Hmmm... I looked at Thinkpads a few months ago, and couldn't find any with that resolution. This may have been because I couldn't find the actual numeric resolution on a lot of the models. It must be because I'm such a dummy that I can't easily follow the confused layout of the vendors' web sites...
(I was also looking for models that came with linux, and generally couldn't find that info in the vendors' sites, either.;-)
"This new technology is sponsored and funded by: Your friendly health and life insurance company, constantly finding new and innovative ways to make sure we never have to pay you a dime since 1666."
For the benefit of those who didn't hear the show, last week NPR had an interesting interview with Karen Tumulty on just this topic. It was about her brother's problems with his insurance company when he started having kidney problems. Part of the story was that the insurance investigators got access to his medical tests, and they found a slight anomaly in one test a few years earlier that the doctors didn't consider significant at the time, but which the insurance company said indicated the kidney disease. Since he was insured by a different policy when that test was done, it means the disease was a "pre-existing condition" for the current policy, so they didn't have to cover it.
This is rapidly becoming a serious problem in the US. The more detailed your medical monitoring is, the more likely that the insurers will collect your money, and they refuse coverage due to a precursor that showed up once far in the past.
Of course, in the long run, the health insurers will shoot themselves in their collective feet, as health insurance slowly ceases to cover anything, and it will become pointless to pay for insurance that doesn't cover anything. Then the only people with medical coverage will be those able to pay the million-dollar charges for treating minor ailments.
(We're not quite there yet, but it's easy enough to be a prophet in this subject area.;-)
It would then be up to me if I wanted to share this information with a doctor or others.
Not necessarily. Here in the US, if you want the possibility of medical treatment for most serious health problems, you have to have insurance. And you wouldn't be able to get insurance unless you gave the insurance salespeople access to your medical data.
Also, you might not be able to get anything but a minimum-wage job without giving your employer (and their insurance investigators) access to your data.
The poorer part of the population has been in this situation for some time, and the fraction of the population that it applies to is slowly increasing. Another decade, and it'll probably include most people like you.
Thanks for the reference. I'll look into that model, see whether linux works on it, etc.
I did notice that the first few google hits for places didn't actually give the screen resolution, just the 17" measurement and some superlatives. But one of the links to within Dell did give the 1900x1200 numbers.
But a lot of the scientific hypothesis we use right now, like the earth is round, seemed completely wacko once too.
Actually, the shape of the Earth is a bad thing to use as such an example. There's lots of historical evidence showing that people have understood the shape of the Earth for as long as we have historical records. Lots of textbooks describe the successful measurement of the Earth's size by those classical Greek guys, and similar things have been done in other cultures. Sailors on large bodies of water have always known that they were sailing on a big curved surface, from the way that things disappear from the bottom up as you sail away from them, and appear from the top down as you approach them.
The idea of a flat Earth historically has belonged to religious and political types that were insulated from real-world experience and could safely believe whatever nonsense they liked on such topics that didn't impact their everyday lives. But the guys down at the docks always knew better, because the curved surface they lived on was a significant part of their everyday lives, and they could see its shape.
It's curious especially that we still have this bizarre myth that Columbus was trying to prove something about the shape of the Earth. At the time, even the religious people accepted that the planet was spherical. The debate wasn't its shape; the debate was over its size. (And Columbus's trips didn't decide that.;-) This is quite well documented in the historical records.
For most purposes, it's the flat-earth theory that should be considered wacko, as most knowledgeable people would have laughed at it at any time in history. Only religious nuts and political leaders believed such nonsense in the face of the evidence that any dummy can pick up in a single reasonably long sailing trip.
Part of the problem is that our main religious texts came from people living out in the desert, which is a terrain where measuring the size and shape of the Earth is both unimportant and very difficult. Now if we could just free our society from the influence of dummies who take some just-so stories written down by a few desert shepherds and insist that they are eternal truths about the whole world.
I have wondered who besides Apple sells a 17" laptop with a screen that's 1900x1200 pixels or better. I haven't seen one. Of course, I could easily have missed it, because with most laptops, it's impressively difficult to discover the pixel count. They tell you the diagonal size, with no other numeric information. Even if you can dig out a "specs" page, often the pixel count isn't anywhwere to be seen.
Maybe they think that we're all too stupid to understand the concept of resolution.
It is clear the slashdot editors run Windows/Mac machines and never view their site on a Linux based machine or they would have at least addressed the Javascript problems.
Nah; they read with NoScript blocking the scripts, too, like most of us do.;-)
I use both this Mac and a linux machine, and I haven't seen that firefox or opera looks significantly different on either of them. Maybe there would be differences if I enabled scripts, but why would I do that?
(That wasn't being facetious. If anyone could give me a good motive to enable scripts, I might do it. But the few times I've experimented, I didn't see any improvement, and downloads were slower, so I turned scripts back off. So if you're allowing/.'s scripts, why are you doing that? Am I missing something useful?)
Every time I come here I wonder what kind of totally pointless misfeature I'll find next.
Hmmm... I don't see any obvious misfeatures in/.'s UI. Of course, that might be because I went to my account page and turned off nearly everything that can be turned off. I also have AdBlock and NoScript installed in the browser that I use for/. (firefox), and have/. scripts are totally blocked.
The only obvious thing wrong is that there's a lot of the window that's just white space. The main slashdot.org page has the rightmost 1/3 blank, so the window covers screen space that could be used for something. And the discussion pages have too-wide indentation, so there's a lot of white space on the left. Anyone know how to cut down on those wastes of screen space?
