This "study" didn't measure browser speed at all. It compared only the speeds of the javascripts that the browsers use. TFA says so fairly clearly.
If you're making heavy use of sites that are mostly javascript, this is a useful study. For the rest of us, it's yet another case of measuring a tiny corner of what is claimed, and then asserting that this measures the whole thing.
Using similar reasoning, we can imagine an oceanographer measuring the parts of the ocean along the beaches where most people are found, and concluding that the oceans average about 2 meters deep. (There's gotta be a good auto analogy here, too.)
As someone else has pointed out, most "power users" of browsers mostly disable java and javascript (and Active-X and any other misfeature that lets strangers run code on their machines). They may use NoScript with FF and enable JS for selected sites. Or they may simply copy the links to another browser such as opera or safari when they want to use JS. So to them, firefox and mozilla may well be the fastest browsers, since they permit easy selective disabling of all scripting features.
And we should also note that the time to render most web pages is mostly the download time. If due to network delays it takes 23 seconds to download a page, and browser X renders it in.001 sec while browser Y renders it in.01 sec, there's no practical meaning to a claim that Y renders 10 times faster than X. If the page takes 23.001 sec to render in X, and 23.01 sec in Y, few people will be able to reliably tell you which is faster.
If this were announced as a comparison of various JS interpret speeds, I'd take it seriously. But claiming that it's about browser speed pretty well discredits the authors (and the editor who wrote the summary).
Funny, I don't see it refreshing at all. I get a pop-down saying:
Authentication Required
A username and password are being requested by http://beta.linux.com./ The site says: "Linux.com is under maintenance. Please check www.linux.com shortly"
This has username and password input widgets, plus Cancel and OK buttons. It doesn't seem to have any way of setting up an account, so I tried a couple of dummy loging, hit OK -- and got back exactly the same pop-down windowlet. So I hit Cancel -- and got the same pop-down windowlet.
I notice that the button to close the window is greyed out...
I'll check them occasionally and see if anything works. At least they're not soaking up all my cpu while they wait for me.
If the NSA (No Such Agency) is in charge, it'll be the same as having no security oversight at all. They naturally keep everything secret, so if they want to tell you to do something, you won't have the security clearance to read the order or any of its details.
Yes, they can write secret orders, not show them to you, and then prosecute you for not obeying them. But this has been true for around a decade now, so it won't be anything new.
Anyway, the main area where security is important is in the corporate world's handling of its comprehensive information about all of us. And in the modern US, agencies of the government don't give orders to corporations; the corporations give orders to the government. So corporate databases will continue to be as insecure as always, which doesn't really matter because the information is always for sale to the highest bidder, secure or not. Security really means that the information can't be read by anyone who hasn't paid for it, y'know.
If there are any changes, the most likely are that the NSA will be forced to adopt corporate-style "security" measures such as 4-digit PINs or password rules so complex that you have to write your passwords down and carry them in your wallet. And they'll routinely leave entire databases in laptops inside parked cars. This will be by policy, not accident. It'll result in more funny news stories; we'll mostly laugh and go about our lives.
I'd add a;-), but I'm not sure that this actually qualifies as humor...
(I'm sure that Jon Stewart and Steven Colbert will explain it much better than I can.)
- set socket into non-blocking mode - call connect() - check for immediate success (happens on solaris over loopback interface)
Well, I've tried that; it only works if the connect() returns in the calling process. With the bug I wrote about, the problem is that this doesn't happen, even with a nonblocking socket. Also, an ALARM is supposed to interrupt a system call and force a failure return, but this doesn't happen, either.
sometimes even a "kill -9" won't kill the process That sounds odd. Which OS was that?
I've documented it on FreeBSD, NetBSD, Solaris and several flavors of linux. It's been a couple of years since I've worked on a project where I could test this, so I can't tell you the releases.
It may not be a single bug, as similar "hangs, unkillable" problems have been reported in all sorts of systems since the dawn of "time sharing" computer systems back in the 1960s. They're usually good classroom studies for subtle kernel bugs, since diagnosing the problem requires a more-or-less complete path analysis of the code to find a "can't get there from here" scenario, where "there" is the code that returns to the process and "here" is the state in some system dumps.
A big part of the problem is that, in my experience, when the problem happens, it's typically after several million connect() calls by the same process. Imagine a search bot, for example, which might have to read millions or billions of URLs, typically with a connect() per URL (though this can be optimized in the obvious way).
Most people treat "one in a million" as an idiomatic way of saying "extremely rare". But there are 86400 seconds in a day, so if you call connect() as few as 12 times per second, you hit a million in a bit under a day. (23 hours, 8 minutes, 53 seconds, actually.;-)
But in my experience, the bug isn't usually reproducible. During testing, I have had the process log all its connect() calls, so I could tell who it was trying to connect to. Usually restarting it for that address will result in success, which doesn't tell you much. I've managed to reproduce the problem maybe a total of a couple of dozen times, and the only thing I find in common is that the remote system has report that it was running some versions of Windows and IIS. But the sample size is far too small to bash Microsoft for this (other than the fun of doing so;-). And most Windows+IIS systems connect just fine. I've seen the problem when the remote system was various unixoid OSs, but that has never been reproducible. And in any case, the culprit could easily be some intermediary bridge or router doing something "intelligent" with the connection attempt.
OTOH, the problem long predates any reports of ISPs like Comcast intentionally screwing with connections in order to cause problems for one end or both, which only go back a few years (to my knowledge). Before that, it almost had to be some incompatibility between the two ends of the connection.
I've worked on any number of projects in which we created a new network API, usually with UI tools to match. Of course, our package always used on sockets as the lower-level "internal" basis.
It's called "layering". Some network programmers have learned that it's a useful approach.
So far, I've never seen a different networking UI that's easier to program (or debug) than sockets. I keep reading articles like this on the topic, but I'm still looking for one that's better for the job. It's possible that the Berkeley people found the best low-level approach.
