The problem is that politics tries to categorise things into good and bad {and, sometimes, indifferent} whereas pure science does not make such distinctions. Pure scientific phenomena are neither good nor bad; they just are.
For instance, there exists a very simple scientific experiment which would determine once and for all whether or not humans an chimpanzees belong to the same genus. It probably will never be conducted, for political reasons.
The question which remains to be answered is, can political methods be used to control the spread of scientific knowledge? Up until a few years ago I would have thought a resounding NO. Today, I'm not so certain. As technology advances inevitably further and further beyond the understanding of the common person, and so deeper into the domain of multi-national corporations, its uses are being controlled by ever more draconian laws. There is a very real possibility that the world could enter a new Dark Age.
And frankly, I'm a lot more afraid of WW3 than global warming.
Typical American attitude. You people are so obsessed with war, you can't imagine any other ways of dying. You will most probably still be hunting for "terrorists" the day a "natural" disaster {probably of human origin} wipes you out.
It's my experience -- working for a company that does some ISP work, and having to speak to other ISPs -- that those ISPs that only support Windows simply would not know know their arse from a hole in the ground. The people who eat at McDeaths every day and drive BMWs will tell you to reboot, reinstall, download some anti-virus software {yes really -- I've been told to download software tools to cure an ISP's DHCP server fault}, anything to get you off the phone and take up some of your time while the real upstream problem is getting sorted, by the people who cook leftovers on a Coleman stove for lunch and whose cars are held together with gaffer tape and string.
If your Ubuntu box can't connect to the internet, but it could before and you haven't deliberately changed anything, then the chances are good that the problem is at the ISP's end {or it's your router; but if you can see the web configurator from inside, you know that it's working}. If you haven't deliberately changed anything on your Windows box, that isn't saying much: Windows boxes sometimes change things all by themselves and for no obvious reason.
It's a common misconception, but you do not need a licence from Microsoft to run Windows. The Law of the Land {in the US and the UK at least} already gives you permission to do that: as long as you acquired Windows legitimately, the right to make a copy of the program in memory for the purpose of execution is part of your right of fair dealing / fair use. The Law also gives you the right to transfer your copy of Windows to another person: that would come under exhaustion of rights / first sale doctrine.
These are statutory rights, given to you by the Law of the Land. No agreement into which you enter can affect your statutory rights and be worth the paper it's printed on.
However..... some countries are known to have corrupt legal systems, where the winner of a civil action may be liable for some costs and/or money may be required to change hands before a verdict is delivered. If you live in such a country, the winner of any civil suit is invariably the person who had more money to begin with.
At any rate, the question you want to be asking yourself is "Do I want these people to be taking their food from the skips behind supermarkets, which is fairly harmless but implicitly depends on the supermarket being there; or do I want them to be growing all their own food, and so not dependent upon anyone else?"
To be sure that they will work, you need to compile your extensions against the Mozilla source tree from which you compiled Firefox. That is not a limitation of any specific technology; rather, it is a limitation of mathematics. You have no more business expecting an extension compiled against 1.0.8 to work with anything except 1.0.8, than you have expecting that a CD will play on a cassette walkman.
Now, if the extensions won't compile against the new source tree, you may need to patch them. Perhaps a variable or function has changed its name. Or a function might require an extra parameter. Read the changelog and get hacking, if you can't wait for your favourite extensions developer to do so.
The article mentioned that shape-recognition technology is being employed. This is especially interesting because the abstract mathematics that underlie shape recognition are the same as the abstract mathematics that underlie another computational problem: decompilation.
Machine-language instructions are vertices. High-level-language loops, functions and similar structures are more or less complex primitive shapes {squares, triangles, circles.....} The question "To what shape do these points belong?" resembles mathematically "What high-level-language control structure does this set of machine-language instructions represent?"
If someone has got it together to manage shape recognition that can detect the human form, then it cannot be long before someone else independently gets it together to recognise the "shapes" of structures within compiled object code. That will have some very profound implications indeed.
I've long suspected a good Vindaloo was good for you! Now they've just got to find a way to prove that having anything covering your lower arms is bad for you.....
Re:Reading anything on tomshardware....
on
The Story of Tron
·
· Score: 2, Informative
Really? I don't get anything like that in Konqueror 3.4.2.
