Yes, but math is all about finding patterns and generalizing, so when mathematicians study the representations of numbers, they don't just study the properties of *one* representation. They study what the properties of entire categories of representations (e.g., standard place-value representations with a natural base) have in common. So yeah, I would expect the formula to be at least generalized enough to handle base n, where n is any natural number. I wouldn't be surprised if it also works for non-integer bases, or even complex bases.
It probably does only work for place-value representations, though. It would be pretty hard to meaningfully generalize such a pattern beyond that.
> Why would anyone in the market have had an interest in loaning to high-risk individuals > if it wasn't for the "affordable housing" and "homeownership for all" agenda
You have a point, but it's also true that the risk was underestimated. Some people who *ought* to have known better had themselves convinced that "property values always rise over time", and so they underestimated the risk associated with subprime mortgages. If it were indeed absolutely true that property values without exception rise over time, the risk would then be minimal, because the bank could always foreclose and sell the house and get its money back. There's some overhead associated with doing this, so the bank would typically prefer that everyone make their payments as agreed, but on the whole mortgages would be a fairly safe investment, IF it were really true that property values always rise. But, of course, while property values tend upward over the long term, sometimes they don't rise over the short term. In that case, the bank gets stuck holding a house that's worth less than the mortgage. That's the risk the investors forgot to properly account for.
When any significant number of people become convinced that the value of Item X always rises (whatever Item X may be), this causes a Item X to become overvalued (i.e., a bubble), a situation that invariably leads to what economists call a "correction", i.e., sooner or later the bubble pops and prices fall back toward (or, in some cases, temporarily below) what they really should have been; thus, the prices are corrected to better reflect the item's actual value. The people who were doing the investing in the collateralized obligations *ought* to have known this. I mean, it's Econ 101 stuff, not something you can afford to be ignorant of if you're in charge of investing risk-averse funds. But property values had been increasing for so long, people started to believe it would continue indefinitely. That was an error in judgment.
The investors *also* should have known better because the investments in question were offering interest rates well above what would be considered normal for a safe investment. Well, the numbers were reasonable if you took them out of context, but the prime rate was low at the time, and so of course interest rates were running low across the board. Yet, these subprime mortgage investments were offering rates double, triple, even quadruple the prime rate, and that's *not* normal for a safe investment. The investors' brains should have been screaming "TOO GOOD TO BE TRUE", but they allowed themselves to be lulled into a false sense of security.
Incidentally, there's nothing per se wrong with investing in subprime mortgages, provided you know what you're doing and have a sufficiently diverse portfolio to tolerate the associated risk. But treating a risky investment as if it were a safe investment, that's dangerous.
Actually, the whole fiasco really boils down to this:
Things (economically speaking) had been so good throughout the nineties, a lot of people had started to actually believe that the economy was always going to keep endlessly spiraling upward like that with no setbacks and no corrections. This is, of course, not possible in the real world, but people had started to believe it anyway, including people who ought to have known better, and so they started to behave as if that were the case.
When people believe that values will always rise, they overestimate the values of things, and that's where bubbles come from. And where there's a bubble, sooner or later it's going to pop. (Economists call this a "correction", because the market is correcting somebody's mistaken, overinflated concept of the value of whatever it is.)
Also, when people believe that values will always rise, they underestimate risk. For example, people who believed that property values would always rise underestimated the risk associated with investing in sub-prime mortgages, and so funds were invested in sub-prime CDOs that should not have been, because they could not tolerate the level of risk involved. (A "sub-prime" loan is a loan to someone who arguably can't really afford to borrow that much money. Typically they have less than 20% for a downpayment, which puts the value of the principal amount dangerously close to the value of the collateral. There are other factors as well.) If property values had continued to only rise all the time, it would have been okay, because in that case when somebody defaults the bank could simply foreclose, sell the house, and get its money back. But the reality is that values don't always rise all the time, and when housing values fell (it happens sometimes, over the short term), the risk inherent in sub-prime loans came home to roost, and the people who had assumed that this could never happen (don't you know, property values always rise?) discovered that they had made an error in judgment.
And they really should have known better. Nothing that promises to pay out more than double the prime rate comes without some risk. But they had sold themselves on the line, "values always go up", because values had *been* going up for a while, and so it seemed like that was the pattern. Like I said, things were so good in the eighties and especially in the nineties, people got to thinking that was normal and expected it to keep on going like that. But that kind of endless upward spiral is not sustainable over the long term.
> There is a reason for the lack of direct entry of Unicode.
I think there's more than one reason. Slashdot is, by design, an English-language-only forum geared primarily toward Americans. Foreign characters aren't really *desirable* on slashdot. They would just make the signal-to-noise ratio worse, for no significant benefit.