However, one thing that they missed that has been sorely needed on a number of projects is a "timeout" parameter to the connect() call. Time and again I've seen cases where an app hangs inside a call to connect() and never returns. Typically setting an alarm won't interrupt the call, and sometimes even a "kill -9" won't kill the process. Sometimes only a reboot will get rid of the zombie process. If the OS no longer gives your process any cpu time, it doesn't matter what clever code you have in it to diagnose problems. This seems to happen under unknown conditions on all OSs (though I haven't actually tested it on all releases of all OSs, so I could be wrong).
But this is an implementation detail of what's basically a fairly sound design.
Copyright is an arbitrary, artificial construct "designed" (I use the word generously) to encourage creation.
Actually, if you dig into the history of copyright, you'll find that this wasn't at all true. The earliest copyrights were given by monarchs to publishers of texts such as the Bible. They clearly didn't do this to encourage the creation of more bibles. They did it to limit the production so the masses couldn't get and learn to read such texts on their own, but had to rely on their betters to read and interpret the holy texts. They also did it to collect the royalties (note the root of that word) from the publishers who were permitted to do the publishing.
The idea that copyright is to encourage creation is a bit of PR created by the publishers to justify their monopoly. But from the beginning of copyright, it was a total lie. And can anyone find a study (with valid scientific methodology) that supports the idea that copyright every has actually resulted in the creation of more literature? Yes, we can find lots of claims that this is true. But has anyone presented actual evidence that it works?
There's a competing hypothesis (which probably hasn't been well-tested either): Creative people like to create things, and will do so even if not rewarded. Copyright was developed by the people who distribute the creations to customers, as a way of creating a monopoly market in creative works, to the financial benefit of the distributors. This is widely believed by many of those creative people, mostly the ones who aren't making a living from their creation because the contracts they have to sign give most of the income to the publishers and distributors.
People still buy books and newspapers because the portability and physical experience is better than a notebook or even a Kindle.
Books perhaps, but at least here in the US we're hearing a lot of analyses of the rapid decline of the newspaper business.
Part of this seems to be due to the fact that the conventional newspaper isn't in fact a very convenient format for readers. The pages are too big and the material is too flimsy. You can hold most books in one hand as you read, with your cup of coffee or whatever in the other hand, but this doesn't work at all with a newspaper. Anywhere but at home on your kitchen table, it's a very difficult format to read easily.
Something that comes up occasionally in the discussions of the newspaper industry's woes is the suggestion that "Nobody ever bought a newspaper because they want piles of low-quality, cheaply-printed paper in their house". People don't want the paper; they want the news. News is mostly a throwaway commodity, so it has to be done cheaply or people won't buy. But the resulting product has never been very good physically.
A funny thing about the discussions I've heard or read: When someone points out the above problem, it is usually ignored. That is, the others don't switch to talking about the "news industry"; they continue to talk about "newspapers", missing the point entirely that nobody actually wants the paper. We're mostly happy to be rid of the stuff. But the pundits seem to be willfully ignoring this point, and continue to talk as if the paper were the important product. As long as the news(paper) companies keep insisting that it's the paper that's important, they don't stand much of a chance of changing their business model over to electronic news distribution.
A few news companies have caught onto this, but most of them don't seem to understand. So their hard-copy news business is dying, and others that learn how to switch to the new, cheaper distribution medium will be the ones who survive. They just need to learn how to handle the advertising part, since so far that seems to be how they make their money (just as with the old newspapers).
First we need some laws on the subject that were not distorted from 200 years ago to fit the internet.
One of the general problems is that as soon as a computer is introduced to a subject area, all precedent is forgotten, chants of "That's different!" are heard repeatedly, and we humans must relearn every social lesson that we so laboriously worked out over the centuries.
Free speech? We're using computers; it isn't "speech", it's email or texting or...
Freedom of the press? We're using computers, not presses; we're producing blogs and RSS feeds.
Fair use? We're using computers, which do copying for even the most trivial operations, so we have to throw the idea out and look for something else.
It'll take us a long time to figure out how to transfer all the rights our ancestors fought and died for to the brave new computerized world. Because none of those lessons are relevant when you use a computer to do something.
So your company doesn't want to hear about bug reports, UI problems, security issues, etc.
How dumb can a company's management be?
On numerous occasions, I've bought something partly because the vendor was open about problems. If a company pretends that there are no problems with its products, I suspect that they're suppressing news about something really awful.
it's important to remember that "monopoly" when used here doesn't mean 100% of the market, but (like MS) enough of the market that it might as well be 100%, or at least large enough that they can exercise anti-competitive behavior.
One of the clichés in economics texts is the "5-50" rule of thumb saying that a "market" acts like a monopoly if 5 or fewer companies get 50% or more of the sales.
Of course, like any rule of thumb, this is basically "economics for dummies", because the reality is that there's a continuum of actual behaviors. Some big companies are run by people with ethics and a long-term view (though they tend to disappear with time). Some markets have sufficient delivery problems that they act like local monopolies even with a hundred companies.
But the point of such things is to debunk the traditional even sillier idea that you only have a "monopoly" if there is just one company. This is called the Etymological Fallacy, the idea that the meaning of a word is defined by the meanings of its parts in the original (long-dead) languages. It's popular with the people who like the idea of unbridled, lassez-faire capitalism. But that's not how economists or most other people use the term in English. In the real world, there are such things as "gentlemen's agreements" that produce monopoly markets even when there are several sellers.
It's fairly clear to nearly everyone that the US retail computer business is a monopoly market, although there are two companies supplying the core hardware and two companies providing the OSs. A small fraction of the population can actually name the second software supplier (though very few can name either hardware supplier). But it's been that way here for a few decades now, so we don't expect that we'll see an actual free market in computer retailing in our lifetime. It's interesting reading about efforts in other parts of the world to do something about the monopoly. It'll be even more interesting if they actually succeed, and make it possible for smaller startups to actually do business.
That's indeed a danger, and made worse by the media's tendency to compare every little search project with google, even when there's little or no overlap in what they do.
OTOH, I've been involved in a few of the thousand or more projects to build highly-specialized search tools that "understand" the data used for specialized purposes. So far, none of them has to my knowledge been approached by anyone from google. In a few cases, the opposite has happened, as the people in a project get to wondering whether any of google's expertise could be useful to them. Again, as far as I can tell, the answer is "No", but I've only seen a few small corners of the field.
