Except that my Ubuntu box is at home and can get dial-up *if* I run a telephone extension cord halfway through the house, whereas at work I have highspeed connection and can burn CDs.
How do I get the extra packages which are not included on the single install CD? I can download the.deb's, and simply transfer them, but I'd need to laboriously look up all of the dependencies to make sure I had all the required packages.
Here's a possible countermeasure. Construct your bomb. Shrink wrap it in plastic, taking care to get as little explosive residue on the outside as possible. Take it away from the bomb construction area, and wash the outside with strong soap etc. Give the result to another person.
They take it to somewhere clean of explosives residue, shrink wrap it in another layer, and carefully wash it, then hand it off to a third person who repeats the entire process again.
If you can reduce the explosives residues detectable by a factor of 100 or 1000 each time you do this, it can't take many iterations to reach undetectability - so long as the plastic is impervious to leakage. (Of course, then you need some way to program your hermetically sealed bomb. Also, you've forced many more people to become involved, which greatly increases the chance of betrayal before the bomb reaches its target.)
If this is practical, it must already have been tried to defeat drug-sniffing dogs. Does anyone have any ideas?
If you block cookies, it just shows you the flash premercial page over and over. (Yes, I block flash also.) I've tested this by accepting the cookie to see the article. I've searched for friendly copies elsewhere on the net, but failed to find any.
Have you considered the following possible features?
* Tech tree varies randomly between games (both prereqs and effects), so that you have to vary your strategy much more from game to game. In previous Civ games, you know that Railroads are amazingly useful, and exactly what techs you need to get there. However, while a 15th centuary genius might speculate on the possibility, they have no way to know how well it will work in practise. Some games, railroads are great. Some games they are a flop. The closer you get to them, the better your estimate of their usefulness becomes. (Another example: in 1950, commercial fusion power was thought to be 20 years away. It still is. What if they 1950's people had turned out to be right?)
* An option to shrink your empire for a greater score. The fun part is expanding, not managing the large empire that results. Allow an option where you can split your empire, keep playing only one half, and get a bonus multiplier to your end-of-game score. (E.g. you start of as England, and in 1776 become USA, and England becomes computer controlled.)
As an aside: I've often played "single city" games of Civ. I play at an easer level than normal, and allow myself only a single city. This allows a game in a single evening. From the Gamespot preview of Civ 4, it looks like you've already made allowance for fast games - thank you.
Yes, this "question" really is just a thinly disguised plug for my own ideas. Feel free to make me a job offer:-).
The manned version has an escape rocket attached to the capsule. I don't know which stages of flight it is effective in, but it may be they can escape during the SRB burn.
"It isn't immediately obvious to me how they can support their conclusions. Therefore, without any attempt to find out what their reasoning was, I declare that their conclusions are unsupportable."
So in a uniformly expanding universe, every point will eventually be observable (although the time until observability is exponential in the distance.)
This can fail for accelerating expansion - for example, if the rubber band doubles in length every second, the relative distance the ant travels halves every second.
My intuition is that the exponential time to arrive of the linear case is so close to never arriving, so even very mild accelerations (e.g constant acceleration, rather than exponential as in my example above) will suffice to prevent the ant reaching the end of a long-enough rubber band. However, I'm too lazy to prove it.
About 13.5 billion light years ago, the universe changed from being opaque to photons to being transparent (an event inappropriately called "recombination"). No photon emitted earlier than this time could reach us, so we cannot observe further than about 13.5 billion light years away. (The photons emitted at that time are the cosmic microwave background.) So the observable universe is 13.5 billion light years in radius. A billion years from now, it will be 14.5 billion light years in radius.
However, it gets more complicated: the universe is expanding, so the space that photon travelled through has got bigger in the meantime.
Imagine two points in the universe. Because the universe is expanding, the distance between them is increasing with time. The rate at which the distance increases is a velocity (which you can think of as causing the red shift of distant galaxies.) Hubble's law says this velocity is proportional to the distance between them. If they are sufficiently distant, the relative velocity is greater than the speed of light.
