colleges in varying categories of selectivity. Applying even to five colleges was pretty rare in my day; a minimum of eight would have seemed overkill.
Thank you for injecting a bit of scientific realism into a discussion that has so far consisted of opinion and anecdote.
Whenever I read studies like these my first question is whether the samples involved are comparable. What kinds of kids were taking the MMPI in 1938? What kinds of kids take them today? Might the differences in the samples be sufficient to account for the observed differences in scores? If there are differences, are they controlled for in the study design? I visited Dr. Twenge's website and didn't see any obvious links to the study in question so we can begin to evaluate its scientific validity. I did skim one study comparing student responses from the "Measuring the Future" surveys in 1975 and 2006, which claims to show that "there has been a small increase in positive self-views across the generations, but that this has not been accompanied by an increase in general self-competence." Yet I saw no sophisticated multivariate analysis that might tell us whether these were "real" changes in attitudes, what other variables were correlated with those attitudes, or how changes in those other variables might affect the observed change in attitudes.
For those interested in methodlogical problems in the social sciences, there's no better place to start than with Donald Campbell and Julian Stanley's Experimental and Quasi-Experimental Designs for Research. For a highly-readable introduction to what they call "threats to validity" in social scientific research, Campbell's "Reforms as Experiments" [pdf] is a good place to start.
I agree. This has to be one of the stupidest articles I've read lately.
I guess in the author's view if the SMTP envelope sender (the value appearing in the "Return-Path" header at the top of each delivered message) doesn't match the From: address, the message is somehow bogus. Try telling that to the thousands of listserver admins around the world. Many listservers preserve the the original message sender's address in the From field, while redistributing the message with an SMTP sender like owner-listname@example.com. That way if you hit reply, it goes back to the original author and not the list. However bounce messages get sent to the envelope sender, which is usually the listserver admin.
Automated web processes have the same feature. I'm careful to specify what I want the envelope sender to be and what I want the From to be, and often they are not the same thing at all. I wrote a variety of applications for organizations where an officer can send mail to a membership list using his or her own address as the From. However the envelope sender is usually something like bounces@example.com so that non-delivery messages go there rather than to the actual author.
I might want to compare the addresses, and maybe give non-matching ones an extra fractional point of spamminess in SpamAssassin, but that's about it. Not delivering messages like these would break an huge portion of the e-mail infrastructure.
IMHO, if Best Buy were charging people $40 to reinstall the OS and drivers from scratch, install all of the applications the customer wanted (and none that they didn't want), and do some basic post-install configuration (set up user accounts, click "OK" on all the EULAs, install all updates, etc.), then I would think that wasn't a bad value.
From the reading the article, it sounds more like they have two different images installed on the computers, one with crapware and one that is "optimized." Apparently what happens is that someone tries to buy a computer at the listed price but is told that's the price for the unoptimized machine which is out of stock. I don't think BB actually does anything to each specific computer to justify the $40, they just sell a machine with a different image.
Now there's a strategy I haven't heard before. Is it considered RFC-compliant to list MX servers you know for sure aren't authoritative for your domain? I guess in practice it only slows your own inbound connections, though it does add marginal overhead to the network as all those servers must issue the initial failing queries. Even with locally cached records there's still computational time involved in asking the question.
Some spammers routinely choose to mail to secondary servers, perhaps because they find those servers have less vigilant defenders athwart the walls. I'm going to try this strategy:
IN MX 0 mail.example.com.
IN MX 10 mail.example.com.
IN MX 20 mail.my_real_primary.net.
IN MX 30 mail.example.com.
IN MX 40 mail.my_real_secondary.net.
IN MX 50 mail.example.com.
The first two and the last protect protect against the most obvious strategies. The 30 record between the real servers may be overkill.
If you don't mind using a different player for those softsubbed MKVs, I strongly recommend SMPlayer, a well-designed front-end for mplayer. The Windows version includes an up-to-date mplayer build as well. I use the Linux version for this task, and it's superb including support for vdpau on my NVIDIA card. I do build my own mplayer binaries from the CVS snapshots, though, then tell smplayer to use the one I have in/usr/local/bin rather than the stock Ubuntu release. A quick "apt-get build-dep mplayer-nogui" will get all the development files you need to compile mplayer.