It does seem fairly clear that we're better off with a flock of independent projects to solve the world's search problems. There's a well-known problem with human organizations, summarized by the old saying that if your only tool is a hammer, everything looks like a nail. When one corporation gobbles up a smaller one because it wants the little one's technology, what often happens is that the management (who are mostly ignorant of the details of all the technology) wants to integrate the newcomers into their culture, and this means replacing the people and mindset that allowed the small company to develop their new technology, thus preventing further development. Even if some managers are aware of this problem, it can be hard to resist the natural pressure towards social integration.
It could be interesting to see if google could successfully incorporate such things as a search tool that deals with a DNA database. Would the result be to "improve" it with software that treats DNA sequences as a kind of spoken human language (without spaces, as in Chinese and Japanese writing)? Would this actually add any useful features, or is DNA's syntax so utterly different from English that the attempt would simply damage the search tool?
There are already some well-known problems with using google to locate things that are written in a highly technical English.
I notice the comment "If you just want to do e-mail and the Web, you pay $40 a month for the service". When I'm online, I use ssh a lot, and from the phrasing of that and other things in the article, I'm guessing that they block ssh. After all, ssh isn't email (SMTP, etc.) or web (HTTP).
Anyone know whether this is true? Do they actively block protocols? How about VoIP? (Not that I'd use it much, but if they do block VoIP, that's an admission that this isn't internet service, it's restricted to a set of protocols chosen by Verizon.
So how much extra would I have to pay to use ssh? How about VPN? Does ICMP work or is it extra?
For that matter, when they say "the Web", is this read-only access, or can I run my own web server? I would like to be able to run a web server from my laptop when I'm away from home, but most ISPs block this, including Verizon. So is it allowed with this gadget?
IBM makes laptops? They don't make Thinkpads, and never have. Those are manufactured by Lenovo, and are now sold by Lenovo.
Of course, it can be hard to keep track of who's manufacturing what, who's installing software, and who's doing the selling (under which brand names.)
I read what I thought was a bit of humor a while ago, which explained that IBM no longer actually makes anything that's concerned with computers. They just live off the billions of $$$ they get annually from their patent portfolio and from royalties for the use of their logo. At least I thought this was humor, but what do I know?
I don't remember ever being asked to show title to a car to anyone doing repairs or selling parts. Here in the US, the general advice is that you keep the title in a safe place where it can't be easily stolen. In particular, don't keep it in the car, because then a car thief could easily forge a sales document and they'd legally own it. (You keep the state registration paper in the car for when a cop stops you; the registrations papers aren't the title deed.)
Anyway, I'll try to remember Alienware as a company to stay away from.
These tissue types can only last hundreds of thousands of years, tops. So... either it's fake, or there's some unknown preservation process at work here, or -
Yeah, this is why there's so little known of the actual tissues of critters that old. But it's really an example of the "long tail" statistical phenomenon. Proteins, DNA, etc usually disappear pretty quickly, but there's no sharp cutoff age at which all samples instantly disintegrate into their constituent atoms. The decay is an exponential process, and no matter what age you pick, there's a small nonzero probability that there are fossils that old, until you get back to an age when there were no "tissues" on Earth. A very few fossils have been found that contain proteins that date to tens of millions of years. The story a couple of years ago about such a T. Rex fossil was an example that got lots of attention, mostly because it's such a popular dinosaur. But there aren't many people studying such fossils, because we haven't found very many of them.
The T. Rex tissues survived because they were inside intact bones buried in a place that has been dry for some 70 million years. The overlying material was never heavy enough to crack the bones, and the internal humidity never got high enough for any embedded bacterial spores to come to life. This is highly unlikely, but in a few places it has happened. Nobody knows whether we'll find more, though. It's possible that we've found the only such fossils that exist on the planet. Or there may be more buried in Montana, where both of these fossils were found. That area has been dry for a rather long time.
People are also considering the possibility of finding some very old frozen fossils under the Antarctic ice. But if they exist, they're in places that are sorta hard to get at. And the researchers want to be extra careful, because they expect that there will also be living spores (and maybe seeds) there, too. They don't want anyone doing the digging until they can be certain that the samples won't be contaminated by surface bacteria. But the digging (or more likely drilling) will probably be tried within the next decade or two.
Wow! Human DNA contains only 20 amino acids. (Actually, there is a 21st, but it's extremely rare.) I wonder what the Hadrosaur was doing with so many of them.
It sounds like our world really lost a lot at the K-T impact event.
(And isn't it wonderful how ambiguous the English language can be, especially in the hands of journalists.;-)
Both Principia Mathematica books, by Isaac Newton and Bertand Russell, have been important to the development of science. These were math books, of course, and as such weren't actually science. But since mathematics is fundamental to understanding most sciences, they should be in any list of Classic Books of Science. We might also note that Russell chose the same title because his intent was to replace Newton's great work with something even more important. Many people think he succeeded, in the same sense that Einstein succeeded in replacing Newtonian physics. But of course this doesn't detract at all from the importance of Newton's works. (Just take care to avoid his theological writings.;-)
Statistics is a branch of mathematics that's very important to science, but a quick attempt to learn the "classic" books in the correct sense of that term didn't work very well. Google just returned lots of ads for books that use "classic" in the marketing sense of "current best seller that has had more than one edition". Does anyone have references to truly classic statistics texts that have been important to scientists in the past?
As they also claim Microsoft Windows is Posix compliant! It is simply to be able to tic a "mandated" requirement in some government procurement, not as something one would actually use or deploy.
Back when the POSIX standard was under development, one of the things that the standards committee did was to put out a call for details of a hypothetical implementation called "Weirdnix". This was defined as an implementation that followed the wording of the spec and passed compliance tests, but did so in a way that prevented interoperability with other implementations. The point, of course, was to debug the spec and decrease the chances that someone like Microsoft would product a valid implementation that didn't play nice with the others.
Some time later, when Microsoft produced their POSIX library and passed government compliance tests, there were a lot of comments to the effect that Microsoft had actually implemented Weirdnix for real. Most simple 10- or 20-line test programs would seem to work correctly, but hardly any real applications would port between MS POSIX and others' POSIXes. Also, POSIX programs on MS Windows often didn't work well with programs that used the MS libraries.
