I think you might be misreading the last graph. They are not directly testing video performance (they say in the text that all platforms could play 720p WMV and mp4 perfectly), they are comparing video via flash. So it's (in a way) a second flash benchmark.
Ubuntu beat win7 on boot and opening office docs. XP is a much simpler system, something like Xubuntu might be a fairer test there.
This study compares the survival of people with similar diseases once they become ill.
Re:Thrash for several hours before benchmarking
on
How Do Browsers Scale?
·
· Score: 3, Informative
It can also be poorly-coded websites.
The BBC's news page used to have an annoying javascript news ticker that ran across the top of the page. As it ran it built a slowly larger and larger array of something or other and memuse would slowly creep up. When I stopped having that as my home page my ff memory problems stopped.
Dead thread, but anyway. There are two issues here I think: 1) Windows is missing a package manager, so maintaining a stack of dependant libraries has to be done manually, and 2) Windows is missing a standard from-source build system (an equivalent to./configure ; make ; make install).
Most things that have been ported to Windows from Unix still work better on Unix. Things like Apache, PHP, MySQL, Python, Ruby, Perl, git, svn,... the list is almost endless. At a user level, gnumeric, pidgin, inkscape and gimp are all available as Windows binaries, but they run poorly and are usually old versions.
Part of the problem is that building software on Windows is monstrously difficult. Try building php (including dependant libraries) from source and see if you can do it in less than a week.
Finland has half the population density of the US and far faster, cheaper broadband. NYC has huge population density, but very slow, very expensive broadband.
The key factor is competition: US infrastructure owners are allowed to block competitors from using their bits of wire. This creates an almost insurmountable barrier to entry on the market and effectively establishes local monopolies. Consumers have little or no choice, usually.
Everywhere else in the world has a regulatory framework that enforces open access: owners of infrastructure have to sell access to their cabling to all comers at non-discriminatory rates. As a result setting up an ISP is cheap and easy, there is enormous competition, and consumers get fast broadband for chickenfeed.
I don't know of a Linux equivalent to this attack. No Linux desktop will download and execute a file when you click on a link. No linux desktop will even let you execute a downloaded file unless you first right-click on it and set the executable bit. And even then it will only run as the user. And even code running as the user won't have access to the users passwords, since they are encrypted.
It's not all roses, of course. GNOME.desktop files will execute even without the x bit set, annoyingly, though I think they are planning to change that. And as you say being limited to the one user is good, but not good enough for the user whose files got hosed.
They mean that the data was widely available from many sources and anyone could download it. The "competent" means selecting and combining data from among the various sources takes a degree of skill and knowledge.
One of the most common allegations made against the CRU scientists was that they blocked access to raw data, drawn from weather stations around the world, and adjusted that data to falsely show a pattern of global warming. There were also complaints that they failed to release on demand the computer code they wrote to analyse the data. Without such information, how could sceptics check the CRU's calculations?
The panel showed that it was relatively straightforward to reproduce the CRU analysis without needing to ask Jones and his colleagues for anything.
They used data from public databanks and wrote their own computer code, which they say could be repeated by any "competent researcher". The results were similar to those of the CRU.
The panel found that they did not subvert the peer review process to censor criticism as alleged, and that the key data needed to reproduce their findings was freely available to any "competent" researcher.
Other materials, such as aluminium, are much more harmful. Aluminium of course is used heavily in all cars. From the article:
... the researchers discovered that only 15% of the total environmental impact of building the car could be attributed to the battery pack. Of that, only 2.3% came from mining and processing raw lithium.
Other materials used in lithium-ion batteries such as copper and aluminium, attributed 7.5% of the environmental burden.
Lead is a very serious toxin, especially in children. It causes measurable impairment of intelligence at even very, very low levels. There does not seem to be a safe level of exposure.
50% of people's lead exposure comes though the environment, so regulatory authorities are working hard to eliminate it where they can.
But WGII is supposed to be a review of all available material, including non-peer-reviewed material. That's why it's 3,000+ pages. It's not supposed to be rigorously scientific. It says so, all over it.
WGI is the science bit and, as you say, it's much less flabby.
It's things like webcams, audio, low-level graphics and sockets, as I recall. On IE8 (the one with a sandbox) the flash plugin is split into two parts. They have an in-sandbox stub which communicates over a channel to an out-of-sandbox process which contains the actual flash player. This out-of-sandbox player runs with full rights.
So even if FF did have a full sandbox for plugins, Adobe would probably make a hole in it for Flash.