Yeah, I admit, I have been annoyed by it at times. I wanted to put the little squiggly mark over the n in the word jalepeno once, and several times I've wanted to post untransliterated Greek words for one reason or another. But the current system is fair: I can't post Greek letters, and you can't post Cyrillic letters, and the Perl6 people can't post weird double-angle-bracket thingies, and the Frenchies can't put little marks on the bottoms of the letter C, and everybody else can't post kana or arabic or devanagari or whatever. For an all-English-language-all-the-time discussion forum like slashdot, all of this isn't really a net loss, IMO.
With the ASCII-only filter, the weirdest writing we have to put up with is 1337 |-|4X0r 5P33|<, and that's bad enough.
> But maybe it's time to replace "duke nukem forever" with "slasdot unicode support" in our posts.
My understanding was that slashdot filters out anything other than ASCII not out of laziness, but because it specifically wants to be an all-English-language-all-the-time forum, and foreign characters would just be illegible and/or distracting to most readers. If slasdot allowed non-ASCII characters, people would use them, and that would make the signal-to-noise ratio even lower than it already is.
Europeans would use diacritical marks that the rest of us don't really know how to read: dots and funny squiggles over the letters and eight different kinds of accent marks and slashes through the letters and all sorts of nonsense. But that's not the worst of it. Language nerds would sling around ancient Greek words _in the Greek alphabet_ (I've caught myself trying to do that in slashdot posts on three separate occasions, even though I theoretically know better), so if you don't know Greek you wouldn't be able to recognize them, even though there are copious English cognates in most cases. Trolls would fill entire posts with cyrillic characters, and the mods who don't read cyrillic (i.e., most of them) wouldn't really be completely sure whether these posts were really trolls or legitimate messages that happen to be in Russian or Ukranian. You just know the anime fanboys would sling kana around in their otherwise English-language posts, for names of characters if nothing else. Wouldn't you rather they be forced to transliterate them into romaji, so that everyone else can follow? And so on and so forth.
Limiting slashdot to ASCII on the one hand makes it less literate in the academic sense, and references to foreign words and names are less authentic, but on the other hand it guarantees a certain minimum level of legibility: everyone in the entire English-speaking world can at least read the individual _letters_ in every single comment. The weirdest writing system you have to figure out is ISO-1337-1, which frankly doesn't exactly require a great deal of study.
I wouldn't make this argument for every site, but for slashdot I think it does make sense.
Unicode adoption is in progress. A lot of software supports it now that didn't five years ago.
But unicode is a large and complex beast, and not entirely a stationary target, and supporting it properly and completely is difficult, and it isn't always the first priority for software developers, even in the open-source world. Unaccountably, sometimes they just want to get the basic functionality of the software working first, figuring that they'll add unicode support later. Subsequently, they discover that unicode support is difficult to add later, because it breaks a number of common assumptions that programmers can and often do make if they only need to support ASCII or individual 8-bit codepages (like ISO-8859-1, for instance). Since it's difficult, it tends to be the object of significant procrastination, and then when it is undertaken, it takes a long time to be completed.
Progress IS being made, but gradually. I figure about 25% of the email I get now uses Unicode (mostly UTF8) for the character encoding. But the percentage increases every year.
I'm just glad we finally got to the point where pretty much all computer systems at least support ASCII at the bare minimum. I believe the last EBCDIC software was decommissioned, along with the ancient IBM mainframe it ran on, several years ago. And there was much rejoicing.
> If you really want to use ruminants to mow for you, sheep are a much better choice than goats.
That depends whether you want to have any grass left at all when they're done. The goats might leave a little behind. The sheep typically won't. They're like the nuclear weapons of anti-grass warfare.
Why on earth would you give antibiotics for a cold? Colds are viral, last I checked.
But yeah, the prescription system is on the whole something society is better off with than without.
And the other poster is completely out of line with the phrase "doctor monopoly". There are millions of doctors, and while small groups of them may have business relationships (shared practices and so forth), they're still not even really an oligopoly, much less a monopoly. (I suppose there may be certain communities where the doctors have all banded together to form a cartel of sorts, but if so that is a local issue in that community.)
If anything the small number of medical insurance outfits, and the unwarranted leverage that they exert, is a larger problem, but the last thing I want is the government trying to straighten that out. The federal government (indirectly) caused that mess in the first place through a fantastically stupid attempt to control inflation by the terribly misguided expedient of simply making it illegal, through wage and price controls -- an approach that has never once worked as intended in the history of the universe and usually causes significant undesirable side effects. Sure enough, employers unable to recruit the workers they needed at the wages the government was willing to let them pay found a loophole and started paying the workers in something other than money. Unfortunately, rather than simple goods that could be bartered off, they chose services like medical insurance for the additional compensation, so to this day your employer gets to decide which insurance company will set the policies regarding what medical procedures you can have done and, worse, which doctors and hospitals you can do business with. Unwarranted government meddling in the economy caused this.