For a few weeks, I had scripts enabled for/., but I was noticing that refreshes were taking a lot longer than before. So I turned scripts back off, and now things load in 1-2 seconds, so I guess I won't try that experiment again for a while.
Anyway, I can read articles and discussions, I can reply to messages, I can moderate when I grudgingly decide to do my social duty, and I can even meta-moderate. Is there anything else/. can do for me that I don't know about and have blocked?
I do sorta miss my message counter, though I'd have to admit that it's not something that's actually useful for anything. But we geeks and nerds do like numbers, y'know.
Yeah, what's the point of "better tab handling" and a "niftier search bar" if the results look like crap because it can't render everything properly?
Ya gotta consider that to their intended audience, proper rendering is defined as "however IE6 (mis)renders a page". Yes, a few people have updated this to mean "however IE7 (mis)renders a page", but that doesn't affect the outcome much.
To most of their intended audience, FF isn't "standard" because on their screen it renders the MS-only sites differently than their current IE does.
IE8 is aimed at people who have no clue about w3c.org, and wouldn't like it if they ever happened to accidentally stumble across it.
Try plugging centigrade temperatures into the ideal gas law and lemme know how it goes.;)
Actually, the fun part is that there's a long history of people doing exactly that. The patent offices in the US and other countries have an ongoing problem of people attempting to patent perpetual-motion machines. In most cases, a "proof" that a particular gadget will work is produced by taking the standard equations and using Celsius/Centigrade numbers when temperature in "degrees" is needed. This seems to be something that a lot of physics cranks can't quite get right. They read that the Kelvin degree is the same as the Celsius degree, and that means that it doesn't matter which you use, right?
I don't even know how many ml there are in a tablespoon.
15.
There, now wasn't that easy? You can even treat the tablespoon as a metric term if you like. It's one of the "extended metric" units widely used in the US.
And there are 5 ml in a teaspoon. (In my experience, most Americans can't tell you how many teaspoons there are in a table spoon. That's one of many such questions that are good tests of how easy the Imperial system is to use. Most people flunk the tests pretty badly.;-)
This is probably mostly due to the fact that the components are no longer manufactured in the US. A lot of final assembly is still done here, but the pieces that actually have to be measured have mostly been manufactured in Asia for a couple of decades.
I've sometimes wondered whether part of the decline in the US auto industry was from the sizable fraction of the population that does a lot of the routine work on their own cars. Back in the 1970s, when Asian imports became common, a lot of Americans realized that "Hey, I don't need to keep two sets of auto tools any more. If I buy Japanese, I can throw out all those "imperial" tools (which the sales folks still insist on calling "standard";-) and just use standard (i.e., metric) tools where the arithmetic is so much simpler. The US auto companies did pick up on this and switch to mostly metric parts, but it took them a decade or so, and they'd lost the battle by then.
Of course, now the media is openly talking about the final death of the US auto companies...
Too many possibilities of Earth origin are rejected with the phrase "it seems unlikely", and there's no mention of the most obvious method by which the micro-organisms get there: random motion (OK, particle velocities in the atmosphere will not be truly random, but you'd still expect a few outliers with very high velocities.)
Also, the "output" of the Earth's atmosphere was measured back in the 1960s and 1970s (and perhaps earlier). I remember reading a couple of articles in the 70s about our planet's "cometary tail", which was described as basically produced by the same solar-wind pressure as a comet's tail, and contains particles ranging from single atoms up to the size of bacterial spores. It was predicted then that when we examined the outer planets in detail, we'd find bacterial life there similar to Earth's, because that's where it came from.
The authors also pointed out that most of this dust tail would escape the solar system. It's a reasonable inference that we've been spewing this dust out, bacterial spores included, for some 4 billion years, which is about 16 trips around the galaxy, and the Earth's dust tail has permeated the galaxy during that time. Of course, this would be true of any other planet in the galaxy that has evolved bacteria-sized life.
The only real question is whether the spores can survive for long periods in space. 30 years ago, nobody had collected any bacteria from the Earth's outer atmosphere or from its dust tail, so survival time in space was purely conjectural. Now we have some samples that appear to be able to survive in space, at least long enough to get far enough away that the UV flux is no longer a danger. Whether these critters can survive the millions of years between stars is still a question.
But the default assumption now should be that some of them can survive. Most will die, or sleep forever, but some will encounter another object that contains edible food. Out between the stars, there doesn't seem to be anything we know of that would destroy something that can survive the UV flux of our stratosphere.
So the panspermia conjecture has just got a serious boost. And we've reached the point where, as the old joke goes, we present the most important statement in any scientific research: Further research is necessary. We know of a mechanism, and we know of a few bacteria that can survive the mechanism. But the numbers are all vague and fuzzy, barely respectable in scientific circles.
It's time to start applying for the funding for exploratory robots that can examine the outer planets' atmospheres and assorted planets' and asteroids' surfaces looking for more bacteria. We should expect to find earthly spores out there, and it'll be a surprise if we don't. Can we find tiny critters that are sufficiently different that we can say definitively that they didn't come from Earth?
All predictions about fast adoption of linux because of it being cheaper have not come true partially because corporate service boys charged a healthy premiums on their linux 'loving' customers.