If you wanted to, you could go to the local newsagent's shop and "pirate" a newspaper using the photocopier. Except that, even if you were prepared to put up with a stack of loose papers with overflows, it would still cost you more than buying the original.
That's the business model the CD companies need to start using if they want to avoid "piracy".
The maths: A CD-R ready for burning costs just slightly more to manufacture than a stamped CD-DA or CD-ROM {assuming you already have a master to work from}. An audio cassette costs three times more to manufacture than a CD, again assuming you have a master or want it blank. Burning a CD-R takes ($DURATION / $SPEED) * (100 / $HIT_RATE_PER_CENT) minutes, during which your computer cannot be used for anything else. The hit rate is not 100% when you use cheap CD-R media at high speeds; you can get better hit rates at lower speeds or on more expensive media, but in the final analysis it works out cheaper to put up with a few beermats {and no worse for the environment, since the manufacturing process hit rate is the same. Just as many or more "expensive" discs end up in landfill as "cheap" -- the only difference is they haven't passed through consumers' hands first}. You then have to print the cover artwork and a label {it used to be possible to get away with the title handwritten onto the disc with an indelible marker and a simple track listing, even handwritten if the burning run was short, for the right price..... ah for the days when people were easier to satisfy} which will take 1.5 * (100 / $HIT_RATE_PER_CENT) A4 pages at medium/high ink coverage. Call it £3 per CD and even though the blank disc is the cheapest item, you're still making barely any profit -- you'll never sell enough "pirated" CDs to pay your living expenses. But it's a good sideline if you're in the kind of job where you literally have to decide each day whether to catch the bus home, or walk home and eat. {Bear in mind you can burn off more calories walking in wet clothing than you'd get from a meal.}
The record companies aren't having to pay for individual inkjet prints; so they can have a CD packaged and ready to sell for a lot less what than the "pirates" can do it for. They can sell it for the same price, maybe a pound more -- for most people, that much money is not worth the effort of saving {especially not for the sake of the lyrics, band photos and other nice stuff you don't get with a "pirated" CD}. As long as they sell enough CDs to cover the one-off costs of mastering {the band's advance and studio time}, everything else is pure profit.
It's possible for any sufficiently-determined Fred in the Shed to build a weapon of mass destruction; be that nuclear, biological or chemical. Radioactive substances can be obtained with a little hard work; even if you don't achieve a nuclear explosion, you will contaminate an impressive area with radiation. You can breed some really evil germs with the right starter culture, and there's probably some somewhere in every woman's house. If you've studied chemistry to A-level then you will have heard mention of some pretty toxic substances. If you happen to be a corrupt government, then you can misappropriate money, industrial facilities and labour as required.
It's not hard to imagine a scenario where some high-school kid actually succeeds in making a really serious weapon with the intention to exact revenge on the school bullies. They would have a hard job ahead of themselves -- there is always the risk of a fatal error when working with deadly materials. It would invariably be easier, if a lot less satisfying, just to get hold of a firearm; and there would almost certainly be a media cover-up if and when it eventually happened. What else are they going to say -- that a seventeen-year-old cultured deadly bacteria in his mum's airing cupboard, enough of the stuff to fell an entire regiment, to deal with a dozen or so idiots who had been calling him a poof twice a day for the last four years?
You'd have to be supremely arrogant to dismiss the idea out-of-hand. Whatever's keeping people from building weapons is not ignorance.
You know, one thing that might really help PostreSQL is if the advocates didn't complain bitterly about how unfairly unpopular they were, and instead built bridges to MySQL.
*sigh* Where are mod points when you need them?
I think all Open Source users are, to some greater or lesser extent, a bit self-righteous about it. Like "My system is purer than yours". It's the same with the BSDs versus Linux. This attitude can be counter-productive because it creates artificial divisions between people who, if it came down to a battle of "good" versus "evil", should be on the same side. And while the Free / Open Source Community descend into factions and squabble amongst ourselves over trivia, The Enemy is making progress.