If the allegations are true and verifiable, drug regulatory agencies (e.g., the FDA in the US) should fine them so hard as to risk wiping them clean out of existence, just to send a clear signal to everyone who might be watching that schenanighans of this sort will NOT be tolerated. (And if they don't want to pay the fine, their other option would be to no longer sell drugs in the country. They'll come up with the money somehow.) Heck, I would think the other drug companies would be in favor of nipping this in the bud, since they already have *enough* trouble with general mistrust of their products even without junk like this taking place.
They won't do that, though. They'll probably fine them, but it'll be an amount that they can pay and recover from and stay in business.
Anyone who says "99% secure is 100% insecure" clearly does not understand even the most basic principles of security. I quote from Bruce Schneier:
"Unbreakable", "absolute", "unforgeable", and "impenetrable" are all words that make no sense when discussing security. If you hear them, you can be sure you're listening to someone who doesn't understand security or is trying to hoodwink you. Good security systems are designed in anticipation of possible failure. -- _Beyond Fear_, pp.57-58.
99% is just about the best security you're ever going to get, and with decent defense in depth strategies, plus detection and response plans, it's very manageable.
> To prevent the car from being stolen, a batter quick disconnect is nice. > This is especially try if you have a car where the batter is mounted in > the boot and is clean as opposed to under the bonnet. There is a light > fused line to all power to electronics, put you pull the big steal pin > out of the positive lead.
Incidentally, what language is this? It bears such a strong resemblance to English, I suspect I could learn it with only a few years of study.
> I don't want to sound troll-ish but it's likely that people who have > auto-update set to "download-and-install-automatically" aren't the more savvy set
I would have said the reverse, since that's not the default setting, and home users never change the defaults. Therefore, people who have it set up that way are most likely people who did so deliberately, presumably for security reasons.
> Firefox makes IE 8 your default browser in Linux? That's kinda odd.
Every distribution is a little bit different. In Gentoo, you choose your own default browser. In Fedora, Firefox asks you the first time it runs if you want it to be the default browser, and the default is yes. In Debian, Firefox is not available in the package list. I don't happen to know what distribution has Firefox making IE8 the default. Probably one of the ones I've never used, like Lycoris or Linspire or something along those lines.
> It amazes me how when firefox has a new version, everyone downloads it > with a warm and fuzzy feeling that it is going to be an improvement.
Actually, this is no longer quite true. It's still true for the _majority_ of Firefox users, but if you've been paying any attention at all you know that the Mozilla folks have been fretting and stewing no end about how a lot of people (a small percentage, but enough to get their attention) are still using Firefox 2.
> However, whenever IE has a new version, people are so reluctant to > download it that MS now has to force the public to upgrade.
Different user base. Most (almost all, until recently) Firefox users are technically inclined and like to try out new software. These are what we in the IT industry call "power users". Most IE users are what we call "end users", and as a general rule they do not want the software on their computer to change, EVER, because it might mean they might have to learn something new or different, and that is, to their way of thinking, only slightly less undesirable than having surgery without anesthesia.
If Firefox market share continues to grow, a larger and larger percentage of users will not upgrade in a timely fashion.
If you ask web developers what we think about IE8, we're all very pleased about it, because it means we will eventually be able to stop working around IE7's inadequacies. (How soon depends on what percentage of end users we're willing to hork off, but having IE8 on Auto Update helps, a lot.)
I'd be less pleased if this purported change in default-browser policy turns out to be true, but one rant about it on a blog doesn't exactly have me frothing at the mouth because, you know, people rant about stuff like this and then turn out to be getting their facts slightly wrong about six times out of seven.
> IE7 and 8 do not natively support xhtml, they treat the > document as SGML and apply SGML rules to it. This means > that namespaces, MathML, etc. cannot be supported by > that browser,
Actually, it's theoretically possible to still support those things without fully embracing all the other XML rules. It won't make the purists happy, but it can be done.
> and that it will not fail on invalid content as an xml parser should.
I consider completely failing on all invalid content to be an undesirable characteristic in a web browser and contrary to the basic principle of graceful degradation that the web is built on. If a browser sees an element it does not recognize, it should just treat it as a generic element (comparable to span) and go on. This allows webmasters to go ahead and use new features before every web browser that's still in use supports them (which can take a decade or more).