(And I wonder what the pseudo-Greek or pseudo-Latin plural of POSIX would really be?;-)
The standards community has long understood that major vendors always have a motive to sabotage standards when possible. This is the origin of such ancient metaphors as "thumb on the scale", and the jokes about a "fishermen's ruler".
Suggesting that this is accidental is just silly. As pointed out in TFA, doing some of the standard wrong is more difficult than doing it right. The default assumption should be that the MS developers knew what they were doing, and they were probably doing it on orders from their managers. They have historically had a lot of success with customers who treat the MS version as "standard", and if it differs from the published standard, then the published standard isn't "standard".
(Actually, this attitude isn't uncommon in the US. Consider how many US companies sell both "Standard" and "Metric" versions of their products, where "Standard" means the US variant on the British Imperial measurements, not the published ISO standards - or even the actual British standards. To my knowledge, no company has ever been punished in the US for this blatant misuse of the term "standard". So Microsoft's approach to standards fits right in with what Americans have long been accustomed to.)
As we feared: he was trying to be an open source. Now he's closed.
Heh. Deserving of the "funny" mod, and like a lot of humor, based on a lot of truth.
Back in the years either side of 1990, I worked on a number of projects building software packages on workstations, and mostly I worked on Suns. One of the fun parts was all the people asking why we'd go with such an expensive machine, when you could get cheaper workstations with the same capabilities. But what happened repeatedly was: The guys working on the cheaper machines would be busy tracking down a bug, and the evidence led them into the system (libraries or OS). They'd ask the support people, and the answer would be "We can't tell you; that's proprietary."
Meanwhile, we Sun geeks would handle such things by asking questions on the various relevant mailing lists and newsgroups. More often than not, answers would be posted by Sun engineers, who were encouraged to follow the forums. Fairly often, a Sun guy would simply offer to send us the relevant source code as an explanation. Often, they'd explain that it was too much code to just post on a list, but anyone who asked would get the code in their email. Sun wasn't officially open-source, but they'd figured out that helping developers understand the innards was good for business. And it was, because inevitably we'd have stuff running on the Sun boxes long before the teams working on the "proprietary" systems. Having working deliverables is always better than not having them, even if they're more expensive.
But during the 1990s, this situation slowly disappeared. Sun slowly took their stuff proprietary, and developing significant software started running up against the brick wall of trade secrecy. But systems like linux and the *BSDs came along to replace Sun, giving us the access to internals that Sun no longer allowed.
Note that part of Sun's early success was that the code was sent "warts and all". Sun engineers were quite open about what it could do and what it couldn't. It was common to get warnings from the Sun guys that certain parts hadn't been tested much, and we should expect bugs. This sort of thing horrifies marketing people, but it was what made us trust Sun and choose their platform for our packages. And again, when Sun drifted toward not being as open about their stuff's problems, that told us developers that we could no longer trust them for the information we needed to get our stuff working. So we moved to the open systems where people post criticisms and bug reports openly.
It doesn't surprise me that Sun has lost its original position as a high-quality deveopment platform. It doesn't surprise me that they wouldn't want their people publishing criticisms of Sun's products. But those things long ago moved Sun from the "friendly and honest little company" category into the "untrustworthy big business" category in my mind and in the minds of a lot of developers.
Someday the same thing will happen with the major linux distros, and we'll respond the same way. Not with a big fuss, but by just quietly walking away and adopting other systems.
If you click the "help" button, it says "This report lists the market share of the top operating systems in use for browsing (not servers)."
And right there is the common clue that tells you how they're generating bogus statistics: the phrase "market share" means that they are only counting things that are purchased. A very high precentage of linux users get their copy via free downloads, and these systems aren't counted as part of the "market". In fact, since free linux systems are often installed on machines that were puchased with MS Windows (due to the difficulty of gettin the hardware in any other form), a significant fraction of running linux systems are counted by the marketeers as Windows systems. I have two linux systems on the shelf next to my desk, and one had Windows installed when it was delivered, so it's counted as a Windows sale. The other was ordered without an OS (which a local shop will do if you ask), so it's probably not counted as a sale of either Windows or linux - unless the vendor reported it to MS as a Windows box to avoid the usual harassment that happens if they openly sell just the bare hardware.
Similarly, I've seen listings of web servers ranked by sales, and Microsoft's IIS was the clear leader. Apache sometimes isn't even on such lists, since hardly anyone actually pays for it. There's no "market" for apache, since anyone can download it, install it, spend a few minutes tweaking the httpd.conf file, and use it without ever getting involved in any software market transactions.
In general, it's a good idea to be extremely skeptical of any figures derived from a "market". This is especially true when the numbers seem to support something you like. Such numbers almost always come from someone trying to sell you something, and the statistics are part of their marketing pitch.
Anyway, to use an obvious transport analogy, if you're looking for a good heavy-duty truck or small airplane, would you care about statistics saying that some kind of automobile is much more popular? Even if the numbers are accurate, they aren't too meaningful if you need the capabilities of a truck or airplane. The idea that some single kind of computer system is "the best" is equally silly. It depends on what you need it to do. If you think that all computer OSs are interchangeable, you simply don't understand what an OS is.
2 girls 1 cup was such a letdown... It was so fake looking as you can clearly tell it was just chocolate soft serve.
Yeah, I thought so, too. But if you view it as humor aimed at young children (or parrots or dogs or monkeys maybe), it's a lot funnier. I remember a similar case when my daughter was maybe four, and I took her into a restroom somewhere. When she was done, she looked into the toilet, and I remarked "Look at all that shit!" She replied with "That's not shit; it's peanut butter!" I was proud of her. Of course, she's an adult now, so her offensive humor is a lot more sophisticated. But that was pretty good for a preschool kid. And lots of preschool teachers make friends with their kids by engaging in the same sort of humor. Peewee Herman made a film career of it. Seeing it on youtube should surprise no-one.
Slashdot Mirror
Bo-o-o-o-gus.
This "study" didn't measure browser speed at all. It compared only the speeds of the javascripts that the browsers use. TFA says so fairly clearly.