Oh, I dunno about that. A few months ago, I ordered the hardware for a new "desktop" system from a local computer assembler, and since I ordered it without the default Vista OS, I got a discount of a few hundred $$$. While talking about it with a rep over the phone just before delivery, he asked what I intended to install on it. I said "The latest Ubuntu release", and he said "We can install that for you, for no extra charge." I said "Huh?", and he said "Yeah; we've found that Ubuntu always installs quickly, with no problems at all. Give us an hour, and we can have it all set up for your." I told him "OK", and I got it with Ubuntu running just fine.
(Well, OK, there was a problem: They forgot to tell me the password that it wanted when I booted it. They were very apologetic about that. They were even more apologetic when I told them that, since they were closed when I got it home, I'd booted a handy knoppix briefly to mount the root partition and set the root password to something I knew.;-)
I do sorta suspect that they wanted to do it as a training exercise for their installer guys, as a response to a good number of customers wanting that system installed. But no matter; the fact is that a local system builder took the attitude that "The customer is always right", and wanted their people to be able to install whatever the customer wanted.
Anyway, this one company didn't charge a healthy premium on a linux-loving customer. They said "We can do that for you for no extra charge." And, needless to say, I told a number of other local friends about it, probably resulting in a few more sales.
... people seem to think throwing money at proprietary software will magically solve their problems...
This attitude is exemplified by the mantra "You get what you pay for". We've read that here in numerous/. discussions. Maybe it has already appeared in this discussion.
IBM and Microsoft are among the many companies that encourage this belief in their marketing, because it's such a good tool to make people spend money without doing the sensible price shopping and product comparisons that intelligent purchasing requires. Anyone who believe this mantra is their natural prey.
It's especially bizarre to hear or read "You get what you pay for" in a computer setting. The history of our industry is full of example of just the opposite. Especially with software, quality tends to be an inverse function of price. The high-priced stuff is mostly designed to be sold at a "management" level, i.e., to people who have no ability to compare the quality of various software products. It's especially unfortunate that this happens with school purchases, since this encourages the "monoculture of the worst" that we see all over. Then people buy the same stuff for home use because it's the only thing they can (barely) use without further training, and also because it's expensive and "You get what you pay for".
Of course, the FOSS crowd did sorta blow it by using the phrase "free software". Yes, they meant "free" as in freedom, not as in beer. But the other 99% of the population understands "free" as a euphemism for "cheap and shoddy", because it's what they've learned it means when salesmen use the term. "You get what you pay for", and if it's free, it must be crap, right?
Fighting this attitude is the biggest marketing problem for anyone selling cheap-but-good stuff. This is especially true for software, for which quality is nearly independent of the time and effort (i.e., the cost) of development, but people still believe that cheap stuff must be crap.
Maybe the solution is to sell expensive editions that only differ from the free stuff in the price (and maybe an expensive-looking splash page).
Every person I know says "March fourteenth two thousand nine", not "fourteenth March two thousand nine".
Typical "strawman" argument, disproof by using the wrong syntax.
I've often heard "the fourteenth of March, two thousand [and] nine." As a native speaker of English, I'd claim that that's every bit as natural as "March [the] fourteenth, two thousand [and] nine".
So dd/mm/yyyy should be as normal in English as the weird middle-endian form used by most Americans.
Myself, I use the ISO format (yyyy-mm-dd hh:mm_ss) whenever I think I can get away with it. That way, you minimize the computer code needed to handle the date. And nobody that I know of uses yyyy-dd-mm, so there isn't the usual annoying ambiguity that software just can't deal with.
The problem with the year-last formats is that when someone writes something like 4/9/2009, you generally have no way of knowing which order they're using and which date they're talking about. It's not just a problem with software; now that we have a single worldwide communication system, it's a problem in human communication, too. The only way to win that game is to not play it.
Let the Pope decide what a sacrament should be and keep any hint of sacrament out of what the government does.
Oh, so now you want to prevent non-Catholics from getting married, too?
If you dig further into the history, you'll find that there's a lot of precedent for that.
I've read a few explanations that one of the many reasons for the US's "separation of church and state" in the Bill of Rights was that the UK had traditionally not recognized marriages outside the established church. This was among the many reasons that various groups (Puritans, Quakers, Jews, etc.) had emigrated to the wilds of North America, so that they could set up local governments that would honor their right to such institutions as marriage (and land ownership, right to enter into contracts, right to persecute people not like them, etc.).
The people in the US claiming ownership of marriage and demanding that "people not like us" not be allowed to marry are carrying on an old tradition. Their predecessors had to grudgingly accept marriages among black people, but for a long time many states refused to permit mixed-race marriages. Then the Supreme Court said that this had to be permitted, but most states still refused to allow marriages in many other cases (e.g., same-sex marriages). Now a few states have permitted that, and the "only people like us" people are getting desperate to find people they can legally exclude.
A funny thing about the religious objectors to same-sex marriage is that my wife and I (;-) attended several same-sex weddings here in Massachusetts before the state legalized same-sex marriage. These weddings were in fact performed in churches (or a synagogue in one case). So it's obvious that some religious groups implemented same-sex (and mixed-race) marriage since long before such things became recognized by the state. The people claiming that this is something new have no knowledge of history. Various religious bodies have been doing "disapproved" marriages for as long as we have any records of such things, and some of those religious bodies have kept records going back centuries. They don't advertise it, of course, for obvious reasons. But a number of historians have found the records and written about the actual practices.