My recollection was that the introduction of Win95 led to a lot of upgrades from Win 3.1 and 3.11 installations. Just having an integrated TCP/IP stack and a web browser was a big deal. Back in 1994 I wrote an article that detailed how to install the Novell TCP/IP stack on machines running in Netware networks. It's hard to recall now, but Netware was networking in the Win 3 era. Before Win95, native support for networking in Windows really only existed in 3.11 ("Workgroups"), and that support largely consisted of SMB over Netbios which didn't help with Internet access.
Runs fine using Virtualbox on top of Ubuntu even in a VM with only 512 MB. I ran a variety of leaked release candidates in that configuration, though now that I have a machine with 4GB I'm giving one of them to the VM and running the actual release.
Fifteen minute boot times? Not here. More like one.
Trust me, it doesn't seem like a common practice in website design either. I don't think many web designers consider color-blindness at all.
Not to sound sexist, but I wonder if this has something to do with the higher proportion of women engaged in graphic design. Color-blindness is so rare in their gender that most women I know never think about it as a potential problem.
I have a mild red/green color blindness and find some games with color codings difficult to navigate. I had an especially hard time distinguishing the green and yellow elements in Chrono Cross where color is a primary component of the game play.
Upgrading to 9.10 killed my wireless as well. 8.10 and 9.04 worked fine with WPA-PSK2. In 9.10 it would hang and never connect to the router.
My biggest gripe about recent releases is the problems they have with certain displays. I have one machine connected to a Sony KDL-V3000 television. All the installation screens appear with unreadably tiny text; even zooming the image with the TV's scaler doesn't help. I've done enough Ubuntu installs to know pretty much every screen, but I wanted to repartition the hard drive on this occasion, and that was simply impossible given how small the text was. I even tried switching to the onboard VGA connector to see if the problem was a function of my NVIDIA card, but that didn't help either.
I can solve the problem after installation by adding the line Options "DPI" "100x100" to xorg.conf, but I obviously can't modify the file on the installation CD. I actually had to connect the machine to another monitor, install, then move the machine back to the TV and update the drivers. Of course, that wasn't too easy to do since I had no working wireless.
Studies conducted by advertising researchers generally find that the audio portion of a television commercial is much more important than the video portion when it comes to influencing what people remember. Another reason commercials are louder is the widespread realization among advertisers than many people leave the room during a commercial break. Not seeing the picture but still hearing the jingle can make an advertisement "effective" (in the advertising research sense of measures like "top-of-mind awareness" or "day-after-recall." Most studies I know of fail to demonstrate any direct connection between advertising and purchase behavior, but my knowledge of the field is now out-of-date. The one commonly-held belief based on experience is that cutting back on or stopping advertising for a product usually results in a decline in its market share. That fact puts advertisers in something of a "prisoner's dilemma.")
TV ratings are collected in two different ways. Some people, like you, fill in diaries to report their viewing, but Nielsen also maintains panels of homes with meters attached to all the video devices in the household. These meters report viewing pretty much on a minute-by-minute (or maybe these days second-by-second) basis. There's a national meter panel, and metered panels in the largest markets as well. National networks (both broadcast and cable) and national advertisers depend on these data from metered households. The diary method is used to measure viewing in local markets during "sweeps" periods (February, May, July, November). Smaller markets don't have the revenues to justify full-time metering and use the cheaper, and obviously somewhat more inaccurate, diary method instead.
Slashdot Mirror
I guess they should start with taking down whitehouse.gov now that the Administration has switched to Drupal.
(Who comes up with these names? "drupal" always makes me think of "droopy" which isn't really what I'd want my web site to be.)
As a public health issue, even a 1% increase in brain cancers might be significant because they are expensive and hard to treat.
As a public health issue, I'd suggest driving while talking on a cell phone poses a much more substantial risk to the user, as well as to people around him or her, that any effects of ionizing radiation.
Did you actually read the article? That quote isn't from McCullagh:
I guess these days the Cato Institute favors freedom in the marketplace, but not quite so much freedom when it comes to mere citizens. Of course, the Cato Institute is full of "policy analysts" with deep understanding of the technologies and difficulties involved in determining the authenticity of a request and the potential for information leakage.
How, precisely, do we determine if a request is "fully documented" when it's coming from someone claiming to be a police officer using a web browser? Wouldn't full documentation require at a minimum that the search warrant somehow be uploaded to the site and reviewed for legitimacy before the information is released?
This follows from my finding that, on the evidence and on a proper interpretation of the law, a person makes each film available online only once through the BitTorrent system and electronically transmits each film only once through that system.