The issue was data protection, not the cards themselves.
UK data protection law (I think this is an EU-wide thing now?) says (among other things) thst you can't use personal information gathered for one purpose for another purpose without the consent of the people involved. This means you can't link databases together. The TV licence database can't be linked to the healthcare database or the police database or... well, anything really.
Two things help enforce this separation. First, it's illegal (heh), and second, it's impossible to do automatically since all these databases have different ways of establishing identity. There's no 'citizen number' that can be used as a common key for a join, and no way to make one (how can you be sure that the JAMES SMITH of 23 Pootle Gardens in the car license DB is the same JIM SMITH of 23 Potle Gdns in the TV license DB?). One of the purposes of the UK ID card scheme was to introduce a robust citizen ID that could be a common way to index databases (and could reduce costs by having a single identity register).
So the concern was that ID cards were a prelude to the more-or-less complete loss of data protection, at least as far as data held by government went. Moves were already being made last year to grant large data protection exemptions to government.
The ID database would no doubt have crept into the private sector too and be used to identify people for bank accounts and internet services and all that stuff as well. It's easy to imagine a future where data protection no longer really exists at all, where even minor government officials (perhaps under an 'anti-terrorist' banner) could browse almost every piece of information held anywhere on any UK citizen.
Anyway, the loss of the national ID register makes this, at least technically, much more difficult.
It's the other way around, actually. 80 - 90% of a vehicle's lifetime energy use is in driving it around. You can google many versions of this calculation, but here's one from Slate.
You might be remembering the report from a few years ago that claimed a Hummer was more efficient than a Prius, but that's been pretty thoroughly debunked many times now.
Ahem, and replying to myself, I think you underestimate just how specialised people get.
To stick with the medical analogy, you might get a very basic knowledge of biology with a few 100 hours of study, but how will you be able to comment on the work of someone who has spent 10 years looking at the role of RhoB in the apoptopic pathway?
Even people at this level are unable to comment on each other's work unless they have studied almost exactly the same problem for a similar length of time. When I submit a paper I can almost always guess the reviewers the journal has selected since there are only a few dozen people who COULD comment on it.
I remember my father (also an academic) telling me a story about a prof who visited his college to give a seminar. My father had to sit next to him at the formal supper that night. The visitor was a very, very boring man, but his redeeming quality was that he was the world's leading authority on spider's webs.
Trying to make conversation, my father asked him how many types of thread went into a typical cobweb. "Sorry old chap," said the visitor, "I'm more of a gossamer man".
Studying for a few 100 hours will get you enough that you can follow the basics of a debate or perhaps make a timid comment if you feel brave, but you will not be an expert and you will not be able to judge the work of experts.
I changed field about 5 years ago to bioinformatics and, having to work with senior medical researchers, I'm made horribly aware every day how limited my knowledge of biology is. And that's with a lot more than a few 100 hours of study on my part.
Fortunately for me they are comically clueless regarding CS, heh, so they keep me around.
Haha, that's amazing, there are actually people there claiming that the scientists have invented the whole thing to get more research funding. It's worse than youtube comments.
That kind of thing could never happen on slashdot!
They didn't try and verify the CRU's results (as you say, it'd be a huge undertaking), they were simply looking for evidence of deliberate fraud or poor technique. After the emails were leaked you'll remember a huge hoo-ha about "hiding the decline", "tricks" and allegations of cherry-picking data, especially in regard to the tree studies.
This committee has concluded that those allegations are unfounded: that there was no deliberate deceit, that data was not cherry-picked, and that, while some parts of their methodology could probably be improved, the science was generally sound.
Whether their conclusions were actually correct... that's another issue!
[Lord Oxburgh] stated: "We found absolutely no evidence of any impropriety whatsoever. That doesn't mean that we agreed with all of their conclusions, but these people were doing their jobs honestly."
Slashdot Mirror
I think you might be misreading the last graph. They are not directly testing video performance (they say in the text that all platforms could play 720p WMV and mp4 perfectly), they are comparing video via flash. So it's (in a way) a second flash benchmark.
Ubuntu beat win7 on boot and opening office docs. XP is a much simpler system, something like Xubuntu might be a fairer test there.
The summary is misleading. Brits, on average, outlive Americans.
http://en.wikipedia.org/wiki/List_of_countries_by_life_expectancy
This study compares the survival of people with similar diseases once they become ill.
It can also be poorly-coded websites.