If you're making heavy use of sites that are mostly javascript, this is a useful study. For the rest of us, it's yet another case of measuring a tiny corner of what is claimed, and then asserting that this measures the whole thing.
Using similar reasoning, we can imagine an oceanographer measuring the parts of the ocean along the beaches where most people are found, and concluding that the oceans average about 2 meters deep. (There's gotta be a good auto analogy here, too.)
As someone else has pointed out, most "power users" of browsers mostly disable java and javascript (and Active-X and any other misfeature that lets strangers run code on their machines). They may use NoScript with FF and enable JS for selected sites. Or they may simply copy the links to another browser such as opera or safari when they want to use JS. So to them, firefox and mozilla may well be the fastest browsers, since they permit easy selective disabling of all scripting features.
And we should also note that the time to render most web pages is mostly the download time. If due to network delays it takes 23 seconds to download a page, and browser X renders it in .001 sec while browser Y renders it in .01 sec, there's no practical meaning to a claim that Y renders 10 times faster than X. If the page takes 23.001 sec to render in X, and 23.01 sec in Y, few people will be able to reliably tell you which is faster.
If this were announced as a comparison of various JS interpret speeds, I'd take it seriously. But claiming that it's about browser speed pretty well discredits the authors (and the editor who wrote the summary).
Funny, I don't see it refreshing at all. I get a pop-down saying:
This has username and password input widgets, plus Cancel and OK buttons. It doesn't seem to have any way of setting up an account, so I tried a couple of dummy loging, hit OK -- and got back exactly the same pop-down windowlet. So I hit Cancel -- and got the same pop-down windowlet.
I notice that the button to close the window is greyed out ...
I'll check them occasionally and see if anything works. At least they're not soaking up all my cpu while they wait for me.
If the NSA (No Such Agency) is in charge, it'll be the same as having no security oversight at all. They naturally keep everything secret, so if they want to tell you to do something, you won't have the security clearance to read the order or any of its details.
Yes, they can write secret orders, not show them to you, and then prosecute you for not obeying them. But this has been true for around a decade now, so it won't be anything new.
Anyway, the main area where security is important is in the corporate world's handling of its comprehensive information about all of us. And in the modern US, agencies of the government don't give orders to corporations; the corporations give orders to the government. So corporate databases will continue to be as insecure as always, which doesn't really matter because the information is always for sale to the highest bidder, secure or not. Security really means that the information can't be read by anyone who hasn't paid for it, y'know.
If there are any changes, the most likely are that the NSA will be forced to adopt corporate-style "security" measures such as 4-digit PINs or password rules so complex that you have to write your passwords down and carry them in your wallet. And they'll routinely leave entire databases in laptops inside parked cars. This will be by policy, not accident. It'll result in more funny news stories; we'll mostly laugh and go about our lives.
I'd add a ;-), but I'm not sure that this actually qualifies as humor ...
(I'm sure that Jon Stewart and Steven Colbert will explain it much better than I can.)
Well, I've tried that; it only works if the connect() returns in the calling process. With the bug I wrote about, the problem is that this doesn't happen, even with a nonblocking socket. Also, an ALARM is supposed to interrupt a system call and force a failure return, but this doesn't happen, either.
I've documented it on FreeBSD, NetBSD, Solaris and several flavors of linux. It's been a couple of years since I've worked on a project where I could test this, so I can't tell you the releases.
It may not be a single bug, as similar "hangs, unkillable" problems have been reported in all sorts of systems since the dawn of "time sharing" computer systems back in the 1960s. They're usually good classroom studies for subtle kernel bugs, since diagnosing the problem requires a more-or-less complete path analysis of the code to find a "can't get there from here" scenario, where "there" is the code that returns to the process and "here" is the state in some system dumps.
A big part of the problem is that, in my experience, when the problem happens, it's typically after several million connect() calls by the same process. Imagine a search bot, for example, which might have to read millions or billions of URLs, typically with a connect() per URL (though this can be optimized in the obvious way).
Most people treat "one in a million" as an idiomatic way of saying "extremely rare". But there are 86400 seconds in a day, so if you call connect() as few as 12 times per second, you hit a million in a bit under a day. (23 hours, 8 minutes, 53 seconds, actually. ;-)
But in my experience, the bug isn't usually reproducible. During testing, I have had the process log all its connect() calls, so I could tell who it was trying to connect to. Usually restarting it for that address will result in success, which doesn't tell you much. I've managed to reproduce the problem maybe a total of a couple of dozen times, and the only thing I find in common is that the remote system has report that it was running some versions of Windows and IIS. But the sample size is far too small to bash Microsoft for this (other than the fun of doing so ;-). And most Windows+IIS systems connect just fine. I've seen the problem when the remote system was various unixoid OSs, but that has never been reproducible. And in any case, the culprit could easily be some intermediary bridge or router doing something "intelligent" with the connection attempt.
OTOH, the problem long predates any reports of ISPs like Comcast intentionally screwing with connections in order to cause problems for one end or both, which only go back a few years (to my knowledge). Before that, it almost had to be some incompatibility between the two ends of the connection.
I've worked on any number of projects in which we created a new network API, usually with UI tools to match. Of course, our package always used on sockets as the lower-level "internal" basis.
It's called "layering". Some network programmers have learned that it's a useful approach.
So far, I've never seen a different networking UI that's easier to program (or debug) than sockets. I keep reading articles like this on the topic, but I'm still looking for one that's better for the job. It's possible that the Berkeley people found the best low-level approach.
However, one thing that they missed that has been sorely needed on a number of projects is a "timeout" parameter to the connect() call. Time and again I've seen cases where an app hangs inside a call to connect() and never returns. Typically setting an alarm won't interrupt the call, and sometimes even a "kill -9" won't kill the process. Sometimes only a reboot will get rid of the zombie process. If the OS no longer gives your process any cpu time, it doesn't matter what clever code you have in it to diagnose problems. This seems to happen under unknown conditions on all OSs (though I haven't actually tested it on all releases of all OSs, so I could be wrong).
But this is an implementation detail of what's basically a fairly sound design.
Copyright is an arbitrary, artificial construct "designed" (I use the word generously) to encourage creation.