It would be nice to see some articles on switching from MySQL to PostgreSQL, covering the how {bridging the dialect gap}, the why {explaining the improvements}, and including a checklist of points to help MySQL users see when its limitations are beginning to bite. And what sort of things can PostgreSQL do for itself, that would need a script to do them in MySQL. I'm sure MySQL is adequate for what most of its users expect {glorified variable persistence, if we're brutally honest}; but there must be a proportion of advanced users who are encountering its limitations and probably don't even realise it. The people who have had projects that started modest and grew slowly but relentlessly, like a family that eventually outgrow their car. These are the people the PostgreSQL folks really need to convince -- which means not getting all snobbish and hostile because of where they've come from, but welcoming them into the community. Try catching flies with honey as opposed to vinegar!
Well, that worked fine for me..... maybe it's a problem your end? If it's worked before and you've since installed something that broke cropping, you probably need to check your libraries. Check the output of ldd on each of the main ImageMagick executables {probably in/usr/X11/bin/ or/usr/bin, but try something like which convert if you're not sure} and make sure the files referred to are OK. In the worst case, you may have to re-compile ImageMagick from source. If you aren't using your own distribution's specific source package, change the PREFIX to match wherever the old versions were installed, so they will be overwritten by the new ones.
Why would you want to run Windows XP on an Apple computer? You might just as well watch a scratchy old worn copy-of-a-copy VHS cassette on a brand spanking new high-definition digital-ready TV set; or go to a posh restaurant, all done up in a tux and bow tie, and order sausage, egg and chips with brown sauce.
You could probably do it by using elilo to bootstrap ntldr.exe, but this is a gross oversimplification and I have not thought it through fully.
I recently acquired a "broken" HP OfficeJet for nothing. Turned out not to be very broken after all; I can't get any yellow ink to come out for love nor money {and it cost me £50 just to find that out}, but it does mono copies fine {though at £25 for an ink cartridge, I can't imagine I will be making many copies} and it's also supported under GNU/Linux.
You just need to get a scanner which is supported by SANE. That rules out a lot of the povvy cheap Windows-only ones. Write a BASH script to perform full-page scans, translate the format if required and dump the files into a directory. Configure an FTP daemon to see that directory. Easy!
That's because the PS/2 mouse hardware generates interrupts; the processor has to stop what it's doing in order to see what was trying to interrupt it and decide if or not it's important enough to deal with there and then or it can wait. This fraction of a second is just long enough to relieve whatever condition was causing the processor to foul up. My guess, since it involves reading and writing large quantities of data in a contiguous fashion, is some sort of race hazard where if two logic inputs change state in the wrong order, the output is not what you were expecting. But I Am No Chip Designer.
Ubuntu have source repositories, you know. An Ubuntu source package {and for that matter, a Debian source package, since Ubuntu is based on Debian} consists of three files. The orig.tar.gz file {as supplied by the author}, the diff.gz file {containing various tweaks required to customise the source tree and Makefile to Debian / Ubuntu's way of doing things} and the.dsc file {containing various metadata}. You can use apt-get or aptitude to download the source files, and dpkg to build a.deb package and then install this as normal. I don't think there's a way to bypass the.deb creation stage and go straight to installing the freshly-compiled binaries; but I'm sure that someone will correct me soon enough, if there is one.
What really makes the difference between a good distribution and a great distribution is the size and quality of the package repository. Problems with repositories might be: packages completely absent, straightforward bugs or unmentioned dependencies -- i.e. where a package foo depends on a package bar but this is not included in the dependency metadata, and all the developers happen to be running bar anyway so nobody notices. This is actually the worst of all potential repository issues, as those most likely to be in a position to notice it are inexperienced users {otherwise they probably would have installed the dependency anyway with something else} and so not used to the whole dependency system {it does take a bit of getting your head round at first, though it's obvious once you come to understand it}.
Back in the day, only a minority of people could afford recording equipment. They soon found that they couldn't make enough money out of charging artists to record albums which would then become the artist's property that they could sell to the public, so they came up with another model: get the artist to sign over their rights -- in exchange for a one-off payment -- to the recording company, who would take care of the whole business of selling records and arrange for the artist to be paid a cut from each one sold.
It's this exclusivity that's the beginning and end of the problem -- the fact that once an artist is having their recordings distributed by one label, no other label can distribute their recordings.