This is a very strange argument. If I torrent a movie and let it seed indefinitely, I will almost certainly have distributed more than one copy of the film. Did the justice really believe that torrenting is a one-for-one kind of activity where a downloaded work is uploaded once and only once? I haven't read the decision, but I wonder how much of it concerned downloading versus uploading.
These comments don't really alter the basic thrust of his decision, but they do give one pause to wonder how much the justice really understood about the mechanisms of the "BitTorrent system."
The problem I have with cap and trade is not that the method being used is inefficient, but that the value of the carbon credits is being set based on political motives.
Isn't the solution to valuation of the credits to allocate them by auction? That's where the EU method failed in my mind. You can't just distribute the credits to existing polluters initially; you need to auction them off. Market mechanisms in situations like these probably come pretty close to the theoretical model, particularly when it comes to "full information." I'd presume a utility has enough knowledge to calculate what they should pay to emit a ton of carbon dioxide. At some price it becomes more economical to invest in mitigation technologies than to buy the permits.
I strongly suspect that the amount that I pay per show is more than the amount advertisers pays per viewer
In prime time, advertisers pay about three cents per HH. See http://www.tvb.org/rcentral/adrevenuetrack/media/media.asp?c=2f.
If you're a network executive trying to evaluate the long-term profitability for a given show, then these Nielsen ratings are, at best, an incomplete picture. You have to look at possible syndication deals, DVD sales, iTunes sales, Netflix licensing, Hulu views, merchandising, and probably some other stuff.
My understanding is that all this stuff constitutes "gravy." Producers look to cover their costs, or perhaps make a small profit, during the original network run. I'd bet that the other sources of revenue you list are pretty highly correlated with ratings during the initial broadcast run. Shows that don't succeed in building an audience on network television aren't very likely to sell lots of DVDs or do well in syndication. There are, of course, exceptions to this, but I'd bet the relationship is pretty strong for most normal types of television programming.
I wondered about this myself, but I think it ultimately comes down to trust. Why should advertisers trust Hulu to provide accurate numbers on viewing when they have a clear incentive to bump up their figures? The value of Nielsen is its impartiality.
Now if there were a way for a third-party to manage the logging of web visits, that might pose an alternative to Nielsen, if the advertisers and agencies would take the service provider seriously. So far, web measurement seems to rely on recruited panels like Alexa uses. Measuring web traffic at the servers themselves doesn't seem to be very common as of yet. Of course the web is incredibly decentralized which makes server-based measurements very difficult to manage. Still, a large number of consumers probably visit a fairly small number of sites (a few hundred sites probably constitute a large fraction of all visits in the US), so installing a server-based measurement technique at these sites could be quite effective.
There's a lot more to Nielsen than television ratings, too. They also measure sales data and correlate exposures with product sales. From the linked article:
And you'd be wrong. They measure DVR usage the same way they measure every other medium connected to a TV including video games. They only count playback, of course, not recording, since advertisers only care about viewers.
Evidence continues to surface about American and other Western firms cooperating with repressive governments in their efforts to censor and eavesdrop on their citizens. Why didn't Mrs. Clinton mention them in her speech?
We have, for instance, Cisco, Nokia/Siemens, Microsoft, and Yahoo, just to name a few.
According to TFA, this vulnerability was in IE6.
No, only IE 5.01 SP4 and IE 8 are not vulnerable without enabling "data execution prevention." The attackers apparently targeted IE 6, but nearly all other versions can be compromised.
From TFA:
"A security feature known as data execution prevention, which prevents data loaded into memory from being executed, will block the particular exploits McAfee has observed. But Kurtz warned the vulnerability exists in all versions of IE except for IE 5.01, service pack 4, and that it would be possible for attackers to work around the protection.
In an advisory, Microsoft recommended people use DEP, which by default is enabled in IE 8 but must be turned on in prior versions. The statement also advised users on Vista and later versions of Windows to run IE in protected mode. The advisory didn't say when an update would be released that patches the vulnerability."
Here's another vote in favor of Sphinx. I recently was presented with an online shopping site whose search functions were pathetically slow and inaccurate. I replaced these with Sphinx and now get incredibly fast results which are nearly always on target. You'll want to play with the weights assigned to fields and other features to optimize the searches, but if your content is already stored in a MySQL or PostgreSQL database, Sphinx should be one of your top contenders.