Yeah. Now if we could just make code that uses such things portable to systems like *BSD, Solaris, OS X, etc. It's a real pain trying to track down how you do it everywhere (except MS systems, of course;-). You have to write code that explores the system to discover which gimmicks they've included and whether they actually work. And you need copies of each kind of system for testing.
Just calling sync() or fsync() is a lot easier, though as others have observed, some systems (and some disk controllers) have found ways of subtly shooting down that approach.
Yeah; I suspected that that's what you meant. But you can get more humor out of a situation by answering an extreme interpretation of what someone just said...
Anyway, I've always sorta thought it was good if a few programs were calling sync() at a moderate rate. On a few systems, I've added a little background process that just sleeps and calls sync() every N minutes. This was usually in response to someone discovering that the system was holding unwritten data in memory for hours because the system load was too light to force flushing. It's always seemed to me that an OS should have a system call that sets the max time that a write buffer can exist before it's flushed. With all the never-used features that are bloating up all our OSs, you'd think they'd have included that one.
Or maybe they have, and they just haven't told us mere users.
... can't imagine many reasons for having hundreds of tiny config files for a single app...
Ooh, ooh, I know one!
Some years back, I stumbled into being the admin of one of what are by now probably several thousand specialized "search engines" for a certain kind of rather technical data whose nature isn't too important here. The "app" includes a search bot and a single-request download program, and both of them have to deal with what is now nearly 400 sites that have the data. Each site has its own web server, and I sometimes suspect that no two of them are running the same server. So there's a "cfg" directory that contains a file for each hostname, and the file contains config info for dealing with that host's web server.
The most important thing is getting the HTTP level right, because a lot of servers will reject requests if you ask for the wrong level. The default is "HTTP/1.1", but a significant number of the sites' servers require a different level. And this can't be set up at the first access to a host, because sometimes people upgrade or reconfigure their servers. So what the download routine does is look for the HTTP/* entry in a host's cfg file, and uses "HTTP/1.1" if it isn't found. Occasionally a download attempt fails, and the routine goes into a "determine HTTP level" mode, in which it tries a list of level numbers looking for one that works. If it succeeds, it updates the host's cfg file to tell later downloads what level to use.
Actually, there aren't "hundreds" of config files (yet). The number is only slightly over 100, because most of the servers work fine with the default settings. Most of the cfg files are tiny, because few servers have more than one or two settings that need to be specified. The biggest is 468 bytes; the rest are under 100. I expect that the number of files could reach 200 in a few more years.
I expect that others will be able to describe some more situations where they use a large flock of tiny config files for some app or set of closely-related apps.
The app updates a config file by building the new contents in memory, and overwriting the file with a single write() call. It doesn't bother with a flush() call, because such updates are very rare, and if the write fails, it just means that the "determine HTTP level" code will be done again the next time that host is accessed. In a decade or so of life, this has never happened. I can say this with some certainty, because such things get written to a global "special events" log file, so I've seen all the cases where an app updates a config file.
Name one scenario where fsync should be necessary other than:... 2) rewriting data back to an existing file... like for instance something a database might do.
Hmmm... The followups imply that some people have a misunderstanding of how the POSIX write() call works, and interpret the above as implying that when process A writes data to a file and process B reads the same spot in the file, B may get a jumbled mixture of the old and new data. But the POSIX spec for write() explicitly states that a write of up to PIPE_BUF bytes is atomic. This was generally true of unix kernels long before the POSIX standard was created. If your kernel requires fsync() in such situations, your kernel isn't POSIX compliant, and you should file a bug report.
OTOH, if you're writing your data in chunks bigger than PIPE_BUF bytes, you probably should consider that an overlapping read may see an inconsistent state. But fsync() won't fix this; it's irrelevant to the issue of write/read overlap integrity.
Why is it unreasonable to expect the OS and the FS to handle this seemingly common case of writing lots of small files, by introducing a new "fsync all" API call if necessary?
Huh? We've had that API call for decades. It's called "sync".;-)
And there are better options than syncing for reliability. For example, rename the file to backup and then write a new file. The backup will still be there and can be used for automated recovery. Come to think of it, any decent text editor does it that way.
Huh? I've rejected several editors due to this behavior, and I know a number of project leaders that do the same thing. If you have any multiply-linked files, it breaks the linking, turning files in different directories into a flock of files, each with a few small changes. Discovering that this has happened and undoing the damage can be a real nightmare. The sensible approach is to only use editors that write when you say to write, to the file you say to write. (The main editors on unix systems have explicit options to control such backups.)
Of course, if you don't (know how to) use multiply-linked files, this isn't an issue for you. Unless you're working on a project that uses multiply-linked files, in which case you're a danger to the project. And if you don't understand when you should (and shouldn't) use multiply-linked files, I don't want you working on any projects of mine.
The right behavior is that when the editor writes the changes to a file and calls fsync(), the system writes the buffers as fast as it can. If it doesn't, fsync() is broken. If it does, but the disk controller holds some of the data in volatile buffers, you bought a bad disk controller.
And, in any case, you should have a UPS that can provide enough power to flush all the file buffers, even if all are marked modified. If not, you didn't buy a good enough UPS. And you should try to use OSs and disk controllers that write the data when the software tells them to. If you have all of these, you shouldn't have problems when the power goes out.
(It used to be common for unix systems to run a little 1-line C program that called sync() once a minute or so. Is this still at all common?)
Slashdot Mirror
Hmmm ... I looked at Thinkpads a few months ago, and couldn't find any with that resolution. This may have been because I couldn't find the actual numeric resolution on a lot of the models. It must be because I'm such a dummy that I can't easily follow the confused layout of the vendors' web sites ...