So (for example) imagine this is twice the speed of light. A photon emitted from one point travels towards the other. After one year, it has travelled one light-year, but the points have got two light-years further apart - clearly it will never arrive. These two points are not in each other's observable universes. The edge of our observable universe are the points which have a recession velocity equal to the speed of light.
The discussion above assumes no acceleration. Of course, astronomers from Hubble onwards knew there would be acceleration, but it wasn't until the mid 1990s that we could measure it.
It turns out, contrary to general expectation, that the expansion of the universe is now accelerating. This means that as time goes on, points don't have to be so far apart before their recession velocity exceeds the speed of light, so in a sense the observable universe is getting smaller. (In the sense that points that were within our observable universe in the past are no longer so. But remember that the points are always getting further apart - the radius of the obserable universe is increasing linearly with time.)
I am an ex-astronomer, not a cosmologist. There may be subtle errors in the above, but I hope not.
There are statistical tests for the "under selection" part. here's the first summary I found. It changes the frequencies of nearby neutral mutations which get to "ride the coat-tails" of the advantageous mutation.
The "related to brain size/function" is somewhat speculative, in that the gene could have additional unknown functions.
That the mutation makes us smarter is much more speculative. (Indeed, I don't think the paper's authors went this far.) It could, for example, make us 0.1% less smart, but reduce the brain's metabolic cost by 0.5%.
(Note: I've only read the linked article, not the scientific paper.)
Re:Waiting for apps isn't annoying, focus stealing
on
GNOME 2.12 Released
·
· Score: 1
Woohoo! It has really been bugging me too. Now all I have to do is wait for Debian to package it and get it stable - only 18 more months of having my focus stolen!
If I were a virus writer, and wanted to infect only a small number of machines, I'd do this following:
(1) Find some seldom used web page somewhere with a hits-counter on it. (2) Store the address of that web page in my virus, along with a limit count (say, 20,000.) (3) When the virus infects a new host, it visits the web page. If the hit counter is greater than the limit count (or the page is unavailable), the virus does not attempt to spread further.
Because the hits-counter was not set up by me, this can't be used to trace back to me. I have to be careful when finding such counters - server logs of who visited before the virus was released are a potential problem. I visit via anonymous proxies and/or using zombies to obscure my trail. Even better, if I find some site which automatically puts separate counters on many pages (e.g. blogs) I can assume the existence of a counter at a given URL without ever having visited the page.
Thank you for that - I had missed that paper. It is exactly the pattern I refered to, although I note that they say the statistics are not yet good enough to reject an out-of-Africa explaination.
For those not familiar with it: The Invisible Pink Unicorns are a counter-argument to a common straw-man attack on atheism.
The attack is "You can't know whether there is a god, so to disbelieve is every bit as much an act of faith as to believe."
The counter goes as following. Atheist (A): "Do you believe there are herds of invisble pink unicorns somewhere unnoticed on the planet?" Believer (B): "No, of course not." A: "So this would be an absolute religious conviction, would it?" B: "Well, no, not really" A: "Right. I don't believe in god in the same way you don't believe in invisible pink unicorns."
Actually, this merely says that any Neanderthal ancestors we may have weren't through the pure-maternal line. It says nothing at all about the nuclear DNA, which is over 99% of our DNA.
True, however:
My guess is that we'll never have good enough evidence of Neanderthal genes to show that there was no interbreeding at all. That requires study of the entire genome, and the fossil record doesn't have to have preserved it for us.
No - there would be a distinctive signal we can detect purely from modern human genomes. Imagine that for gene X, 1% of Europeans have a Neandertal gene, and everyone else (including all non-Europeans) have the Sapiens gene. We sequence this gene from 10000 people, 1000 of whom are European, 10 of whom have the Neandertal gene. Those 10 Europeans have sequences which are similar to each other, but are much more different from the consensus than any other gene sequences - and most significantly, much more different than any of the African samples. (Africa being the homeland of Homo Sapiens means it has the largest genetic variability.)