As the parent says, the indexing isn't real-time, but Sphinx has features to enable you to keep live indexes active while you reindex. The frequency of re-indexing will obviously depend on how important recency is for your users.
If your content is just text files, I'd consider htdig as well. While it's no longer being actively maintained, I've used it for years to index web archives of listserver postings with great success.
I don't know what parallel universe most of the commenters are coming from -- whether most of them are childless or just get their version of reality from FOX News, I don't know -- but the environment in which my teenager finds herself is highly competitive, not remotely cocooning or coddling, and in many ways significantly more stressful than the one I grew up in. And I don't have her on any medication.
I gave up my mod points to make another comment in this thread, but I have to endorse the parent comment as spot on. My daughter, a senior in high school, works harder than I ever did and carries a much larger load of homework than I ever did, despite the fact that I graduated from what was then a competitive private high school and attended an Ivy League college. We're just wrapping up the college admissions grind, and it's far more competitive and stressful out there than it was when I graduated in 1967. In her suburban Boston public high school, kids are told they must apply to at least eight colleges in varying categories of selectivity. Applying even to five colleges was pretty rare in my day; a minimum of eight would have seemed overkill.
Thank you for injecting a bit of scientific realism into a discussion that has so far consisted of opinion and anecdote.
Whenever I read studies like these my first question is whether the samples involved are comparable. What kinds of kids were taking the MMPI in 1938? What kinds of kids take them today? Might the differences in the samples be sufficient to account for the observed differences in scores? If there are differences, are they controlled for in the study design? I visited Dr. Twenge's website and didn't see any obvious links to the study in question so we can begin to evaluate its scientific validity. I did skim one study comparing student responses from the "Measuring the Future" surveys in 1975 and 2006, which claims to show that "there has been a small increase in positive self-views across the generations, but that this has not been accompanied by an increase in general self-competence." Yet I saw no sophisticated multivariate analysis that might tell us whether these were "real" changes in attitudes, what other variables were correlated with those attitudes, or how changes in those other variables might affect the observed change in attitudes.
For those interested in methodlogical problems in the social sciences, there's no better place to start than with Donald Campbell and Julian Stanley's Experimental and Quasi-Experimental Designs for Research . For a highly-readable introduction to what they call "threats to validity" in social scientific research, Campbell's "Reforms as Experiments" [pdf] is a good place to start.
I agree. This has to be one of the stupidest articles I've read lately.
I guess in the author's view if the SMTP envelope sender (the value appearing in the "Return-Path" header at the top of each delivered message) doesn't match the From: address, the message is somehow bogus. Try telling that to the thousands of listserver admins around the world. Many listservers preserve the the original message sender's address in the From field, while redistributing the message with an SMTP sender like owner-listname@example.com. That way if you hit reply, it goes back to the original author and not the list. However bounce messages get sent to the envelope sender, which is usually the listserver admin.
Automated web processes have the same feature. I'm careful to specify what I want the envelope sender to be and what I want the From to be, and often they are not the same thing at all. I wrote a variety of applications for organizations where an officer can send mail to a membership list using his or her own address as the From. However the envelope sender is usually something like bounces@example.com so that non-delivery messages go there rather than to the actual author.
I might want to compare the addresses, and maybe give non-matching ones an extra fractional point of spamminess in SpamAssassin, but that's about it. Not delivering messages like these would break an huge portion of the e-mail infrastructure.
IMHO, if Best Buy were charging people $40 to reinstall the OS and drivers from scratch, install all of the applications the customer wanted (and none that they didn't want), and do some basic post-install configuration (set up user accounts, click "OK" on all the EULAs, install all updates, etc.), then I would think that wasn't a bad value.
From the reading the article, it sounds more like they have two different images installed on the computers, one with crapware and one that is "optimized." Apparently what happens is that someone tries to buy a computer at the listed price but is told that's the price for the unoptimized machine which is out of stock. I don't think BB actually does anything to each specific computer to justify the $40, they just sell a machine with a different image.
Now there's a strategy I haven't heard before. Is it considered RFC-compliant to list MX servers you know for sure aren't authoritative for your domain? I guess in practice it only slows your own inbound connections, though it does add marginal overhead to the network as all those servers must issue the initial failing queries. Even with locally cached records there's still computational time involved in asking the question.