(I was also looking for models that came with linux, and generally couldn't find that info in the vendors' sites, either. ;-)
"This new technology is sponsored and funded by:
Your friendly health and life insurance company, constantly finding new and innovative ways to make sure we never have to pay you a dime since 1666."
For the benefit of those who didn't hear the show, last week NPR had an interesting interview with Karen Tumulty on just this topic. It was about her brother's problems with his insurance company when he started having kidney problems. Part of the story was that the insurance investigators got access to his medical tests, and they found a slight anomaly in one test a few years earlier that the doctors didn't consider significant at the time, but which the insurance company said indicated the kidney disease. Since he was insured by a different policy when that test was done, it means the disease was a "pre-existing condition" for the current policy, so they didn't have to cover it.
This is rapidly becoming a serious problem in the US. The more detailed your medical monitoring is, the more likely that the insurers will collect your money, and they refuse coverage due to a precursor that showed up once far in the past.
Of course, in the long run, the health insurers will shoot themselves in their collective feet, as health insurance slowly ceases to cover anything, and it will become pointless to pay for insurance that doesn't cover anything. Then the only people with medical coverage will be those able to pay the million-dollar charges for treating minor ailments.
(We're not quite there yet, but it's easy enough to be a prophet in this subject area. ;-)
It would then be up to me if I wanted to share this information with a doctor or others.
Not necessarily. Here in the US, if you want the possibility of medical treatment for most serious health problems, you have to have insurance. And you wouldn't be able to get insurance unless you gave the insurance salespeople access to your medical data.
Also, you might not be able to get anything but a minimum-wage job without giving your employer (and their insurance investigators) access to your data.
The poorer part of the population has been in this situation for some time, and the fraction of the population that it applies to is slowly increasing. Another decade, and it'll probably include most people like you.
Thanks for the reference. I'll look into that model, see whether linux works on it, etc.
I did notice that the first few google hits for places didn't actually give the screen resolution, just the 17" measurement and some superlatives. But one of the links to within Dell did give the 1900x1200 numbers.
But a lot of the scientific hypothesis we use right now, like the earth is round, seemed completely wacko once too.
Actually, the shape of the Earth is a bad thing to use as such an example. There's lots of historical evidence showing that people have understood the shape of the Earth for as long as we have historical records. Lots of textbooks describe the successful measurement of the Earth's size by those classical Greek guys, and similar things have been done in other cultures. Sailors on large bodies of water have always known that they were sailing on a big curved surface, from the way that things disappear from the bottom up as you sail away from them, and appear from the top down as you approach them.
The idea of a flat Earth historically has belonged to religious and political types that were insulated from real-world experience and could safely believe whatever nonsense they liked on such topics that didn't impact their everyday lives. But the guys down at the docks always knew better, because the curved surface they lived on was a significant part of their everyday lives, and they could see its shape.
It's curious especially that we still have this bizarre myth that Columbus was trying to prove something about the shape of the Earth. At the time, even the religious people accepted that the planet was spherical. The debate wasn't its shape; the debate was over its size. (And Columbus's trips didn't decide that. ;-) This is quite well documented in the historical records.
For most purposes, it's the flat-earth theory that should be considered wacko, as most knowledgeable people would have laughed at it at any time in history. Only religious nuts and political leaders believed such nonsense in the face of the evidence that any dummy can pick up in a single reasonably long sailing trip.
Part of the problem is that our main religious texts came from people living out in the desert, which is a terrain where measuring the size and shape of the Earth is both unimportant and very difficult. Now if we could just free our society from the influence of dummies who take some just-so stories written down by a few desert shepherds and insist that they are eternal truths about the whole world.
I have wondered who besides Apple sells a 17" laptop with a screen that's 1900x1200 pixels or better. I haven't seen one. Of course, I could easily have missed it, because with most laptops, it's impressively difficult to discover the pixel count. They tell you the diagonal size, with no other numeric information. Even if you can dig out a "specs" page, often the pixel count isn't anywhwere to be seen.
Maybe they think that we're all too stupid to understand the concept of resolution.
It is clear the slashdot editors run Windows/Mac machines and never view their site on a Linux based machine or they would have at least addressed the Javascript problems.
Nah; they read with NoScript blocking the scripts, too, like most of us do. ;-)
I use both this Mac and a linux machine, and I haven't seen that firefox or opera looks significantly different on either of them. Maybe there would be differences if I enabled scripts, but why would I do that?
(That wasn't being facetious. If anyone could give me a good motive to enable scripts, I might do it. But the few times I've experimented, I didn't see any improvement, and downloads were slower, so I turned scripts back off. So if you're allowing /.'s scripts, why are you doing that? Am I missing something useful?)
Every time I come here I wonder what kind of totally pointless misfeature I'll find next.
Hmmm ... I don't see any obvious misfeatures in /.'s UI. Of course, that might be because I went to my account page and turned off nearly everything that can be turned off. I also have AdBlock and NoScript installed in the browser that I use for /. (firefox), and have /. scripts are totally blocked.
The only obvious thing wrong is that there's a lot of the window that's just white space. The main slashdot.org page has the rightmost 1/3 blank, so the window covers screen space that could be used for something. And the discussion pages have too-wide indentation, so there's a lot of white space on the left. Anyone know how to cut down on those wastes of screen space?