A web browser is not a validator. That isn't its purpose. Web developers need to know how to use a validator, yes, but the web browser shouldn't be it.
> Nope... Many corporations built their intranet around > IE6 and changing browser will break it. Rather than > spend buckets of money revamping their intranet, they > are just more likely to keep going with IE6...
That's fine, as long as they only have to access their own intranet sites and nothing else.
But given the difficulty of keeping multiple versions of IE around (it's much harder than with, say, Firefox), not very many web developers are still testing with IE6 at this point. IE7 will be headed the same direction in the near future, at which point browsing the web with IE6 will become increasingly an exercise in frustration, much like trying to browse with IE5 is now.
With IE5, it's a royal pain just to get to a search engine, look for a more recent version of IE, download, and install it. Really, try it. Trying to actually use it as your regular browser is just futile. In another year or two, IE6 will be like that. Unusable.
We just did a major website revamp here, so I tested in IE7 because it's still current. We'll stick with this design revision for a while now, and just change the information on individual pages as necessary. But the next time I do a design reorg, I *won't* still be testing in IE7.
My recommendation for "our intranet app is IE6-only" sites is to deploy an alternate web browser (perhaps Firefox) for regular web browsing, and keep IE6 around just for your intranet only. Take the IE6 shortcut off the desktop and replace it with a shortcut that directly opens your intranet in IE6, plus another that opens Firefox for actual web browsing.
But, you know, you can do what you want. You can try to browse the web in IE4 if you want. Just don't expect to actually, you know, *work*.
> Most large companies do not use torrents because they are a little complex for most users
Yeah, but you can put that under the hood and the user never has to know the details. You give them a normal http link to download an executable "installer" which downloads the rest of the thing using whatever protocol you like. A few years ago most large software companies were doing this to distribute large freely-downloadable stuff. The protocol under the hood obviously wasn't BitTorrent at that time, but the software could do things like resume an interrupted download (which web browsers of the day couldn't do) and was simpler for the user than working with a real ftp client.
However, for someone like Microsoft distributing something like a Windows 7 beta build, you're still going to want to spread the load across multiple servers on multiple continents and so on and so forth, which, yeah, is sort of what services like Akamai are all about. If Microsoft doesn't want to contract out like that, they could probably just do something similar with their own resources. I'm pretty sure they're big enough to be able to handle that.
> CLIs are great IF you know the command to launch it.
Exactly. Investing a moment or so, once, to learn the name of the command saves you 5-10 seconds every single time you want to do the thing from then on; your initial investment of time is paid back with interest in a few days. And that's for programs that you just want to launch (no significant options, no real interaction).
But a lot of people aren't just ignorant: they are *willfully* ignorant and *determined* to remain that way. These people are so averse to ever *learning* anything, they will retype whole paragraphs to avoid learning to copy and paste, and they'll do it all with 1-2 fingers, every time, for their whole lives, rather than learn to actually type. They *know* that there are faster and better ways, because they have met people who zip right along on the computer, typing 30+ words a minute, magically moving paragraphs around with a couple of keystrokes, and just generally doing everything in less than a quarter of the time. They *know* this is possible, but they DO NOT WANT to do it themselves, because it would require learning.
Took them long enough. I've been systematically turning AutoRun off on every Windows system I touch since 1998.
> Why wasn't this the default to begin with?
Same reason automatically downloading and executing any ActiveX code a website asks for was originally the default. Same reason automatically launching executable attachments when you preview a message in Outlook Express was originally the default. Same reason saving downloaded executables from the web on the desktop of all places was originally the default. Because Microsoft does not think about security issues until *after* the problem is exploited in the wild, and they also don't consider usability issues until *after* it becomes obvious that users are having a problem.
> There's no good reason to automatically run anything on media like hard > disks or flash drives. It's an obvious virus vector.
There's no really compellingly good reason to automatically run anything on a CD-ROM disk either, as far as that goes. Any user who can't figure out how to browse the disk in Windows Explorer and double-click on SETUP.EXE probably should not be installing software without help, anyway. And don't say "but the installation instructions indicate it will automatically launch when you insert the disk". The instructions were written like that *because* that's the default behavior, not the other way around.
Slashdot Mirror
Yes, but math is all about finding patterns and generalizing, so when mathematicians study the representations of numbers, they don't just study the properties of *one* representation. They study what the properties of entire categories of representations (e.g., standard place-value representations with a natural base) have in common. So yeah, I would expect the formula to be at least generalized enough to handle base n, where n is any natural number. I wouldn't be surprised if it also works for non-integer bases, or even complex bases.
It probably does only work for place-value representations, though. It would be pretty hard to meaningfully generalize such a pattern beyond that.