Putting it another way: if we created a phylogenetic tree of the genes, we would observe some of the European genes being basal (separated from the bulk of the sequences by the first bifurcation on the tree), and by a large margin (after this bifurcation, there is a long time before the next bifurcation on the main branch.)
We haven't yet observed such a pattern, although I think people have looked. We may yet find this, but the longer we look without finding it, the less likely the interbreeding hypothesis becomes.
IAATMP. (I am a theoretical molecular phylogenist.)
I don't get your point. Whatever we spend 6.5M pounds on (more lifeguards, overpriced military hardware, whatever), it doesn't disappear. If we don't spend it (lowering pool admission fees, or rates if it is a municipal pool), then it still doesn't disappear but gets spent on ice-creams, bigger mortgages etc.
Spending money is an allocation of limited resources (primarily labour) to a certain cause. If, for example we spend 65k pounds on an anti-drowning system and it is completely ineffective, then all the labour and resources that went into that system have indeed effectively disappeared. (Except that we now know what not to do again.)
I was comparing the cost/benefit to other systems - roading and airlines. In 1st world countries, safety upgrades are expected to save about one life per $1M (roughly). This system comes in at a similar (same order of magnitude) cost.
100 systems installed, 65k pounds per system = 6.5M pounds. Five lives saved (according to the article) = 1.3M pounds per life.
+: The systems are only recently installed, and have years of use yet, so should save many more. If they are 20% through their life-cycle, we can expect final cost around 260k pounds/life. +?: Perhaps the system will allow cost savings through fewer lifeguards. -: We're not 100% sure those people wouldn't have been saved anyway without the system. -: I haven't accounted for running costs, just purchase cost.
It is at least in the ball-park of cost-per-life-saved for other safety expenditure such as on airlines and roads - and it will get cheaper. So we can expect these to become wide-spread in the next decade.
For about 2-3 years (ending 2 1/2 years ago) I was administering ClearCase for about 100 developers. (I did other things as well, and other people helped with ClearCase, but I was the closest to an expert that they had.) Previously to this, we'd used SCCS. I have no experience with other SCM packages.
CC had some really neat features - the virtual file system, rigourous auditing of what files went into a build*, versioning of directories** etc. My major objections was that it seemed overly complicated (advanced features imply complexity, but it seemed to me that it had more than it needed) and opaque. There were times when something stuffed up, and I'd have loved to be able to look into the underlying database to see what was wrong, but the database is not directly accessible.
It is also extremely expensive.
We had about 3 million lines of code, and I think about 6000 source files. We did not use the "multisite" capabilities. As I remember it, performance only occasionally sucked.
Big merges were always a pain, causing lots of breakage. I'm not sure that any SCM package can avoid this.
* We didn't use the build-audit feature - there were some large library files (not standard.a or.so format) which would change pretty much every build and be linked by pretty much every compile, plus it took longer to check whether a file needed to be built than to just rebuild it anyhow. ** Versioning directories I expect is in any modern SCM package, but it wasn't in SCCS, which is what we moved from.
The ClearCase* source control system has a nifty way of making a database appear to be a file system. You tell the "file system" version of the source tree you want to see, and it makes it appear automagically.
Is this anything like what WinFS will be able to do? The one-line description (NTFS with metadata on files) doesn't make it sound like it, but it is hard to be sure.
If not like ClearCase, will it nevertheless be useful for source control?
I've thought for some time that the best managers are those who see their jobs backwards from the way most managers see their jobs: they act like they work for the people they manage. They help the employees work well together. They organize and make sure their different employees understand what is going on with the other employees. They evaluate the various obstacles that their employees are facing, and they try to remove those obstacles. They deal with executives and customers so you don't have to.
Well said, thank you. I have the same opinion, but hadn't thought of such a coherent framework for expressing it.
