Another way in which I think the study may be flawed is by asking people to self-define their position in the political spectrum -- a one-dimensional political spectrum. What guarantees do you have that participants really are "conservative" or "liberal" (whatever that means to you), and have actually thought about the political issues involved in each "choice" (as if there weren't many, many greys)?. A 2-dimensional political positioning would provide more insight. A short questionnaire where participants actually had to think, instead of "choosing their favorite color" would have been even better.
But the quoted excerpt from the article noted that this single-item measure accounted for a whopping 85% of the variance in Presidential voting intentions in surveys across three decades, and in this study was correlatedstrongly with reported vote for Kerry or Bush. That strong evidence of the measure's validity should outweigh an intuitive sense that the measure "may be flawed," no?
If someone gains physical access to the machine, well... if someone gains physical access to their TV, they can just walk off with it, and you don't see people chaining TVs to their walls to avoid the scenario. So really, it's a nonissue for 99%+ of the userbase
Bad analogy. Even ignoring the fact that the computer costs more than the TV.
1. The TV theft costs me a few hundred dollars, worst case (if not covered by insurance). The computer theft may cost me all my family photos, years of archived emails, difficult-to-reconstruct work documents, financial records, and even potentially embarrassing materials in some unknown person's hands. A much greater downside risk means greater security is appropriate.
2. If someone steals the TV I know it. If someone dumps my data to a CD (or accesses my data remotely over the network) I may not even know it. See #1: data theft can be massively damaging to me, perhaps even more if I am unaware (so I can't cancel my credit cards etc.)
My longstanding beef (yes, I have complained to them) is that this doesn't work for currency units. Try "240 euros in dollars" -- it does not recognize this as a unit conversion.
1. Nested lists: so I could have one list that says "if genre = rock", then a sublist that just has "if My Rating is > 3" or "if year published is 2" and the other "if My Rating is > 3" (which I use to differentiate between "Background work music" and "Driving kick ass music".
Nested lists are truly a top wish-list item, but the poster doesn't even mention the most important potential use -- for people (like me) who have music libraries larger than their iPod capacity. You can set up a size-limited playlist in iTunes, using whatever criteria you want, and autosync that with the iPod. But suppose you also want a playlist like "Not played in last 6 months" or "Recently added" or "Rated over 3"? You can set those up as Smart Lists in iTunes but can't sync them with the iPod because they will include tunes that aren't in the iPod.
Nested lists are exactly what is needed, to create a list of "Not played in last 6 months" or whatever within the iPod size-limited and autosynced list. Please, Apple...
OK, there are a lot of misconceptions here about how scientific publishing works. Let's try to straighten some out.
1. What is the current system? Most journals don't charge authors to publish, some do, some charge only for material that is particularly costly to reproduce (color photos as opposed to text or B&W graphs). In no reputable journal does author payment determine whether or not a paper is published (only, sometimes, how speedily). The decision as to publication is made by an editor assisted by peer reviewers (experts in the area of the paper). Submitted papers can be processed without much concrete expense (some clerical/administrative costs, a few thousand to tens of thousands honorarium to journal editors, no payment to peer reviewers). Typesetting, printing, and mailing cost money too. As a rough approximation to those concrete costs, consider that most journals are offered to individual subscribers at rates of $30-a few hundred per year. The big bucks come from library or institutional subscriptions, which range from a few hundred to many tens of thousands a year.
2. Why does this insane system survive, where scientific publishers basically have a license to print money at the expense of the libraries? Branding. The top journals have hard-earned reputations, due to consistently publishing high-quality, high-impact papers over the years. Authors want to publish in those journals, to advance their own reputations, and to reach the widest range of readers. Others in the field want to subscribe to those journals or have their libraries do so. Without this factor, internet distribution could make the whole peer-review, editing, and distribution process much cheaper (although not zero-cost), bypassing the scientific publishers.
3. Why should authors pay? Because they are in the best position to cover those remaining unavoidable expenses. Consider that scientific research can't be done for free; it is currently funded, often by government or industry grants, sometimes by universities themselves. They could simply add the costs of publication (which might be 5% of the cost of conducting the research in the first place) to their funding for research. Subsidies would have to be provided for papers from poor countries or from independent (non-university-affiliated) researchers. If author payments mean that the product of research is generally available on the internet, the resulting savings to universities in library subscription costs would be considerably greater than the extra costs of subsidizing publication.
4. Who would suffer under an author-pays, free-subscription system? The scientific publishers (commercial enterprises), and many scientific societies, who often own the journals that are published by commercial publishers and share in their profits. These are real issues. But on balance, the broader distribution of authors' ideas and the faster advancement of science due to more open communication is probably a more important consideration.
These comments don't address more radical changes in the system, e.g., going from the current peer-review arrangements to a Slashdot-style moderation plan. But I hope this post will help orient readers to some of the facts and issues.