No, it means Microsoft views Google as a bigger threat than open source.
> Why would anyone in the market have had an interest in loaning to high-risk individuals
> if it wasn't for the "affordable housing" and "homeownership for all" agenda
You have a point, but it's also true that the risk was underestimated. Some people who *ought* to have known better had themselves convinced that "property values always rise over time", and so they underestimated the risk associated with subprime mortgages. If it were indeed absolutely true that property values without exception rise over time, the risk would then be minimal, because the bank could always foreclose and sell the house and get its money back. There's some overhead associated with doing this, so the bank would typically prefer that everyone make their payments as agreed, but on the whole mortgages would be a fairly safe investment, IF it were really true that property values always rise. But, of course, while property values tend upward over the long term, sometimes they don't rise over the short term. In that case, the bank gets stuck holding a house that's worth less than the mortgage. That's the risk the investors forgot to properly account for.
When any significant number of people become convinced that the value of Item X always rises (whatever Item X may be), this causes a Item X to become overvalued (i.e., a bubble), a situation that invariably leads to what economists call a "correction", i.e., sooner or later the bubble pops and prices fall back toward (or, in some cases, temporarily below) what they really should have been; thus, the prices are corrected to better reflect the item's actual value. The people who were doing the investing in the collateralized obligations *ought* to have known this. I mean, it's Econ 101 stuff, not something you can afford to be ignorant of if you're in charge of investing risk-averse funds. But property values had been increasing for so long, people started to believe it would continue indefinitely. That was an error in judgment.
The investors *also* should have known better because the investments in question were offering interest rates well above what would be considered normal for a safe investment. Well, the numbers were reasonable if you took them out of context, but the prime rate was low at the time, and so of course interest rates were running low across the board. Yet, these subprime mortgage investments were offering rates double, triple, even quadruple the prime rate, and that's *not* normal for a safe investment. The investors' brains should have been screaming "TOO GOOD TO BE TRUE", but they allowed themselves to be lulled into a false sense of security.
Incidentally, there's nothing per se wrong with investing in subprime mortgages, provided you know what you're doing and have a sufficiently diverse portfolio to tolerate the associated risk. But treating a risky investment as if it were a safe investment, that's dangerous.
Actually, the whole fiasco really boils down to this:
Things (economically speaking) had been so good throughout the nineties, a lot of people had started to actually believe that the economy was always going to keep endlessly spiraling upward like that with no setbacks and no corrections. This is, of course, not possible in the real world, but people had started to believe it anyway, including people who ought to have known better, and so they started to behave as if that were the case.
When people believe that values will always rise, they overestimate the values of things, and that's where bubbles come from. And where there's a bubble, sooner or later it's going to pop. (Economists call this a "correction", because the market is correcting somebody's mistaken, overinflated concept of the value of whatever it is.)
Also, when people believe that values will always rise, they underestimate risk. For example, people who believed that property values would always rise underestimated the risk associated with investing in sub-prime mortgages, and so funds were invested in sub-prime CDOs that should not have been, because they could not tolerate the level of risk involved. (A "sub-prime" loan is a loan to someone who arguably can't really afford to borrow that much money. Typically they have less than 20% for a downpayment, which puts the value of the principal amount dangerously close to the value of the collateral. There are other factors as well.) If property values had continued to only rise all the time, it would have been okay, because in that case when somebody defaults the bank could simply foreclose, sell the house, and get its money back. But the reality is that values don't always rise all the time, and when housing values fell (it happens sometimes, over the short term), the risk inherent in sub-prime loans came home to roost, and the people who had assumed that this could never happen (don't you know, property values always rise?) discovered that they had made an error in judgment.
And they really should have known better. Nothing that promises to pay out more than double the prime rate comes without some risk. But they had sold themselves on the line, "values always go up", because values had *been* going up for a while, and so it seemed like that was the pattern. Like I said, things were so good in the eighties and especially in the nineties, people got to thinking that was normal and expected it to keep on going like that. But that kind of endless upward spiral is not sustainable over the long term.
> There is a reason for the lack of direct entry of Unicode.
I think there's more than one reason. Slashdot is, by design, an English-language-only forum geared primarily toward Americans. Foreign characters aren't really *desirable* on slashdot. They would just make the signal-to-noise ratio worse, for no significant benefit.
Yeah, I admit, I have been annoyed by it at times. I wanted to put the little squiggly mark over the n in the word jalepeno once, and several times I've wanted to post untransliterated Greek words for one reason or another. But the current system is fair: I can't post Greek letters, and you can't post Cyrillic letters, and the Perl6 people can't post weird double-angle-bracket thingies, and the Frenchies can't put little marks on the bottoms of the letter C, and everybody else can't post kana or arabic or devanagari or whatever. For an all-English-language-all-the-time discussion forum like slashdot, all of this isn't really a net loss, IMO.