Actually, if you dig into the history of copyright, you'll find that this wasn't at all true. The earliest copyrights were given by monarchs to publishers of texts such as the Bible. They clearly didn't do this to encourage the creation of more bibles. They did it to limit the production so the masses couldn't get and learn to read such texts on their own, but had to rely on their betters to read and interpret the holy texts. They also did it to collect the royalties (note the root of that word) from the publishers who were permitted to do the publishing.
The idea that copyright is to encourage creation is a bit of PR created by the publishers to justify their monopoly. But from the beginning of copyright, it was a total lie. And can anyone find a study (with valid scientific methodology) that supports the idea that copyright every has actually resulted in the creation of more literature? Yes, we can find lots of claims that this is true. But has anyone presented actual evidence that it works?
There's a competing hypothesis (which probably hasn't been well-tested either): Creative people like to create things, and will do so even if not rewarded. Copyright was developed by the people who distribute the creations to customers, as a way of creating a monopoly market in creative works, to the financial benefit of the distributors. This is widely believed by many of those creative people, mostly the ones who aren't making a living from their creation because the contracts they have to sign give most of the income to the publishers and distributors.
People still buy books and newspapers because the portability and physical experience is better than a notebook or even a Kindle.
Books perhaps, but at least here in the US we're hearing a lot of analyses of the rapid decline of the newspaper business.
Part of this seems to be due to the fact that the conventional newspaper isn't in fact a very convenient format for readers. The pages are too big and the material is too flimsy. You can hold most books in one hand as you read, with your cup of coffee or whatever in the other hand, but this doesn't work at all with a newspaper. Anywhere but at home on your kitchen table, it's a very difficult format to read easily.
Something that comes up occasionally in the discussions of the newspaper industry's woes is the suggestion that "Nobody ever bought a newspaper because they want piles of low-quality, cheaply-printed paper in their house". People don't want the paper; they want the news. News is mostly a throwaway commodity, so it has to be done cheaply or people won't buy. But the resulting product has never been very good physically.
A funny thing about the discussions I've heard or read: When someone points out the above problem, it is usually ignored. That is, the others don't switch to talking about the "news industry"; they continue to talk about "newspapers", missing the point entirely that nobody actually wants the paper. We're mostly happy to be rid of the stuff. But the pundits seem to be willfully ignoring this point, and continue to talk as if the paper were the important product. As long as the news(paper) companies keep insisting that it's the paper that's important, they don't stand much of a chance of changing their business model over to electronic news distribution.
A few news companies have caught onto this, but most of them don't seem to understand. So their hard-copy news business is dying, and others that learn how to switch to the new, cheaper distribution medium will be the ones who survive. They just need to learn how to handle the advertising part, since so far that seems to be how they make their money (just as with the old newspapers).
First we need some laws on the subject that were not distorted from 200 years ago to fit the internet.
One of the general problems is that as soon as a computer is introduced to a subject area, all precedent is forgotten, chants of "That's different!" are heard repeatedly, and we humans must relearn every social lesson that we so laboriously worked out over the centuries.
Free speech? We're using computers; it isn't "speech", it's email or texting or ...
Freedom of the press? We're using computers, not presses; we're producing blogs and RSS feeds.
Fair use? We're using computers, which do copying for even the most trivial operations, so we have to throw the idea out and look for something else.
It'll take us a long time to figure out how to transfer all the rights our ancestors fought and died for to the brave new computerized world. Because none of those lessons are relevant when you use a computer to do something.
So your company doesn't want to hear about bug reports, UI problems, security issues, etc.
How dumb can a company's management be?
On numerous occasions, I've bought something partly because the vendor was open about problems. If a company pretends that there are no problems with its products, I suspect that they're suppressing news about something really awful.
(Note that I've avoided any Microsoft bashing. ;-)
Wait ... Cory Doctorow thinks he's obscure? If he's obscure, what writer is well known?
Is there any SF reader who doesn't know his name?
it's important to remember that "monopoly" when used here doesn't mean 100% of the market, but (like MS) enough of the market that it might as well be 100%, or at least large enough that they can exercise anti-competitive behavior.
One of the clichés in economics texts is the "5-50" rule of thumb saying that a "market" acts like a monopoly if 5 or fewer companies get 50% or more of the sales.
Of course, like any rule of thumb, this is basically "economics for dummies", because the reality is that there's a continuum of actual behaviors. Some big companies are run by people with ethics and a long-term view (though they tend to disappear with time). Some markets have sufficient delivery problems that they act like local monopolies even with a hundred companies.
But the point of such things is to debunk the traditional even sillier idea that you only have a "monopoly" if there is just one company. This is called the Etymological Fallacy, the idea that the meaning of a word is defined by the meanings of its parts in the original (long-dead) languages. It's popular with the people who like the idea of unbridled, lassez-faire capitalism. But that's not how economists or most other people use the term in English. In the real world, there are such things as "gentlemen's agreements" that produce monopoly markets even when there are several sellers.
It's fairly clear to nearly everyone that the US retail computer business is a monopoly market, although there are two companies supplying the core hardware and two companies providing the OSs. A small fraction of the population can actually name the second software supplier (though very few can name either hardware supplier). But it's been that way here for a few decades now, so we don't expect that we'll see an actual free market in computer retailing in our lifetime. It's interesting reading about efforts in other parts of the world to do something about the monopoly. It'll be even more interesting if they actually succeed, and make it possible for smaller startups to actually do business.
Don't worry, google will buy it soon enough
That's indeed a danger, and made worse by the media's tendency to compare every little search project with google, even when there's little or no overlap in what they do.
OTOH, I've been involved in a few of the thousand or more projects to build highly-specialized search tools that "understand" the data used for specialized purposes. So far, none of them has to my knowledge been approached by anyone from google. In a few cases, the opposite has happened, as the people in a project get to wondering whether any of google's expertise could be useful to them. Again, as far as I can tell, the answer is "No", but I've only seen a few small corners of the field.