In the old days, this same point was made about stereo gear (receivers, pre-amps, what have you). One time a columnist wrote in Stereo Review or whatever mag I was getting at the time to defend the magazine's practice of reviewing models sent by the manufacturers. The argument was that it would be just silly for a manufacturer to send a "specially tuned" unit to the magazine for testing and review because the special manual tuning would as a rule make performance worse than a random production machine! This claim always sounded fishy to me -- a slur on the competence of the engineer who was told to do the tuning?
In any case, with computer prices as low as they are today why on earth wouldn't a publication dispose of this worry and buy a unit for review off the shelf? With staff members' salaries and all, this would certainly not be the majority of the cost of generating a full review article. And, yes, with some components like LCDs, it would be worth buying three to look at manufacturing variances (probably not necessary with CPUs, motherboards, disks, however).
Then there are Windows switchers - they got fed up with the Wintel PC, some it was system crashes, some it's more religous reasons.
Linux switchers, often those who were working in Windows/Linux for various reasons. Lots have PowerBooks.
Then there are old NeXT users (not many of us actually!).
And others I'm too stupid to identify. I'm not sure that the Mac is a single culture anymore. I hope this is healthy for the platform.
I'm an OS/2 user who switched! I think all three of us did. Seriously, for many years (92-maybe 98 or 99) OS/2 was a more-or-less viable alternative to Winxx on the desktop. For a while I was waiting for Linux to "get there" but it didn't happen. OS X came along and it was easy to decide -- Unix underpinnings, modern GUI.
And loads of software, free and commercial! Mac people who complain they don't have the range of choice Win users do should try OS/2 for a month:-(
This review glossed over what was for me the most fascinating theme of the book: faith
and loyalty.
This is a highly significant theme in the Barrayar series as well. Not from the religious
angle (Miles Vorkosigan isn't religious although
he does burn an offering to the dead once or twice). But if the series is "about" any one thing, it's about faith and loyalty: upward to the Emperor and (perhaps especially) downward
to the troops Miles commands. In fact, all in
all it gives as good a portrayal as any fiction
I've ever read of the human side of
leadership: loyalty, relationship-building,
and, dare I say, moral character -- as opposed
to the technical-competence side (e.g., how to
attack a wormhole).
One thread that runs through most if not all of these hacks is that they make a computer work in some way that was never intended by the original designers. That, to me, is a key ingredient that distinguishes a hack from a non-hack.
In this spirit, how about "proxy ARP," which I have certainly heard called "the ARP hack." Imagine a scenario where machines A and B are on an Ethernet wire (a broadcast medium). B has a second interface and machine C is located there. A can't directly address C on the Ethernet; B must act as gateway.
In the old days before subnet masks and all, situations arose where machine A thought it should be able to address C (e.g., because C's IP address is in a range that A believes it can see). So when A wished to communicate with C, A would issue an ARP (a broadcast packet essentially asking all computers that hear it "Whose IP is this? If it's yours, return your hardware Ethernet address.") But C would never see the broadcast packet and would never reply to the ARP, so connectivity would fail.
The "ARP hack" solves this problem. Machine B answers the ARP request for C giving its own (B's) Ethernet address. A then happily sends packets intended for C to that address -- they arrive at B, who reads the IP address and happily forwards them (out the other interface) to C. The neat thing is that nothing special needed to be programmed in machines A or C at all, and nothing in B except for the special module that knew to answer ARPS for machines on the second interface. The IP forwarding machinery was all there anyway.
Slashdot Mirror
But the quoted excerpt from the article noted that this single-item measure accounted for a whopping
85% of the variance in Presidential voting intentions in surveys across three decades, and in this study
was correlatedstrongly with reported vote for Kerry or Bush. That strong evidence of the measure's
validity should outweigh an intuitive sense that the measure "may be flawed," no?
Bad analogy. Even ignoring the fact that the computer costs more than the TV.
1. The TV theft costs me a few hundred dollars, worst case (if not covered by insurance). The computer theft may cost me all my family photos, years of archived emails, difficult-to-reconstruct work documents, financial records, and even potentially embarrassing materials in some unknown person's hands. A much greater downside risk means greater security is appropriate.
2. If someone steals the TV I know it. If someone dumps my data to a CD (or accesses my data remotely over the network) I may not even know it. See #1: data theft can be massively damaging to me, perhaps even more if I am unaware (so I can't cancel my credit cards etc.)
My longstanding beef (yes, I have complained to them) is that this doesn't work for currency units. Try "240 euros in dollars" -- it does not recognize this as a unit conversion.
Nested lists are truly a top wish-list item, but the poster doesn't even mention the most important potential use -- for people (like me) who have music libraries larger than their iPod capacity. You can set up a size-limited playlist in iTunes, using whatever criteria you want, and autosync that with the iPod. But suppose you also want a playlist like "Not played in last 6 months" or "Recently added" or "Rated over 3"? You can set those up as Smart Lists in iTunes but can't sync them with the iPod because they will include tunes that aren't in the iPod.