With the ASCII-only filter, the weirdest writing we have to put up with is 1337 |-|4X0r 5P33|<, and that's bad enough.
> But maybe it's time to replace "duke nukem forever" with "slasdot unicode support" in our posts.
My understanding was that slashdot filters out anything other than ASCII not out of laziness, but because it specifically wants to be an all-English-language-all-the-time forum, and foreign characters would just be illegible and/or distracting to most readers. If slasdot allowed non-ASCII characters, people would use them, and that would make the signal-to-noise ratio even lower than it already is.
Europeans would use diacritical marks that the rest of us don't really know how to read: dots and funny squiggles over the letters and eight different kinds of accent marks and slashes through the letters and all sorts of nonsense. But that's not the worst of it. Language nerds would sling around ancient Greek words _in the Greek alphabet_ (I've caught myself trying to do that in slashdot posts on three separate occasions, even though I theoretically know better), so if you don't know Greek you wouldn't be able to recognize them, even though there are copious English cognates in most cases. Trolls would fill entire posts with cyrillic characters, and the mods who don't read cyrillic (i.e., most of them) wouldn't really be completely sure whether these posts were really trolls or legitimate messages that happen to be in Russian or Ukranian. You just know the anime fanboys would sling kana around in their otherwise English-language posts, for names of characters if nothing else. Wouldn't you rather they be forced to transliterate them into romaji, so that everyone else can follow? And so on and so forth.
Limiting slashdot to ASCII on the one hand makes it less literate in the academic sense, and references to foreign words and names are less authentic, but on the other hand it guarantees a certain minimum level of legibility: everyone in the entire English-speaking world can at least read the individual _letters_ in every single comment. The weirdest writing system you have to figure out is ISO-1337-1, which frankly doesn't exactly require a great deal of study.
I wouldn't make this argument for every site, but for slashdot I think it does make sense.
> I guess unicode hasn't been adopted yet.
Unicode adoption is in progress. A lot of software supports it now that didn't five years ago.
But unicode is a large and complex beast, and not entirely a stationary target, and supporting it properly and completely is difficult, and it isn't always the first priority for software developers, even in the open-source world. Unaccountably, sometimes they just want to get the basic functionality of the software working first, figuring that they'll add unicode support later. Subsequently, they discover that unicode support is difficult to add later, because it breaks a number of common assumptions that programmers can and often do make if they only need to support ASCII or individual 8-bit codepages (like ISO-8859-1, for instance). Since it's difficult, it tends to be the object of significant procrastination, and then when it is undertaken, it takes a long time to be completed.
Progress IS being made, but gradually. I figure about 25% of the email I get now uses Unicode (mostly UTF8) for the character encoding. But the percentage increases every year.
I'm just glad we finally got to the point where pretty much all computer systems at least support ASCII at the bare minimum. I believe the last EBCDIC software was decommissioned, along with the ancient IBM mainframe it ran on, several years ago. And there was much rejoicing.
> The phrasing "gone missing" makes him sound like he's from somewhere in the United Kingdom
Not necessarily. We say "gone missing" in Ohio.
But yes, employees of the victim organization are the first people you investigate.
> If you really want to use ruminants to mow for you, sheep are a much better choice than goats.
That depends whether you want to have any grass left at all when they're done. The goats might leave a little behind. The sheep typically won't. They're like the nuclear weapons of anti-grass warfare.
Why on earth would you give antibiotics for a cold? Colds are viral, last I checked.
But yeah, the prescription system is on the whole something society is better off with than without.
And the other poster is completely out of line with the phrase "doctor monopoly". There are millions of doctors, and while small groups of them may have business relationships (shared practices and so forth), they're still not even really an oligopoly, much less a monopoly. (I suppose there may be certain communities where the doctors have all banded together to form a cartel of sorts, but if so that is a local issue in that community.)
If anything the small number of medical insurance outfits, and the unwarranted leverage that they exert, is a larger problem, but the last thing I want is the government trying to straighten that out. The federal government (indirectly) caused that mess in the first place through a fantastically stupid attempt to control inflation by the terribly misguided expedient of simply making it illegal, through wage and price controls -- an approach that has never once worked as intended in the history of the universe and usually causes significant undesirable side effects. Sure enough, employers unable to recruit the workers they needed at the wages the government was willing to let them pay found a loophole and started paying the workers in something other than money. Unfortunately, rather than simple goods that could be bartered off, they chose services like medical insurance for the additional compensation, so to this day your employer gets to decide which insurance company will set the policies regarding what medical procedures you can have done and, worse, which doctors and hospitals you can do business with. Unwarranted government meddling in the economy caused this.