It does seem fairly clear that we're better off with a flock of independent projects to solve the world's search problems. There's a well-known problem with human organizations, summarized by the old saying that if your only tool is a hammer, everything looks like a nail. When one corporation gobbles up a smaller one because it wants the little one's technology, what often happens is that the management (who are mostly ignorant of the details of all the technology) wants to integrate the newcomers into their culture, and this means replacing the people and mindset that allowed the small company to develop their new technology, thus preventing further development. Even if some managers are aware of this problem, it can be hard to resist the natural pressure towards social integration.
It could be interesting to see if google could successfully incorporate such things as a search tool that deals with a DNA database. Would the result be to "improve" it with software that treats DNA sequences as a kind of spoken human language (without spaces, as in Chinese and Japanese writing)? Would this actually add any useful features, or is DNA's syntax so utterly different from English that the attempt would simply damage the search tool?
There are already some well-known problems with using google to locate things that are written in a highly technical English.
I notice the comment "If you just want to do e-mail and the Web, you pay $40 a month for the service". When I'm online, I use ssh a lot, and from the phrasing of that and other things in the article, I'm guessing that they block ssh. After all, ssh isn't email (SMTP, etc.) or web (HTTP).
Anyone know whether this is true? Do they actively block protocols? How about VoIP? (Not that I'd use it much, but if they do block VoIP, that's an admission that this isn't internet service, it's restricted to a set of protocols chosen by Verizon.
So how much extra would I have to pay to use ssh? How about VPN? Does ICMP work or is it extra?
For that matter, when they say "the Web", is this read-only access, or can I run my own web server? I would like to be able to run a web server from my laptop when I'm away from home, but most ISPs block this, including Verizon. So is it allowed with this gadget?
IBM make some of the best laptops around.
IBM makes laptops? They don't make Thinkpads, and never have. Those are manufactured by Lenovo, and are now sold by Lenovo.
Of course, it can be hard to keep track of who's manufacturing what, who's installing software, and who's doing the selling (under which brand names.)
I read what I thought was a bit of humor a while ago, which explained that IBM no longer actually makes anything that's concerned with computers. They just live off the billions of $$$ they get annually from their patent portfolio and from royalties for the use of their logo. At least I thought this was humor, but what do I know?
I don't remember ever being asked to show title to a car to anyone doing repairs or selling parts. Here in the US, the general advice is that you keep the title in a safe place where it can't be easily stolen. In particular, don't keep it in the car, because then a car thief could easily forge a sales document and they'd legally own it. (You keep the state registration paper in the car for when a cop stops you; the registrations papers aren't the title deed.)
Anyway, I'll try to remember Alienware as a company to stay away from.
I don't see the problem here. I have a lot of free software on my web site, and it all comes with a money-back guarantee if it doesn't work for you.
Looks like we've gone from throwing money at the problem to throwing computers at the problem.
And as a result, the schools (and many of the kids) will have two problems.
These tissue types can only last hundreds of thousands of years, tops. So ... either it's fake, or there's some unknown preservation process at work here, or -
Yeah, this is why there's so little known of the actual tissues of critters that old. But it's really an example of the "long tail" statistical phenomenon. Proteins, DNA, etc usually disappear pretty quickly, but there's no sharp cutoff age at which all samples instantly disintegrate into their constituent atoms. The decay is an exponential process, and no matter what age you pick, there's a small nonzero probability that there are fossils that old, until you get back to an age when there were no "tissues" on Earth. A very few fossils have been found that contain proteins that date to tens of millions of years. The story a couple of years ago about such a T. Rex fossil was an example that got lots of attention, mostly because it's such a popular dinosaur. But there aren't many people studying such fossils, because we haven't found very many of them.
The T. Rex tissues survived because they were inside intact bones buried in a place that has been dry for some 70 million years. The overlying material was never heavy enough to crack the bones, and the internal humidity never got high enough for any embedded bacterial spores to come to life. This is highly unlikely, but in a few places it has happened. Nobody knows whether we'll find more, though. It's possible that we've found the only such fossils that exist on the planet. Or there may be more buried in Montana, where both of these fossils were found. That area has been dry for a rather long time.
People are also considering the possibility of finding some very old frozen fossils under the Antarctic ice. But if they exist, they're in places that are sorta hard to get at. And the researchers want to be extra careful, because they expect that there will also be living spores (and maybe seeds) there, too. They don't want anyone doing the digging until they can be certain that the samples won't be contaminated by surface bacteria. But the digging (or more likely drilling) will probably be tried within the next decade or two.
Wow! Human DNA contains only 20 amino acids. (Actually, there is a 21st, but it's extremely rare.) I wonder what the Hadrosaur was doing with so many of them.
It sounds like our world really lost a lot at the K-T impact event.
(And isn't it wonderful how ambiguous the English language can be, especially in the hands of journalists. ;-)
Both Principia Mathematica books, by Isaac Newton and Bertand Russell, have been important to the development of science. These were math books, of course, and as such weren't actually science. But since mathematics is fundamental to understanding most sciences, they should be in any list of Classic Books of Science. We might also note that Russell chose the same title because his intent was to replace Newton's great work with something even more important. Many people think he succeeded, in the same sense that Einstein succeeded in replacing Newtonian physics. But of course this doesn't detract at all from the importance of Newton's works. (Just take care to avoid his theological writings. ;-)
Statistics is a branch of mathematics that's very important to science, but a quick attempt to learn the "classic" books in the correct sense of that term didn't work very well. Google just returned lots of ads for books that use "classic" in the marketing sense of "current best seller that has had more than one edition". Does anyone have references to truly classic statistics texts that have been important to scientists in the past?
Why the fuck would you link to wikipedia for the definition of a word?
Because you weren't aware of the existence of wiktionary.com?
As they also claim Microsoft Windows is Posix compliant! It is simply to be able to tic a "mandated" requirement in some government procurement, not as something one would actually use or deploy.
Back when the POSIX standard was under development, one of the things that the standards committee did was to put out a call for details of a hypothetical implementation called "Weirdnix". This was defined as an implementation that followed the wording of the spec and passed compliance tests, but did so in a way that prevented interoperability with other implementations. The point, of course, was to debug the spec and decrease the chances that someone like Microsoft would product a valid implementation that didn't play nice with the others.