This view also short-circuits another possible problem: the idea that many managers have that they must be paid more than anyone in their team, which in turn leads to talented techies changing into poor managers, because it is the only way to advance. (One would still expect the mangers to be getting above median pay for the company, however.)
The last (only) time I worked in industry, one of my team-mates was the company's alpha geek. He had no underlings, was probably paid about twice what my team-leader was, and more than twice what I was, and was worth more than 4 times as much to the company. This is sensible.
Re:Bookmarks are better
on
Lucene in Action
·
· Score: 2, Insightful
Yep, I also think the poster's abandonment of bookmarks is truely bizzare. I have top-level folders of bookmarks, each of which becomes an instantly available pull-down menu. Most have subfolders.
I like your automatic classification ideas.
A complaint about Firefox: when I choose "bookmark this page" it comes up with a little dialog. This dialog has a one-line selector for where I want to create the bookmark (default being the folder named "bookmarks") and a little button to expand this one line into a screen. NOT ONCE in hundreds of times have I not clicked that little button to expand this display. The default behaviour assumes you don't classify your bookmarks, which is likely to only apply to people who seldom use bookmarks. I'd expect 80%+ of the time this dialog is invoked, the little button is clicked. So why doesn't it open the big display by default?
Truthfully, AMD could do it, even without Intel's permission. Just go grab a chip off the shelf and let loose.
But the credibility of the results would suffer. If Intel are producing the box, with their reputation on the line, you know they'll have the best possible motherboard, memory etc. for the purpose. If AMD built the Intel box, you don't have this confidence.
In short, the bill is bad. It reflects the worst kind of special interest law-making that hurts us all. And I mean REALLY hurts us because it will only act to discourage inventors. Record and movie companies beating-up on music and film pirates don't save or cost lives, but discouraging new medical inventions literally does cost lives.
Slashdot Mirror
- 1 install CD instead of 3 to 6
.deb's, and simply transfer them, but I'd need to laboriously look up all of the dependencies to make sure I had all the required packages.
Except that my Ubuntu box is at home and can get dial-up *if* I run a telephone extension cord halfway through the house, whereas at work I have highspeed connection and can burn CDs.
How do I get the extra packages which are not included on the single install CD? I can download the
Here's a possible countermeasure.
Construct your bomb. Shrink wrap it in plastic, taking care to get as little explosive residue on the outside as possible. Take it away from the bomb construction area, and wash the outside with strong soap etc. Give the result to another person.
They take it to somewhere clean of explosives residue, shrink wrap it in another layer, and carefully wash it, then hand it off to a third person who repeats the entire process again.
If you can reduce the explosives residues detectable by a factor of 100 or 1000 each time you do this, it can't take many iterations to reach undetectability - so long as the plastic is impervious to leakage. (Of course, then you need some way to program your hermetically sealed bomb. Also, you've forced many more people to become involved, which greatly increases the chance of betrayal before the bomb reaches its target.)
If this is practical, it must already have been tried to defeat drug-sniffing dogs. Does anyone have any ideas?
Here's the heatsink. The manufacturer has several similar products.
But I only have a Radeon 9200. (It does just great on Railroad Tycoon II and Master of Orion II. I think my AMD-64 3000+ may be overkill however.)
If you block cookies, it just shows you the flash premercial page over and over. (Yes, I block flash also.) I've tested this by accepting the cookie to see the article. I've searched for friendly copies elsewhere on the net, but failed to find any.
Have you considered the following possible features?
:-).
* Tech tree varies randomly between games (both prereqs and effects), so that you have to vary your strategy much more from game to game. In previous Civ games, you know that Railroads are amazingly useful, and exactly what techs you need to get there. However, while a 15th centuary genius might speculate on the possibility, they have no way to know how well it will work in practise. Some games, railroads are great. Some games they are a flop. The closer you get to them, the better your estimate of their usefulness becomes. (Another example: in 1950, commercial fusion power was thought to be 20 years away. It still is. What if they 1950's people had turned out to be right?)