The BBC's news page used to have an annoying javascript news ticker that ran across the top of the page. As it ran it built a slowly larger and larger array of something or other and memuse would slowly creep up. When I stopped having that as my home page my ff memory problems stopped.
You need click2flash: http://github.com/rentzsch/clicktoflash
Dead thread, but anyway. There are two issues here I think: 1) Windows is missing a package manager, so maintaining a stack of dependant libraries has to be done manually, and 2) Windows is missing a standard from-source build system (an equivalent to ./configure ; make ; make install).
I'm hoping CoApp will fix this. Fingers crossed!
Do you mean this?
http://www.amazon.com/Samsung-7-Inch-UMPC-with-Touchscreen/dp/B001TZVW8E 1 new from $835.18
That's not the same product, that's a winxp thing with a hard drive.
Most things that have been ported to Windows from Unix still work better on Unix. Things like Apache, PHP, MySQL, Python, Ruby, Perl, git, svn, ... the list is almost endless. At a user level, gnumeric, pidgin, inkscape and gimp are all available as Windows binaries, but they run poorly and are usually old versions.
Part of the problem is that building software on Windows is monstrously difficult. Try building php (including dependant libraries) from source and see if you can do it in less than a week.
Finland has half the population density of the US and far faster, cheaper broadband. NYC has huge population density, but very slow, very expensive broadband.
The key factor is competition: US infrastructure owners are allowed to block competitors from using their bits of wire. This creates an almost insurmountable barrier to entry on the market and effectively establishes local monopolies. Consumers have little or no choice, usually.
Everywhere else in the world has a regulatory framework that enforces open access: owners of infrastructure have to sell access to their cabling to all comers at non-discriminatory rates. As a result setting up an ISP is cheap and easy, there is enormous competition, and consumers get fast broadband for chickenfeed.
Here's a lecture by Lessig on the subject:
http://lessig.blip.tv/file/3485790
I checked again, and it sounds like this has been addressed. Phew!
https://wiki.ubuntu.com/SecurityTeam/Policies#Execute-Permission%20Bit%20Required
I don't know of a Linux equivalent to this attack. No Linux desktop will download and execute a file when you click on a link. No linux desktop will even let you execute a downloaded file unless you first right-click on it and set the executable bit. And even then it will only run as the user. And even code running as the user won't have access to the users passwords, since they are encrypted.
It's not all roses, of course. GNOME .desktop files will execute even without the x bit set, annoyingly, though I think they are planning to change that. And as you say being limited to the one user is good, but not good enough for the user whose files got hosed.
They mean that the data was widely available from many sources and anyone could download it. The "competent" means selecting and combining data from among the various sources takes a degree of skill and knowledge.
From http://www.guardian.co.uk/environment/2010/jul/07/climategate-scientists-main-points:
One of the most common allegations made against the CRU scientists was that they blocked access to raw data, drawn from weather stations around the world, and adjusted that data to falsely show a pattern of global warming. There were also complaints that they failed to release on demand the computer code they wrote to analyse the data. Without such information, how could sceptics check the CRU's calculations?
The panel showed that it was relatively straightforward to reproduce the CRU analysis without needing to ask Jones and his colleagues for anything.
They used data from public databanks and wrote their own computer code, which they say could be repeated by any "competent researcher". The results were similar to those of the CRU.
They've been cleared of this allegation. From the linked wiki article:
The panel found that they did not subvert the peer review process to censor criticism as alleged, and that the key data needed to reproduce their findings was freely available to any "competent" researcher.
Other materials, such as aluminium, are much more harmful. Aluminium of course is used heavily in all cars. From the article:
It's in the article:
Honestly, it's worth a read. You need to understand the 'establishment' case if you plan to criticise it.
This is a standard list of objections, all of which are addressed by every "top ten climate myths" list every science magazine has ever published.
For example, here's the New Scientist (the UK equivalent of Scientific American) list:
http://www.newscientist.com/article/dn11462-climate-change-a-guide-for-the-perplexed.html
It answers all your points (I think) and several others as well.
Lead is a very serious toxin, especially in children. It causes measurable impairment of intelligence at even very, very low levels. There does not seem to be a safe level of exposure.
50% of people's lead exposure comes though the environment, so regulatory authorities are working hard to eliminate it where they can.
But WGII is supposed to be a review of all available material, including non-peer-reviewed material. That's why it's 3,000+ pages. It's not supposed to be rigorously scientific. It says so, all over it.