Most other things are available from several sources. For instance, I can buy a loaf of bread baked by Sunblest, Kingsmill or Warburtons; or I can buy flour and yeast and bake my own. There is, in principle, no artificial barrier to a new player entering the market; if their product offers value, as judged by those who buy it, then it will be successful. "Value" is of course a nebulous concept, and so it should be; but in this case it is likely to mean a loaf of bread that tastes better, or costs less than what is already available.
The same holds true for things such as standardised industrial components. If I'm buying M4 x 20 steel bolts with a raised Posidriv head, or 4.7k ohm 0.25 watt resistors, or 80 gsm A4 paper in packs of 500 sheets, or 15mm. end-feed plumbers' tee-pieces, I still have a choice of suppliers; but there is no subjective assessment of value, and anyone's part will do the same job as anyone else's.
I think there is a market about to emerge for a new way of running a record label. The steps would be as follows:
Artist obtains a loan, using the rights over their work as collateral.
Record company manufactures CDs for a fee, which Artist pays using part of loan -- Lender has a lien over CDs.
Artist uses remainder of loan to promote CD.
Until loan is repaid, Lender can exercise control over certain uses of recording.
As soon as loan is repaid, rights in recording revert to Artist.
There's no reason in principle why an artist shouldn't have the same album available on different labels, possibly even selling at different RRPs, at the same time; they would be competing with one another strictly on the basis of the services they offered {studio facilities, in-house producers and session musicians, pressing capacity and so forth}. Some labels would specialise in quick turnaround at a slight premium. Supermarkets probably would have their own record pressing operations, but no studio facilities -- they would work strictly from masters.
What is really required is for the law to state unequivocally, once and for all that, as long as you own a recording on a legitimately-acquired medium sanctioned by the copyright holder, your "fair dealing" rights include the right to make an unlimited number of copies of that recording in alternative formats for your own use, and to perform any necessary step in the process: copyright would not be infringed unless you used a copy you had made in some way that you would not be permitted to use the original.
What would be even better would be a ban on DRM systems that prevent absolutely the exercise of Fair Dealing rights and/or copying under Special Licence {e.g., I have permission from the band Ocean Colour Scene to make copies of any of their work for my own use; any DRM system that does not take this into account, perhaps by requiring a password to enable copying, should be illegal}.
At the end of the day, you're damned if you do try to imitate Windows and you're damned if you don't. If you try to be like Windows, that can mean copying inherently broken behaviour -- and will lose you friends in the "keep it pure" camp. If you don't try to be like Windows, somebody will complain that Linux is "too hard" {i.e. "not like Windows"}. If you try to make an application finely customisable, you end up driving people away because it's "hard to use"; if you don't include options to change things, you end up accused of "dumbing down".
I can only really attribute the "problem" to Microsoft's dominance in the marketplace combined with the popular mindset, which deems that "ignorance is bliss" and eschews learning to do something very hard, very well in favour of instant gratification with a half-arsed job.
That's why I think it's important for distributions to specialise. At the moment we have Ubuntu and Mandriva for people who want everything easy; Slackware and Debian for server administrators who feel the need to ride the metal; and Red Hat and SUSE for people who would rather pay someone else to do the donkey work. Not to mention hundreds if not thousands of less well-known distributions, catering to niche markets {self-booting mini-CDs, distributions tailored for antique hardware, retro gaming kits, movies on a self-booting CD, Linux on a USB stick and so forth}. One distribution simply can't be all things to all people.
One thing I would like to see would be a GUI front-end to the configure, make, make install process. It's distribution-agnostic, sometimes even architecture-agnostic. Now that processor power is so cheap, the only compelling reason not to compile locally has been mitigated. A graphical front-end would look a little bit like a Windows InstallShield installer. What puts people off source tarballs isn't so much the idea of compilation {though that's where they will inevitably transfer the blame}, as the thought of unresolved dependencies breaking the process. There's no reason why a properly-put-together automake/autoconf package should not be able to detect everything it needs at the configure stage. Linux allows you to mix and match libraries to an extent; so if a particular application imperatively requires newer libraries than are already installed, that need not be a problem. The installer should be able to determine for itself whether it's possible to download and install its own dependencies, and proceed automatically if it is safe to do so.