* An option to shrink your empire for a greater score. The fun part is expanding, not managing the large empire that results. Allow an option where you can split your empire, keep playing only one half, and get a bonus multiplier to your end-of-game score. (E.g. you start of as England, and in 1776 become USA, and England becomes computer controlled.)
As an aside: I've often played "single city" games of Civ. I play at an easer level than normal, and allow myself only a single city. This allows a game in a single evening. From the Gamespot preview of Civ 4, it looks like you've already made allowance for fast games - thank you.
Yes, this "question" really is just a thinly disguised plug for my own ideas. Feel free to make me a job offer
The manned version has an escape rocket attached to the capsule. I don't know which stages of flight it is effective in, but it may be they can escape during the SRB burn.
No assumption, your rants abundantly reveal it.
Quick summary of the troll:
"It isn't immediately obvious to me how they can support their conclusions. Therefore, without any attempt to find out what their reasoning was, I declare that their conclusions are unsupportable."
Good analogy and analysis, thank you.
So in a uniformly expanding universe, every point will eventually be observable (although the time until observability is exponential in the distance.)
This can fail for accelerating expansion - for example, if the rubber band doubles in length every second, the relative distance the ant travels halves every second.
My intuition is that the exponential time to arrive of the linear case is so close to never arriving, so even very mild accelerations (e.g constant acceleration, rather than exponential as in my example above) will suffice to prevent the ant reaching the end of a long-enough rubber band. However, I'm too lazy to prove it.
Here's the simple answer:
About 13.5 billion light years ago, the universe changed from being opaque to photons to being transparent (an event inappropriately called "recombination"). No photon emitted earlier than this time could reach us, so we cannot observe further than about 13.5 billion light years away. (The photons emitted at that time are the cosmic microwave background.) So the observable universe is 13.5 billion light years in radius. A billion years from now, it will be 14.5 billion light years in radius.
However, it gets more complicated: the universe is expanding, so the space that photon travelled through has got bigger in the meantime.
Imagine two points in the universe. Because the universe is expanding, the distance between them is increasing with time. The rate at which the distance increases is a velocity (which you can think of as causing the red shift of distant galaxies.) Hubble's law says this velocity is proportional to the distance between them. If they are sufficiently distant, the relative velocity is greater than the speed of light.
So (for example) imagine this is twice the speed of light. A photon emitted from one point travels towards the other. After one year, it has travelled one light-year, but the points have got two light-years further apart - clearly it will never arrive. These two points are not in each other's observable universes. The edge of our observable universe are the points which have a recession velocity equal to the speed of light.
The discussion above assumes no acceleration. Of course, astronomers from Hubble onwards knew there would be acceleration, but it wasn't until the mid 1990s that we could measure it.
It turns out, contrary to general expectation, that the expansion of the universe is now accelerating. This means that as time goes on, points don't have to be so far apart before their recession velocity exceeds the speed of light, so in a sense the observable universe is getting smaller. (In the sense that points that were within our observable universe in the past are no longer so. But remember that the points are always getting further apart - the radius of the obserable universe is increasing linearly with time.)
I am an ex-astronomer, not a cosmologist. There may be subtle errors in the above, but I hope not.
There are statistical tests for the "under selection" part. here's the first summary I found. It changes the frequencies of nearby neutral mutations which get to "ride the coat-tails" of the advantageous mutation.
The "related to brain size/function" is somewhat speculative, in that the gene could have additional unknown functions.
That the mutation makes us smarter is much more speculative. (Indeed, I don't think the paper's authors went this far.) It could, for example, make us 0.1% less smart, but reduce the brain's metabolic cost by 0.5%.
(Note: I've only read the linked article, not the scientific paper.)
Woohoo! It has really been bugging me too. Now all I have to do is wait for Debian to package it and get it stable - only 18 more months of having my focus stolen!
If I were a virus writer, and wanted to infect only a small number of machines, I'd do this following:
(1) Find some seldom used web page somewhere with a hits-counter on it.