For a few weeks, I had scripts enabled for /., but I was noticing that refreshes were taking a lot longer than before. So I turned scripts back off, and now things load in 1-2 seconds, so I guess I won't try that experiment again for a while.
Anyway, I can read articles and discussions, I can reply to messages, I can moderate when I grudgingly decide to do my social duty, and I can even meta-moderate. Is there anything else /. can do for me that I don't know about and have blocked?
I do sorta miss my message counter, though I'd have to admit that it's not something that's actually useful for anything. But we geeks and nerds do like numbers, y'know.
Yeah, what's the point of "better tab handling" and a "niftier search bar" if the results look like crap because it can't render everything properly?
Ya gotta consider that to their intended audience, proper rendering is defined as "however IE6 (mis)renders a page". Yes, a few people have updated this to mean "however IE7 (mis)renders a page", but that doesn't affect the outcome much.
To most of their intended audience, FF isn't "standard" because on their screen it renders the MS-only sites differently than their current IE does.
IE8 is aimed at people who have no clue about w3c.org, and wouldn't like it if they ever happened to accidentally stumble across it.
Try plugging centigrade temperatures into the ideal gas law and lemme know how it goes. ;)
Actually, the fun part is that there's a long history of people doing exactly that. The patent offices in the US and other countries have an ongoing problem of people attempting to patent perpetual-motion machines. In most cases, a "proof" that a particular gadget will work is produced by taking the standard equations and using Celsius/Centigrade numbers when temperature in "degrees" is needed. This seems to be something that a lot of physics cranks can't quite get right. They read that the Kelvin degree is the same as the Celsius degree, and that means that it doesn't matter which you use, right?
I don't even know how many ml there are in a tablespoon.
15.
There, now wasn't that easy? You can even treat the tablespoon as a metric term if you like. It's one of the "extended metric" units widely used in the US.
And there are 5 ml in a teaspoon. (In my experience, most Americans can't tell you how many teaspoons there are in a table spoon. That's one of many such questions that are good tests of how easy the Imperial system is to use. Most people flunk the tests pretty badly. ;-)
American cars are all metric.
This is probably mostly due to the fact that the components are no longer manufactured in the US. A lot of final assembly is still done here, but the pieces that actually have to be measured have mostly been manufactured in Asia for a couple of decades.
I've sometimes wondered whether part of the decline in the US auto industry was from the sizable fraction of the population that does a lot of the routine work on their own cars. Back in the 1970s, when Asian imports became common, a lot of Americans realized that "Hey, I don't need to keep two sets of auto tools any more. If I buy Japanese, I can throw out all those "imperial" tools (which the sales folks still insist on calling "standard" ;-) and just use standard (i.e., metric) tools where the arithmetic is so much simpler. The US auto companies did pick up on this and switch to mostly metric parts, but it took them a decade or so, and they'd lost the battle by then.
Of course, now the media is openly talking about the final death of the US auto companies ...
Too many possibilities of Earth origin are rejected with the phrase "it seems unlikely", and there's no mention of the most obvious method by which the micro-organisms get there: random motion (OK, particle velocities in the atmosphere will not be truly random, but you'd still expect a few outliers with very high velocities.)
Also, the "output" of the Earth's atmosphere was measured back in the 1960s and 1970s (and perhaps earlier). I remember reading a couple of articles in the 70s about our planet's "cometary tail", which was described as basically produced by the same solar-wind pressure as a comet's tail, and contains particles ranging from single atoms up to the size of bacterial spores. It was predicted then that when we examined the outer planets in detail, we'd find bacterial life there similar to Earth's, because that's where it came from.
The authors also pointed out that most of this dust tail would escape the solar system. It's a reasonable inference that we've been spewing this dust out, bacterial spores included, for some 4 billion years, which is about 16 trips around the galaxy, and the Earth's dust tail has permeated the galaxy during that time. Of course, this would be true of any other planet in the galaxy that has evolved bacteria-sized life.
The only real question is whether the spores can survive for long periods in space. 30 years ago, nobody had collected any bacteria from the Earth's outer atmosphere or from its dust tail, so survival time in space was purely conjectural. Now we have some samples that appear to be able to survive in space, at least long enough to get far enough away that the UV flux is no longer a danger. Whether these critters can survive the millions of years between stars is still a question.
But the default assumption now should be that some of them can survive. Most will die, or sleep forever, but some will encounter another object that contains edible food. Out between the stars, there doesn't seem to be anything we know of that would destroy something that can survive the UV flux of our stratosphere.
So the panspermia conjecture has just got a serious boost. And we've reached the point where, as the old joke goes, we present the most important statement in any scientific research: Further research is necessary. We know of a mechanism, and we know of a few bacteria that can survive the mechanism. But the numbers are all vague and fuzzy, barely respectable in scientific circles.
It's time to start applying for the funding for exploratory robots that can examine the outer planets' atmospheres and assorted planets' and asteroids' surfaces looking for more bacteria. We should expect to find earthly spores out there, and it'll be a surprise if we don't. Can we find tiny critters that are sufficiently different that we can say definitively that they didn't come from Earth?
I like it better when the customer is not always right.
What are you, some sort of realist? And who let you in here?
All predictions about fast adoption of linux because of it being cheaper have not come true partially because corporate service boys charged a healthy premiums on their linux 'loving' customers.