Of course, probably before the GUI source installer goes mainstream, we will need a reliable developer tool for creating self-installing packages; analysing libraries and creating a dependency database. Although this sounds like a huge effort, it probably will be more likely to succeed than any attempt towards achieving cross-distribution binary compatibility; binaries were never really meant to be compatible, source was always meant to be compatible.
That's hardly surprising in a metropolitan area. Whenever one train has to be re-platformed, the train that was going to use the platform that train is now using also has to be re-platformed, and so on, until there is a big enough time window to get one train out of the way before the xext comes in.
It's worse in the South because there are two electrification systems in use; the old Southern Electric, third-rail DC system and the modern, overhead AC system. Not all vehicles are dual-powered, and neither are all tracks, so re-platforming options are limited. Of course, since {as every Londoner knows} there is no electricity in the North of England, this will not be a problem in Leeds:)
Slashdot Mirror
Science and politics don't get on well.
The problem is that politics tries to categorise things into good and bad {and, sometimes, indifferent} whereas pure science does not make such distinctions. Pure scientific phenomena are neither good nor bad; they just are.
For instance, there exists a very simple scientific experiment which would determine once and for all whether or not humans an chimpanzees belong to the same genus. It probably will never be conducted, for political reasons.
The question which remains to be answered is, can political methods be used to control the spread of scientific knowledge? Up until a few years ago I would have thought a resounding NO. Today, I'm not so certain. As technology advances inevitably further and further beyond the understanding of the common person, and so deeper into the domain of multi-national corporations, its uses are being controlled by ever more draconian laws. There is a very real possibility that the world could enter a new Dark Age.
It's my experience -- working for a company that does some ISP work, and having to speak to other ISPs -- that those ISPs that only support Windows simply would not know know their arse from a hole in the ground. The people who eat at McDeaths every day and drive BMWs will tell you to reboot, reinstall, download some anti-virus software {yes really -- I've been told to download software tools to cure an ISP's DHCP server fault}, anything to get you off the phone and take up some of your time while the real upstream problem is getting sorted, by the people who cook leftovers on a Coleman stove for lunch and whose cars are held together with gaffer tape and string.
If your Ubuntu box can't connect to the internet, but it could before and you haven't deliberately changed anything, then the chances are good that the problem is at the ISP's end {or it's your router; but if you can see the web configurator from inside, you know that it's working}. If you haven't deliberately changed anything on your Windows box, that isn't saying much: Windows boxes sometimes change things all by themselves and for no obvious reason.
It's a common misconception, but you do not need a licence from Microsoft to run Windows. The Law of the Land {in the US and the UK at least} already gives you permission to do that: as long as you acquired Windows legitimately, the right to make a copy of the program in memory for the purpose of execution is part of your right of fair dealing / fair use. The Law also gives you the right to transfer your copy of Windows to another person: that would come under exhaustion of rights / first sale doctrine.
..... some countries are known to have corrupt legal systems, where the winner of a civil action may be liable for some costs and/or money may be required to change hands before a verdict is delivered. If you live in such a country, the winner of any civil suit is invariably the person who had more money to begin with.
These are statutory rights, given to you by the Law of the Land. No agreement into which you enter can affect your statutory rights and be worth the paper it's printed on.
However
At any rate, the question you want to be asking yourself is "Do I want these people to be taking their food from the skips behind supermarkets, which is fairly harmless but implicitly depends on the supermarket being there; or do I want them to be growing all their own food, and so not dependent upon anyone else?"
To be sure that they will work, you need to compile your extensions against the Mozilla source tree from which you compiled Firefox. That is not a limitation of any specific technology; rather, it is a limitation of mathematics. You have no more business expecting an extension compiled against 1.0.8 to work with anything except 1.0.8, than you have expecting that a CD will play on a cassette walkman.
Now, if the extensions won't compile against the new source tree, you may need to patch them. Perhaps a variable or function has changed its name. Or a function might require an extra parameter. Read the changelog and get hacking, if you can't wait for your favourite extensions developer to do so.
The article mentioned that shape-recognition technology is being employed. This is especially interesting because the abstract mathematics that underlie shape recognition are the same as the abstract mathematics that underlie another computational problem: decompilation.