Nested lists are exactly what is needed, to create a list of "Not played in last 6 months" or whatever within the iPod size-limited and autosynced list. Please, Apple...
OK, there are a lot of misconceptions here about how scientific publishing works. Let's try to straighten some out.
1. What is the current system? Most journals don't charge authors to publish, some do, some charge only for material that is particularly costly to reproduce (color photos as opposed to text or B&W graphs). In no reputable journal does author payment determine whether or not a paper is published (only, sometimes, how speedily). The decision as to publication is made by an editor assisted by peer reviewers (experts in the area of the paper). Submitted papers can be processed without much concrete expense (some clerical/administrative costs, a few thousand to tens of thousands honorarium to journal editors, no payment to peer reviewers). Typesetting, printing, and mailing cost money too. As a rough approximation to those concrete costs, consider that most journals are offered to individual subscribers at rates of $30-a few hundred per year. The big bucks come from library or institutional subscriptions, which range from a few hundred to many tens of thousands a year.
2. Why does this insane system survive, where scientific publishers basically have a license to print money at the expense of the libraries? Branding. The top journals have hard-earned reputations, due to consistently publishing high-quality, high-impact papers over the years. Authors want to publish in those journals, to advance their own reputations, and to reach the widest range of readers. Others in the field want to subscribe to those journals or have their libraries do so. Without this factor, internet distribution could make the whole peer-review, editing, and distribution process much cheaper (although not zero-cost), bypassing the scientific publishers.
3. Why should authors pay? Because they are in the best position to cover those remaining unavoidable expenses. Consider that scientific research can't be done for free; it is currently funded, often by government or industry grants, sometimes by universities themselves. They could simply add the costs of publication (which might be 5% of the cost of conducting the research in the first place) to their funding for research. Subsidies would have to be provided for papers from poor countries or from independent (non-university-affiliated) researchers. If author payments mean that the product of research is generally available on the internet, the resulting savings to universities in library subscription costs would be considerably greater than the extra costs of subsidizing publication.
4. Who would suffer under an author-pays, free-subscription system? The scientific publishers (commercial enterprises), and many scientific societies, who often own the journals that are published by commercial publishers and share in their profits. These are real issues. But on balance, the broader distribution of authors' ideas and the faster advancement of science due to more open communication is probably a more important consideration.
These comments don't address more radical changes in the system, e.g., going from the current peer-review arrangements to a Slashdot-style moderation plan. But I hope this post will help orient readers to some of the facts and issues.
In any case, with computer prices as low as they are today why on earth wouldn't a publication dispose of this worry and buy a unit for review off the shelf? With staff members' salaries and all, this would certainly not be the majority of the cost of generating a full review article. And, yes, with some components like LCDs, it would be worth buying three to look at manufacturing variances (probably not necessary with CPUs, motherboards, disks, however).
I'm an OS/2 user who switched! I think all three of us did. Seriously, for many years (92-maybe 98 or 99) OS/2 was a more-or-less viable alternative to Winxx on the desktop. For a while I was waiting for Linux to "get there" but it didn't happen. OS X came along and it was easy to decide -- Unix underpinnings, modern GUI.
And loads of software, free and commercial! Mac people who complain they don't have the range of choice Win users do should try OS/2 for a month
This is a highly significant theme in the Barrayar series as well. Not from the religious angle (Miles Vorkosigan isn't religious although he does burn an offering to the dead once or twice). But if the series is "about" any one thing, it's about faith and loyalty: upward to the Emperor and (perhaps especially) downward to the troops Miles commands. In fact, all in all it gives as good a portrayal as any fiction I've ever read of the human side of leadership: loyalty, relationship-building, and, dare I say, moral character -- as opposed to the technical-competence side (e.g., how to attack a wormhole).
In this spirit, how about "proxy ARP," which I have certainly heard called "the ARP hack." Imagine a scenario where machines A and B are on an Ethernet wire (a broadcast medium). B has a second interface and machine C is located there. A can't directly address C on the Ethernet; B must act as gateway.
In the old days before subnet masks and all, situations arose where machine A thought it should be able to address C (e.g., because C's IP address is in a range that A believes it can see). So when A wished to communicate with C, A would issue an ARP (a broadcast packet essentially asking all computers that hear it "Whose IP is this? If it's yours, return your hardware Ethernet address.") But C would never see the broadcast packet and would never reply to the ARP, so connectivity would fail.
The "ARP hack" solves this problem. Machine B answers the ARP request for C giving its own (B's) Ethernet address. A then happily sends packets intended for C to that address -- they arrive at B, who reads the IP address and happily forwards them (out the other interface) to C. The neat thing is that nothing special needed to be programmed in machines A or C at all, and nothing in B except for the special module that knew to answer ARPS for machines on the second interface. The IP forwarding machinery was all there anyway.