Oh, I dunno about that. A few months ago, I ordered the hardware for a new "desktop" system from a local computer assembler, and since I ordered it without the default Vista OS, I got a discount of a few hundred $$$. While talking about it with a rep over the phone just before delivery, he asked what I intended to install on it. I said "The latest Ubuntu release", and he said "We can install that for you, for no extra charge." I said "Huh?", and he said "Yeah; we've found that Ubuntu always installs quickly, with no problems at all. Give us an hour, and we can have it all set up for your." I told him "OK", and I got it with Ubuntu running just fine.
(Well, OK, there was a problem: They forgot to tell me the password that it wanted when I booted it. They were very apologetic about that. They were even more apologetic when I told them that, since they were closed when I got it home, I'd booted a handy knoppix briefly to mount the root partition and set the root password to something I knew. ;-)
I do sorta suspect that they wanted to do it as a training exercise for their installer guys, as a response to a good number of customers wanting that system installed. But no matter; the fact is that a local system builder took the attitude that "The customer is always right", and wanted their people to be able to install whatever the customer wanted.
Anyway, this one company didn't charge a healthy premium on a linux-loving customer. They said "We can do that for you for no extra charge." And, needless to say, I told a number of other local friends about it, probably resulting in a few more sales.
YMMV, of course.
... people seem to think throwing money at proprietary software will magically solve their problems ...
This attitude is exemplified by the mantra "You get what you pay for". We've read that here in numerous /. discussions. Maybe it has already appeared in this discussion.
IBM and Microsoft are among the many companies that encourage this belief in their marketing, because it's such a good tool to make people spend money without doing the sensible price shopping and product comparisons that intelligent purchasing requires. Anyone who believe this mantra is their natural prey.
It's especially bizarre to hear or read "You get what you pay for" in a computer setting. The history of our industry is full of example of just the opposite. Especially with software, quality tends to be an inverse function of price. The high-priced stuff is mostly designed to be sold at a "management" level, i.e., to people who have no ability to compare the quality of various software products. It's especially unfortunate that this happens with school purchases, since this encourages the "monoculture of the worst" that we see all over. Then people buy the same stuff for home use because it's the only thing they can (barely) use without further training, and also because it's expensive and "You get what you pay for".
Of course, the FOSS crowd did sorta blow it by using the phrase "free software". Yes, they meant "free" as in freedom, not as in beer. But the other 99% of the population understands "free" as a euphemism for "cheap and shoddy", because it's what they've learned it means when salesmen use the term. "You get what you pay for", and if it's free, it must be crap, right?
Fighting this attitude is the biggest marketing problem for anyone selling cheap-but-good stuff. This is especially true for software, for which quality is nearly independent of the time and effort (i.e., the cost) of development, but people still believe that cheap stuff must be crap.
Maybe the solution is to sell expensive editions that only differ from the free stuff in the price (and maybe an expensive-looking splash page).
Every person I know says "March fourteenth two thousand nine", not "fourteenth March two thousand nine".
Typical "strawman" argument, disproof by using the wrong syntax.
I've often heard "the fourteenth of March, two thousand [and] nine." As a native speaker of English, I'd claim that that's every bit as natural as "March [the] fourteenth, two thousand [and] nine".
So dd/mm/yyyy should be as normal in English as the weird middle-endian form used by most Americans.
Myself, I use the ISO format (yyyy-mm-dd hh:mm_ss) whenever I think I can get away with it. That way, you minimize the computer code needed to handle the date. And nobody that I know of uses yyyy-dd-mm, so there isn't the usual annoying ambiguity that software just can't deal with.
The problem with the year-last formats is that when someone writes something like 4/9/2009, you generally have no way of knowing which order they're using and which date they're talking about. It's not just a problem with software; now that we have a single worldwide communication system, it's a problem in human communication, too. The only way to win that game is to not play it.
If you dig further into the history, you'll find that there's a lot of precedent for that.
I've read a few explanations that one of the many reasons for the US's "separation of church and state" in the Bill of Rights was that the UK had traditionally not recognized marriages outside the established church. This was among the many reasons that various groups (Puritans, Quakers, Jews, etc.) had emigrated to the wilds of North America, so that they could set up local governments that would honor their right to such institutions as marriage (and land ownership, right to enter into contracts, right to persecute people not like them, etc.).
The people in the US claiming ownership of marriage and demanding that "people not like us" not be allowed to marry are carrying on an old tradition. Their predecessors had to grudgingly accept marriages among black people, but for a long time many states refused to permit mixed-race marriages. Then the Supreme Court said that this had to be permitted, but most states still refused to allow marriages in many other cases (e.g., same-sex marriages). Now a few states have permitted that, and the "only people like us" people are getting desperate to find people they can legally exclude.
A funny thing about the religious objectors to same-sex marriage is that my wife and I (;-) attended several same-sex weddings here in Massachusetts before the state legalized same-sex marriage. These weddings were in fact performed in churches (or a synagogue in one case). So it's obvious that some religious groups implemented same-sex (and mixed-race) marriage since long before such things became recognized by the state. The people claiming that this is something new have no knowledge of history. Various religious bodies have been doing "disapproved" marriages for as long as we have any records of such things, and some of those religious bodies have kept records going back centuries. They don't advertise it, of course, for obvious reasons. But a number of historians have found the records and written about the actual practices.
Yeah. Now if we could just make code that uses such things portable to systems like *BSD, Solaris, OS X, etc. It's a real pain trying to track down how you do it everywhere (except MS systems, of course ;-). You have to write code that explores the system to discover which gimmicks they've included and whether they actually work. And you need copies of each kind of system for testing.