Some spammers routinely choose to mail to secondary servers, perhaps because they find those servers have less vigilant defenders athwart the walls. I'm going to try this strategy:
IN MX 0 mail.example.com.
IN MX 10 mail.example.com.
IN MX 20 mail.my_real_primary.net.
IN MX 30 mail.example.com.
IN MX 40 mail.my_real_secondary.net.
IN MX 50 mail.example.com.
The first two and the last protect protect against the most obvious strategies. The 30 record between the real servers may be overkill.
If you don't mind using a different player for those softsubbed MKVs, I strongly recommend SMPlayer, a well-designed front-end for mplayer. The Windows version includes an up-to-date mplayer build as well. I use the Linux version for this task, and it's superb including support for vdpau on my NVIDIA card. I do build my own mplayer binaries from the CVS snapshots, though, then tell smplayer to use the one I have in /usr/local/bin rather than the stock Ubuntu release. A quick "apt-get build-dep mplayer-nogui" will get all the development files you need to compile mplayer.
My recollection was that the introduction of Win95 led to a lot of upgrades from Win 3.1 and 3.11 installations. Just having an integrated TCP/IP stack and a web browser was a big deal. Back in 1994 I wrote an article that detailed how to install the Novell TCP/IP stack on machines running in Netware networks. It's hard to recall now, but Netware was networking in the Win 3 era. Before Win95, native support for networking in Windows really only existed in 3.11 ("Workgroups"), and that support largely consisted of SMB over Netbios which didn't help with Internet access.
Runs fine using Virtualbox on top of Ubuntu even in a VM with only 512 MB. I ran a variety of leaked release candidates in that configuration, though now that I have a machine with 4GB I'm giving one of them to the VM and running the actual release.
Fifteen minute boot times? Not here. More like one.
Trust me, it doesn't seem like a common practice in website design either. I don't think many web designers consider color-blindness at all.
Not to sound sexist, but I wonder if this has something to do with the higher proportion of women engaged in graphic design. Color-blindness is so rare in their gender that most women I know never think about it as a potential problem.
I have a mild red/green color blindness and find some games with color codings difficult to navigate. I had an especially hard time distinguishing the green and yellow elements in Chrono Cross where color is a primary component of the game play.
Upgrading to 9.10 killed my wireless as well. 8.10 and 9.04 worked fine with WPA-PSK2. In 9.10 it would hang and never connect to the router.
My biggest gripe about recent releases is the problems they have with certain displays. I have one machine connected to a Sony KDL-V3000 television. All the installation screens appear with unreadably tiny text; even zooming the image with the TV's scaler doesn't help. I've done enough Ubuntu installs to know pretty much every screen, but I wanted to repartition the hard drive on this occasion, and that was simply impossible given how small the text was. I even tried switching to the onboard VGA connector to see if the problem was a function of my NVIDIA card, but that didn't help either.
I can solve the problem after installation by adding the line
Options "DPI" "100x100"
to xorg.conf, but I obviously can't modify the file on the installation CD. I actually had to connect the machine to another monitor, install, then move the machine back to the TV and update the drivers. Of course, that wasn't too easy to do since I had no working wireless.
I reverted back to 9.04.
Studies conducted by advertising researchers generally find that the audio portion of a television commercial is much more important than the video portion when it comes to influencing what people remember. Another reason commercials are louder is the widespread realization among advertisers than many people leave the room during a commercial break. Not seeing the picture but still hearing the jingle can make an advertisement "effective" (in the advertising research sense of measures like "top-of-mind awareness" or "day-after-recall." Most studies I know of fail to demonstrate any direct connection between advertising and purchase behavior, but my knowledge of the field is now out-of-date. The one commonly-held belief based on experience is that cutting back on or stopping advertising for a product usually results in a decline in its market share. That fact puts advertisers in something of a "prisoner's dilemma.")
TV ratings are collected in two different ways. Some people, like you, fill in diaries to report their viewing, but Nielsen also maintains panels of homes with meters attached to all the video devices in the household. These meters report viewing pretty much on a minute-by-minute (or maybe these days second-by-second) basis. There's a national meter panel, and metered panels in the largest markets as well. National networks (both broadcast and cable) and national advertisers depend on these data from metered households. The diary method is used to measure viewing in local markets during "sweeps" periods (February, May, July, November). Smaller markets don't have the revenues to justify full-time metering and use the cheaper, and obviously somewhat more inaccurate, diary method instead.