(2) Store the address of that web page in my virus, along with a limit count (say, 20,000.)
(3) When the virus infects a new host, it visits the web page. If the hit counter is greater than the limit count (or the page is unavailable), the virus does not attempt to spread further.
Because the hits-counter was not set up by me, this can't be used to trace back to me. I have to be careful when finding such counters - server logs of who visited before the virus was released are a potential problem. I visit via anonymous proxies and/or using zombies to obscure my trail. Even better, if I find some site which automatically puts separate counters on many pages (e.g. blogs) I can assume the existence of a counter at a given URL without ever having visited the page.
Thank you for that - I had missed that paper. It is exactly the pattern I refered to, although I note that they say the statistics are not yet good enough to reject an out-of-Africa explaination.
For those not familiar with it: The Invisible Pink Unicorns are a counter-argument to a common straw-man attack on atheism.
The attack is "You can't know whether there is a god, so to disbelieve is every bit as much an act of faith as to believe."
The counter goes as following.
Atheist (A): "Do you believe there are herds of invisble pink unicorns somewhere unnoticed on the planet?"
Believer (B): "No, of course not."
A: "So this would be an absolute religious conviction, would it?"
B: "Well, no, not really"
A: "Right. I don't believe in god in the same way you don't believe in invisible pink unicorns."
Actually, this merely says that any Neanderthal ancestors we may have weren't through the pure-maternal line. It says nothing at all about the nuclear DNA, which is over 99% of our DNA.
True, however:
My guess is that we'll never have good enough evidence of Neanderthal genes to show that there was no interbreeding at all. That requires study of the entire genome, and the fossil record doesn't have to have preserved it for us.
No - there would be a distinctive signal we can detect purely from modern human genomes. Imagine that for gene X, 1% of Europeans have a Neandertal gene, and everyone else (including all non-Europeans) have the Sapiens gene. We sequence this gene from 10000 people, 1000 of whom are European, 10 of whom have the Neandertal gene. Those 10 Europeans have sequences which are similar to each other, but are much more different from the consensus than any other gene sequences - and most significantly, much more different than any of the African samples. (Africa being the homeland of Homo Sapiens means it has the largest genetic variability.)
Putting it another way: if we created a phylogenetic tree of the genes, we would observe some of the European genes being basal (separated from the bulk of the sequences by the first bifurcation on the tree), and by a large margin (after this bifurcation, there is a long time before the next bifurcation on the main branch.)
We haven't yet observed such a pattern, although I think people have looked. We may yet find this, but the longer we look without finding it, the less likely the interbreeding hypothesis becomes.
IAATMP. (I am a theoretical molecular phylogenist.)
IANAL either, but I had the same reaction.
I don't get your point. Whatever we spend 6.5M pounds on (more lifeguards, overpriced military hardware, whatever), it doesn't disappear. If we don't spend it (lowering pool admission fees, or rates if it is a municipal pool), then it still doesn't disappear but gets spent on ice-creams, bigger mortgages etc.
Spending money is an allocation of limited resources (primarily labour) to a certain cause. If, for example we spend 65k pounds on an anti-drowning system and it is completely ineffective, then all the labour and resources that went into that system have indeed effectively disappeared. (Except that we now know what not to do again.)
I was comparing the cost/benefit to other systems - roading and airlines. In 1st world countries, safety upgrades are expected to save about one life per $1M (roughly). This system comes in at a similar (same order of magnitude) cost.
Back-of-the-envelope:
100 systems installed, 65k pounds per system = 6.5M pounds.
Five lives saved (according to the article) = 1.3M pounds per life.
+: The systems are only recently installed, and have years of use yet, so should save many more. If they are 20% through their life-cycle, we can expect final cost around 260k pounds/life.
+?: Perhaps the system will allow cost savings through fewer lifeguards.
-: We're not 100% sure those people wouldn't have been saved anyway without the system.
-: I haven't accounted for running costs, just purchase cost.