WGI is the science bit and, as you say, it's much less flabby.
It's things like webcams, audio, low-level graphics and sockets, as I recall. On IE8 (the one with a sandbox) the flash plugin is split into two parts. They have an in-sandbox stub which communicates over a channel to an out-of-sandbox process which contains the actual flash player. This out-of-sandbox player runs with full rights.
So even if FF did have a full sandbox for plugins, Adobe would probably make a hole in it for Flash.
The issue was data protection, not the cards themselves.
UK data protection law (I think this is an EU-wide thing now?) says (among other things) thst you can't use personal information gathered for one purpose for another purpose without the consent of the people involved. This means you can't link databases together. The TV licence database can't be linked to the healthcare database or the police database or ... well, anything really.
Two things help enforce this separation. First, it's illegal (heh), and second, it's impossible to do automatically since all these databases have different ways of establishing identity. There's no 'citizen number' that can be used as a common key for a join, and no way to make one (how can you be sure that the JAMES SMITH of 23 Pootle Gardens in the car license DB is the same JIM SMITH of 23 Potle Gdns in the TV license DB?). One of the purposes of the UK ID card scheme was to introduce a robust citizen ID that could be a common way to index databases (and could reduce costs by having a single identity register).
So the concern was that ID cards were a prelude to the more-or-less complete loss of data protection, at least as far as data held by government went. Moves were already being made last year to grant large data protection exemptions to government.
The ID database would no doubt have crept into the private sector too and be used to identify people for bank accounts and internet services and all that stuff as well. It's easy to imagine a future where data protection no longer really exists at all, where even minor government officials (perhaps under an 'anti-terrorist' banner) could browse almost every piece of information held anywhere on any UK citizen.
Anyway, the loss of the national ID register makes this, at least technically, much more difficult.
Ooops, sorry. The wiki page for the prius suggested this PDF report but I didn't want to link to a PDF :-(
The CNW study has also been covered here at /..
It's the other way around, actually. 80 - 90% of a vehicle's lifetime energy use is in driving it around. You can google many versions of this calculation, but here's one from Slate.
You might be remembering the report from a few years ago that claimed a Hummer was more efficient than a Prius, but that's been pretty thoroughly debunked many times now.
Ahem, and replying to myself, I think you underestimate just how specialised people get.
To stick with the medical analogy, you might get a very basic knowledge of biology with a few 100 hours of study, but how will you be able to comment on the work of someone who has spent 10 years looking at the role of RhoB in the apoptopic pathway?
Even people at this level are unable to comment on each other's work unless they have studied almost exactly the same problem for a similar length of time. When I submit a paper I can almost always guess the reviewers the journal has selected since there are only a few dozen people who COULD comment on it.
I remember my father (also an academic) telling me a story about a prof who visited his college to give a seminar. My father had to sit next to him at the formal supper that night. The visitor was a very, very boring man, but his redeeming quality was that he was the world's leading authority on spider's webs.
Trying to make conversation, my father asked him how many types of thread went into a typical cobweb. "Sorry old chap," said the visitor, "I'm more of a gossamer man".
Studying for a few 100 hours will get you enough that you can follow the basics of a debate or perhaps make a timid comment if you feel brave, but you will not be an expert and you will not be able to judge the work of experts.
I changed field about 5 years ago to bioinformatics and, having to work with senior medical researchers, I'm made horribly aware every day how limited my knowledge of biology is. And that's with a lot more than a few 100 hours of study on my part.
Fortunately for me they are comically clueless regarding CS, heh, so they keep me around.
Haha, that's amazing, there are actually people there claiming that the scientists have invented the whole thing to get more research funding. It's worse than youtube comments.
That kind of thing could never happen on slashdot!
They didn't try and verify the CRU's results (as you say, it'd be a huge undertaking), they were simply looking for evidence of deliberate fraud or poor technique. After the emails were leaked you'll remember a huge hoo-ha about "hiding the decline", "tricks" and allegations of cherry-picking data, especially in regard to the tree studies.
This committee has concluded that those allegations are unfounded: that there was no deliberate deceit, that data was not cherry-picked, and that, while some parts of their methodology could probably be improved, the science was generally sound.
Whether their conclusions were actually correct ... that's another issue!
[Lord Oxburgh] stated: "We found absolutely no evidence of any impropriety whatsoever. That doesn't mean that we agreed with all of their conclusions, but these people were doing their jobs honestly."
http://news.bbc.co.uk/1/hi/8618024.stm