Some time later, when Microsoft produced their POSIX library and passed government compliance tests, there were a lot of comments to the effect that Microsoft had actually implemented Weirdnix for real. Most simple 10- or 20-line test programs would seem to work correctly, but hardly any real applications would port between MS POSIX and others' POSIXes. Also, POSIX programs on MS Windows often didn't work well with programs that used the MS libraries.
(And I wonder what the pseudo-Greek or pseudo-Latin plural of POSIX would really be? ;-)
The standards community has long understood that major vendors always have a motive to sabotage standards when possible. This is the origin of such ancient metaphors as "thumb on the scale", and the jokes about a "fishermen's ruler".
Suggesting that this is accidental is just silly. As pointed out in TFA, doing some of the standard wrong is more difficult than doing it right. The default assumption should be that the MS developers knew what they were doing, and they were probably doing it on orders from their managers. They have historically had a lot of success with customers who treat the MS version as "standard", and if it differs from the published standard, then the published standard isn't "standard".
(Actually, this attitude isn't uncommon in the US. Consider how many US companies sell both "Standard" and "Metric" versions of their products, where "Standard" means the US variant on the British Imperial measurements, not the published ISO standards - or even the actual British standards. To my knowledge, no company has ever been punished in the US for this blatant misuse of the term "standard". So Microsoft's approach to standards fits right in with what Americans have long been accustomed to.)
As we feared: he was trying to be an open source.
Now he's closed.
Heh. Deserving of the "funny" mod, and like a lot of humor, based on a lot of truth.
Back in the years either side of 1990, I worked on a number of projects building software packages on workstations, and mostly I worked on Suns. One of the fun parts was all the people asking why we'd go with such an expensive machine, when you could get cheaper workstations with the same capabilities. But what happened repeatedly was: The guys working on the cheaper machines would be busy tracking down a bug, and the evidence led them into the system (libraries or OS). They'd ask the support people, and the answer would be "We can't tell you; that's proprietary."
Meanwhile, we Sun geeks would handle such things by asking questions on the various relevant mailing lists and newsgroups. More often than not, answers would be posted by Sun engineers, who were encouraged to follow the forums. Fairly often, a Sun guy would simply offer to send us the relevant source code as an explanation. Often, they'd explain that it was too much code to just post on a list, but anyone who asked would get the code in their email. Sun wasn't officially open-source, but they'd figured out that helping developers understand the innards was good for business. And it was, because inevitably we'd have stuff running on the Sun boxes long before the teams working on the "proprietary" systems. Having working deliverables is always better than not having them, even if they're more expensive.
But during the 1990s, this situation slowly disappeared. Sun slowly took their stuff proprietary, and developing significant software started running up against the brick wall of trade secrecy. But systems like linux and the *BSDs came along to replace Sun, giving us the access to internals that Sun no longer allowed.
Note that part of Sun's early success was that the code was sent "warts and all". Sun engineers were quite open about what it could do and what it couldn't. It was common to get warnings from the Sun guys that certain parts hadn't been tested much, and we should expect bugs. This sort of thing horrifies marketing people, but it was what made us trust Sun and choose their platform for our packages. And again, when Sun drifted toward not being as open about their stuff's problems, that told us developers that we could no longer trust them for the information we needed to get our stuff working. So we moved to the open systems where people post criticisms and bug reports openly.
It doesn't surprise me that Sun has lost its original position as a high-quality deveopment platform. It doesn't surprise me that they wouldn't want their people publishing criticisms of Sun's products. But those things long ago moved Sun from the "friendly and honest little company" category into the "untrustworthy big business" category in my mind and in the minds of a lot of developers.
Someday the same thing will happen with the major linux distros, and we'll respond the same way. Not with a big fuss, but by just quietly walking away and adopting other systems.
If you click the "help" button, it says "This report lists the market share of the top operating systems in use for browsing (not servers)."
And right there is the common clue that tells you how they're generating bogus statistics: the phrase "market share" means that they are only counting things that are purchased. A very high precentage of linux users get their copy via free downloads, and these systems aren't counted as part of the "market". In fact, since free linux systems are often installed on machines that were puchased with MS Windows (due to the difficulty of gettin the hardware in any other form), a significant fraction of running linux systems are counted by the marketeers as Windows systems. I have two linux systems on the shelf next to my desk, and one had Windows installed when it was delivered, so it's counted as a Windows sale. The other was ordered without an OS (which a local shop will do if you ask), so it's probably not counted as a sale of either Windows or linux - unless the vendor reported it to MS as a Windows box to avoid the usual harassment that happens if they openly sell just the bare hardware.
Similarly, I've seen listings of web servers ranked by sales, and Microsoft's IIS was the clear leader. Apache sometimes isn't even on such lists, since hardly anyone actually pays for it. There's no "market" for apache, since anyone can download it, install it, spend a few minutes tweaking the httpd.conf file, and use it without ever getting involved in any software market transactions.
In general, it's a good idea to be extremely skeptical of any figures derived from a "market". This is especially true when the numbers seem to support something you like. Such numbers almost always come from someone trying to sell you something, and the statistics are part of their marketing pitch.
Anyway, to use an obvious transport analogy, if you're looking for a good heavy-duty truck or small airplane, would you care about statistics saying that some kind of automobile is much more popular? Even if the numbers are accurate, they aren't too meaningful if you need the capabilities of a truck or airplane. The idea that some single kind of computer system is "the best" is equally silly. It depends on what you need it to do. If you think that all computer OSs are interchangeable, you simply don't understand what an OS is.
2 girls 1 cup was such a letdown ... It was so fake looking as you can clearly tell it was just chocolate soft serve.
Yeah, I thought so, too. But if you view it as humor aimed at young children (or parrots or dogs or monkeys maybe), it's a lot funnier. I remember a similar case when my daughter was maybe four, and I took her into a restroom somewhere. When she was done, she looked into the toilet, and I remarked "Look at all that shit!" She replied with "That's not shit; it's peanut butter!" I was proud of her. Of course, she's an adult now, so her offensive humor is a lot more sophisticated. But that was pretty good for a preschool kid. And lots of preschool teachers make friends with their kids by engaging in the same sort of humor. Peewee Herman made a film career of it. Seeing it on youtube should surprise no-one.