Just calling sync() or fsync() is a lot easier, though as others have observed, some systems (and some disk controllers) have found ways of subtly shooting down that approach.
I checked the top 10 (you can check the rest), and they *ALL* have JavaScript.
Not when I visit them with firefox, they don't.
I have NoScript installed. So those sites load really fast in firefox.
(And I'm not at all sorry that I'm missing the javascript "active" ads. ;-)
Yeah; I suspected that that's what you meant. But you can get more humor out of a situation by answering an extreme interpretation of what someone just said ...
Anyway, I've always sorta thought it was good if a few programs were calling sync() at a moderate rate. On a few systems, I've added a little background process that just sleeps and calls sync() every N minutes. This was usually in response to someone discovering that the system was holding unwritten data in memory for hours because the system load was too light to force flushing. It's always seemed to me that an OS should have a system call that sets the max time that a write buffer can exist before it's flushed. With all the never-used features that are bloating up all our OSs, you'd think they'd have included that one.
Or maybe they have, and they just haven't told us mere users.
... can't imagine many reasons for having hundreds of tiny config files for a single app ...
Ooh, ooh, I know one!
Some years back, I stumbled into being the admin of one of what are by now probably several thousand specialized "search engines" for a certain kind of rather technical data whose nature isn't too important here. The "app" includes a search bot and a single-request download program, and both of them have to deal with what is now nearly 400 sites that have the data. Each site has its own web server, and I sometimes suspect that no two of them are running the same server. So there's a "cfg" directory that contains a file for each hostname, and the file contains config info for dealing with that host's web server.
The most important thing is getting the HTTP level right, because a lot of servers will reject requests if you ask for the wrong level. The default is "HTTP/1.1", but a significant number of the sites' servers require a different level. And this can't be set up at the first access to a host, because sometimes people upgrade or reconfigure their servers. So what the download routine does is look for the HTTP/* entry in a host's cfg file, and uses "HTTP/1.1" if it isn't found. Occasionally a download attempt fails, and the routine goes into a "determine HTTP level" mode, in which it tries a list of level numbers looking for one that works. If it succeeds, it updates the host's cfg file to tell later downloads what level to use.
Actually, there aren't "hundreds" of config files (yet). The number is only slightly over 100, because most of the servers work fine with the default settings. Most of the cfg files are tiny, because few servers have more than one or two settings that need to be specified. The biggest is 468 bytes; the rest are under 100. I expect that the number of files could reach 200 in a few more years.
I expect that others will be able to describe some more situations where they use a large flock of tiny config files for some app or set of closely-related apps.
The app updates a config file by building the new contents in memory, and overwriting the file with a single write() call. It doesn't bother with a flush() call, because such updates are very rare, and if the write fails, it just means that the "determine HTTP level" code will be done again the next time that host is accessed. In a decade or so of life, this has never happened. I can say this with some certainty, because such things get written to a global "special events" log file, so I've seen all the cases where an app updates a config file.
Name one scenario where fsync should be necessary other than: ...
2) rewriting data back to an existing file... like for instance something a database might do.
Hmmm ... The followups imply that some people have a misunderstanding of how the POSIX write() call works, and interpret the above as implying that when process A writes data to a file and process B reads the same spot in the file, B may get a jumbled mixture of the old and new data. But the POSIX spec for write() explicitly states that a write of up to PIPE_BUF bytes is atomic. This was generally true of unix kernels long before the POSIX standard was created. If your kernel requires fsync() in such situations, your kernel isn't POSIX compliant, and you should file a bug report.
OTOH, if you're writing your data in chunks bigger than PIPE_BUF bytes, you probably should consider that an overlapping read may see an inconsistent state. But fsync() won't fix this; it's irrelevant to the issue of write/read overlap integrity.
Why is it unreasonable to expect the OS and the FS to handle this seemingly common case of writing lots of small files, by introducing a new "fsync all" API call if necessary?
Huh? We've had that API call for decades. It's called "sync". ;-)
And there are better options than syncing for reliability. For example, rename the file to backup and then write a new file. The backup will still be there and can be used for automated recovery. Come to think of it, any decent text editor does it that way.
Huh? I've rejected several editors due to this behavior, and I know a number of project leaders that do the same thing. If you have any multiply-linked files, it breaks the linking, turning files in different directories into a flock of files, each with a few small changes. Discovering that this has happened and undoing the damage can be a real nightmare. The sensible approach is to only use editors that write when you say to write, to the file you say to write. (The main editors on unix systems have explicit options to control such backups.)
Of course, if you don't (know how to) use multiply-linked files, this isn't an issue for you. Unless you're working on a project that uses multiply-linked files, in which case you're a danger to the project. And if you don't understand when you should (and shouldn't) use multiply-linked files, I don't want you working on any projects of mine.
The right behavior is that when the editor writes the changes to a file and calls fsync(), the system writes the buffers as fast as it can. If it doesn't, fsync() is broken. If it does, but the disk controller holds some of the data in volatile buffers, you bought a bad disk controller.
And, in any case, you should have a UPS that can provide enough power to flush all the file buffers, even if all are marked modified. If not, you didn't buy a good enough UPS. And you should try to use OSs and disk controllers that write the data when the software tells them to. If you have all of these, you shouldn't have problems when the power goes out.
(It used to be common for unix systems to run a little 1-line C program that called sync() once a minute or so. Is this still at all common?)