It is at least in the ball-park of cost-per-life-saved for other safety expenditure such as on airlines and roads - and it will get cheaper. So we can expect these to become wide-spread in the next decade.
For about 2-3 years (ending 2 1/2 years ago) I was administering ClearCase for about 100 developers. (I did other things as well, and other people helped with ClearCase, but I was the closest to an expert that they had.) Previously to this, we'd used SCCS. I have no experience with other SCM packages.
.a or .so format) which would change pretty much every build and be linked by pretty much every compile, plus it took longer to check whether a file needed to be built than to just rebuild it anyhow.
CC had some really neat features - the virtual file system, rigourous auditing of what files went into a build*, versioning of directories** etc. My major objections was that it seemed overly complicated (advanced features imply complexity, but it seemed to me that it had more than it needed) and opaque. There were times when something stuffed up, and I'd have loved to be able to look into the underlying database to see what was wrong, but the database is not directly accessible.
It is also extremely expensive.
We had about 3 million lines of code, and I think about 6000 source files. We did not use the "multisite" capabilities. As I remember it, performance only occasionally sucked.
Big merges were always a pain, causing lots of breakage. I'm not sure that any SCM package can avoid this.
* We didn't use the build-audit feature - there were some large library files (not standard
** Versioning directories I expect is in any modern SCM package, but it wasn't in SCCS, which is what we moved from.
The ClearCase* source control system has a nifty way of making a database appear to be a file system. You tell the "file system" version of the source tree you want to see, and it makes it appear automagically.
Is this anything like what WinFS will be able to do? The one-line description (NTFS with metadata on files) doesn't make it sound like it, but it is hard to be sure.
If not like ClearCase, will it nevertheless be useful for source control?
* by Atria, then Rational and now IBM.
I've thought for some time that the best managers are those who see their jobs backwards from the way most managers see their jobs: they act like they work for the people they manage. They help the employees work well together. They organize and make sure their different employees understand what is going on with the other employees. They evaluate the various obstacles that their employees are facing, and they try to remove those obstacles. They deal with executives and customers so you don't have to.
Well said, thank you. I have the same opinion, but hadn't thought of such a coherent framework for expressing it.
This view also short-circuits another possible problem: the idea that many managers have that they must be paid more than anyone in their team, which in turn leads to talented techies changing into poor managers, because it is the only way to advance. (One would still expect the mangers to be getting above median pay for the company, however.)
The last (only) time I worked in industry, one of my team-mates was the company's alpha geek. He had no underlings, was probably paid about twice what my team-leader was, and more than twice what I was, and was worth more than 4 times as much to the company. This is sensible.
Yep, I also think the poster's abandonment of bookmarks is truely bizzare. I have top-level folders of bookmarks, each of which becomes an instantly available pull-down menu. Most have subfolders.
I like your automatic classification ideas.
A complaint about Firefox: when I choose "bookmark this page" it comes up with a little dialog. This dialog has a one-line selector for where I want to create the bookmark (default being the folder named "bookmarks") and a little button to expand this one line into a screen. NOT ONCE in hundreds of times have I not clicked that little button to expand this display. The default behaviour assumes you don't classify your bookmarks, which is likely to only apply to people who seldom use bookmarks. I'd expect 80%+ of the time this dialog is invoked, the little button is clicked. So why doesn't it open the big display by default?
Sorry, I have no mod points for you today.
Truthfully, AMD could do it, even without Intel's permission. Just go grab a chip off the shelf and let loose.
But the credibility of the results would suffer. If Intel are producing the box, with their reputation on the line, you know they'll have the best possible motherboard, memory etc. for the purpose. If AMD built the Intel box, you don't have this confidence.
(also negative) here.
In short, the bill is bad. It reflects the worst kind of special interest law-making that hurts us all. And I mean REALLY hurts us because it will only act to discourage inventors. Record and movie companies beating-up on music and film pirates don't save or cost lives, but discouraging new medical inventions literally does cost lives.