If the allegations are true and verifiable, drug regulatory agencies (e.g., the FDA in the US) should fine them so hard as to risk wiping them clean out of existence, just to send a clear signal to everyone who might be watching that schenanighans of this sort will NOT be tolerated. (And if they don't want to pay the fine, their other option would be to no longer sell drugs in the country. They'll come up with the money somehow.) Heck, I would think the other drug companies would be in favor of nipping this in the bud, since they already have *enough* trouble with general mistrust of their products even without junk like this taking place.
They won't do that, though. They'll probably fine them, but it'll be an amount that they can pay and recover from and stay in business.
Anyone who says "99% secure is 100% insecure" clearly does not understand even the most basic principles of security. I quote from Bruce Schneier:
"Unbreakable", "absolute", "unforgeable", and "impenetrable" are all words that make no sense when discussing security. If you hear them, you can be sure you're listening to someone who doesn't understand security or is trying to hoodwink you. Good security systems are designed in anticipation of possible failure. -- _Beyond Fear_, pp.57-58.
99% is just about the best security you're ever going to get, and with decent defense in depth strategies, plus detection and response plans, it's very manageable.
> To prevent the car from being stolen, a batter quick disconnect is nice.
> This is especially try if you have a car where the batter is mounted in
> the boot and is clean as opposed to under the bonnet. There is a light
> fused line to all power to electronics, put you pull the big steal pin
> out of the positive lead.
Incidentally, what language is this? It bears such a strong resemblance to English, I suspect I could learn it with only a few years of study.
> Is it not a bit early to be deciding which browsers are more secure than IE8?
Well, I don't have actual data on that yet, but I know which way I'd lay odds.
> I don't want to sound troll-ish but it's likely that people who have
> auto-update set to "download-and-install-automatically" aren't the more savvy set
I would have said the reverse, since that's not the default setting, and home users never change the defaults. Therefore, people who have it set up that way are most likely people who did so deliberately, presumably for security reasons.
> Firefox makes IE 8 your default browser in Linux? That's kinda odd.
Every distribution is a little bit different. In Gentoo, you choose your own default browser. In Fedora, Firefox asks you the first time it runs if you want it to be the default browser, and the default is yes. In Debian, Firefox is not available in the package list. I don't happen to know what distribution has Firefox making IE8 the default. Probably one of the ones I've never used, like Lycoris or Linspire or something along those lines.
HTH.HAND.
That website bothers me. It frightens me because I'm worried some idiot might think it's serious or, worse, actually agree with its deranged logic.
> It amazes me how when firefox has a new version, everyone downloads it
> with a warm and fuzzy feeling that it is going to be an improvement.
Actually, this is no longer quite true. It's still true for the _majority_ of Firefox users, but if you've been paying any attention at all you know that the Mozilla folks have been fretting and stewing no end about how a lot of people (a small percentage, but enough to get their attention) are still using Firefox 2.
> However, whenever IE has a new version, people are so reluctant to
> download it that MS now has to force the public to upgrade.
Different user base. Most (almost all, until recently) Firefox users are technically inclined and like to try out new software. These are what we in the IT industry call "power users". Most IE users are what we call "end users", and as a general rule they do not want the software on their computer to change, EVER, because it might mean they might have to learn something new or different, and that is, to their way of thinking, only slightly less undesirable than having surgery without anesthesia.
If Firefox market share continues to grow, a larger and larger percentage of users will not upgrade in a timely fashion.
If you ask web developers what we think about IE8, we're all very pleased about it, because it means we will eventually be able to stop working around IE7's inadequacies. (How soon depends on what percentage of end users we're willing to hork off, but having IE8 on Auto Update helps, a lot.)
I'd be less pleased if this purported change in default-browser policy turns out to be true, but one rant about it on a blog doesn't exactly have me frothing at the mouth because, you know, people rant about stuff like this and then turn out to be getting their facts slightly wrong about six times out of seven.
> Sometime around 2004 ... most of our potential customers were running Netscape 4
Yeah? In August of 2005, we finally stopped using a particularly mission-critical piece of software that ran on DOS 6. No fooling.
> IE7 and 8 do not natively support xhtml, they treat the
> document as SGML and apply SGML rules to it. This means
> that namespaces, MathML, etc. cannot be supported by
> that browser,
Actually, it's theoretically possible to still support those things without fully embracing all the other XML rules. It won't make the purists happy, but it can be done.
> and that it will not fail on invalid content as an xml parser should.
I consider completely failing on all invalid content to be an undesirable characteristic in a web browser and contrary to the basic principle of graceful degradation that the web is built on. If a browser sees an element it does not recognize, it should just treat it as a generic element (comparable to span) and go on. This allows webmasters to go ahead and use new features before every web browser that's still in use supports them (which can take a decade or more).
A web browser is not a validator. That isn't its purpose. Web developers need to know how to use a validator, yes, but the web browser shouldn't be it.
> Nope... Many corporations built their intranet around
> IE6 and changing browser will break it. Rather than
> spend buckets of money revamping their intranet, they
> are just more likely to keep going with IE6...
That's fine, as long as they only have to access their own intranet sites and nothing else.
But given the difficulty of keeping multiple versions of IE around (it's much harder than with, say, Firefox), not very many web developers are still testing with IE6 at this point. IE7 will be headed the same direction in the near future, at which point browsing the web with IE6 will become increasingly an exercise in frustration, much like trying to browse with IE5 is now.
With IE5, it's a royal pain just to get to a search engine, look for a more recent version of IE, download, and install it. Really, try it. Trying to actually use it as your regular browser is just futile. In another year or two, IE6 will be like that. Unusable.
We just did a major website revamp here, so I tested in IE7 because it's still current. We'll stick with this design revision for a while now, and just change the information on individual pages as necessary. But the next time I do a design reorg, I *won't* still be testing in IE7.
My recommendation for "our intranet app is IE6-only" sites is to deploy an alternate web browser (perhaps Firefox) for regular web browsing, and keep IE6 around just for your intranet only. Take the IE6 shortcut off the desktop and replace it with a shortcut that directly opens your intranet in IE6, plus another that opens Firefox for actual web browsing.
But, you know, you can do what you want. You can try to browse the web in IE4 if you want. Just don't expect to actually, you know, *work*.
If the technical name is A(H1N1), running it through a reverse-1337 filter yeilds Achini. I say we call it the Achini Flu.
What? Why are you looking at me like that?
> Most large companies do not use torrents because they are a little complex for most users
Yeah, but you can put that under the hood and the user never has to know the details. You give them a normal http link to download an executable "installer" which downloads the rest of the thing using whatever protocol you like. A few years ago most large software companies were doing this to distribute large freely-downloadable stuff. The protocol under the hood obviously wasn't BitTorrent at that time, but the software could do things like resume an interrupted download (which web browsers of the day couldn't do) and was simpler for the user than working with a real ftp client.
However, for someone like Microsoft distributing something like a Windows 7 beta build, you're still going to want to spread the load across multiple servers on multiple continents and so on and so forth, which, yeah, is sort of what services like Akamai are all about. If Microsoft doesn't want to contract out like that, they could probably just do something similar with their own resources. I'm pretty sure they're big enough to be able to handle that.
> CLIs are great IF you know the command to launch it.
Exactly. Investing a moment or so, once, to learn the name of the command saves you 5-10 seconds every single time you want to do the thing from then on; your initial investment of time is paid back with interest in a few days. And that's for programs that you just want to launch (no significant options, no real interaction).
But a lot of people aren't just ignorant: they are *willfully* ignorant and *determined* to remain that way. These people are so averse to ever *learning* anything, they will retype whole paragraphs to avoid learning to copy and paste, and they'll do it all with 1-2 fingers, every time, for their whole lives, rather than learn to actually type. They *know* that there are faster and better ways, because they have met people who zip right along on the computer, typing 30+ words a minute, magically moving paragraphs around with a couple of keystrokes, and just generally doing everything in less than a quarter of the time. They *know* this is possible, but they DO NOT WANT to do it themselves, because it would require learning.
Took them long enough. I've been systematically turning AutoRun off on every Windows system I touch since 1998.
> Why wasn't this the default to begin with?
Same reason automatically downloading and executing any ActiveX code a website asks for was originally the default. Same reason automatically launching executable attachments when you preview a message in Outlook Express was originally the default. Same reason saving downloaded executables from the web on the desktop of all places was originally the default. Because Microsoft does not think about security issues until *after* the problem is exploited in the wild, and they also don't consider usability issues until *after* it becomes obvious that users are having a problem.
> There's no good reason to automatically run anything on media like hard
> disks or flash drives. It's an obvious virus vector.
There's no really compellingly good reason to automatically run anything on a CD-ROM disk either, as far as that goes. Any user who can't figure out how to browse the disk in Windows Explorer and double-click on SETUP.EXE probably should not be installing software without help, anyway. And don't say "but the installation instructions indicate it will automatically launch when you insert the disk". The instructions were written like that *because* that's the default behavior, not the other way around.