.....} The question "To what shape do these points belong?" resembles mathematically "What high-level-language control structure does this set of machine-language instructions represent?"
Machine-language instructions are vertices. High-level-language loops, functions and similar structures are more or less complex primitive shapes {squares, triangles, circles
If someone has got it together to manage shape recognition that can detect the human form, then it cannot be long before someone else independently gets it together to recognise the "shapes" of structures within compiled object code. That will have some very profound implications indeed.
I've long suspected a good Vindaloo was good for you! Now they've just got to find a way to prove that having anything covering your lower arms is bad for you .....
Really? I don't get anything like that in Konqueror 3.4.2.
That's nothing ..... on 03/10/2005, I got offered half-price tickets to see The Two Ronnies live at the Haymarket Theatre!
If you didn't laugh, you'd have to cry.
If you wanted to, you could go to the local newsagent's shop and "pirate" a newspaper using the photocopier. Except that, even if you were prepared to put up with a stack of loose papers with overflows, it would still cost you more than buying the original.
..... ah for the days when people were easier to satisfy} which will take 1.5 * (100 / $HIT_RATE_PER_CENT) A4 pages at medium/high ink coverage. Call it £3 per CD and even though the blank disc is the cheapest item, you're still making barely any profit -- you'll never sell enough "pirated" CDs to pay your living expenses. But it's a good sideline if you're in the kind of job where you literally have to decide each day whether to catch the bus home, or walk home and eat. {Bear in mind you can burn off more calories walking in wet clothing than you'd get from a meal.}
That's the business model the CD companies need to start using if they want to avoid "piracy".
The maths: A CD-R ready for burning costs just slightly more to manufacture than a stamped CD-DA or CD-ROM {assuming you already have a master to work from}. An audio cassette costs three times more to manufacture than a CD, again assuming you have a master or want it blank. Burning a CD-R takes ($DURATION / $SPEED) * (100 / $HIT_RATE_PER_CENT) minutes, during which your computer cannot be used for anything else. The hit rate is not 100% when you use cheap CD-R media at high speeds; you can get better hit rates at lower speeds or on more expensive media, but in the final analysis it works out cheaper to put up with a few beermats {and no worse for the environment, since the manufacturing process hit rate is the same. Just as many or more "expensive" discs end up in landfill as "cheap" -- the only difference is they haven't passed through consumers' hands first}. You then have to print the cover artwork and a label {it used to be possible to get away with the title handwritten onto the disc with an indelible marker and a simple track listing, even handwritten if the burning run was short, for the right price
The record companies aren't having to pay for individual inkjet prints; so they can have a CD packaged and ready to sell for a lot less what than the "pirates" can do it for. They can sell it for the same price, maybe a pound more -- for most people, that much money is not worth the effort of saving {especially not for the sake of the lyrics, band photos and other nice stuff you don't get with a "pirated" CD}. As long as they sell enough CDs to cover the one-off costs of mastering {the band's advance and studio time}, everything else is pure profit.
Printer n A machine for generating scrap paper. Laser ~ A machine for generating lots of scrap paper very quickly.
Unfortunately it is not that simple.
It's possible for any sufficiently-determined Fred in the Shed to build a weapon of mass destruction; be that nuclear, biological or chemical. Radioactive substances can be obtained with a little hard work; even if you don't achieve a nuclear explosion, you will contaminate an impressive area with radiation. You can breed some really evil germs with the right starter culture, and there's probably some somewhere in every woman's house. If you've studied chemistry to A-level then you will have heard mention of some pretty toxic substances. If you happen to be a corrupt government, then you can misappropriate money, industrial facilities and labour as required.
It's not hard to imagine a scenario where some high-school kid actually succeeds in making a really serious weapon with the intention to exact revenge on the school bullies. They would have a hard job ahead of themselves -- there is always the risk of a fatal error when working with deadly materials. It would invariably be easier, if a lot less satisfying, just to get hold of a firearm; and there would almost certainly be a media cover-up if and when it eventually happened. What else are they going to say -- that a seventeen-year-old cultured deadly bacteria in his mum's airing cupboard, enough of the stuff to fell an entire regiment, to deal with a dozen or so idiots who had been calling him a poof twice a day for the last four years?
You'd have to be supremely arrogant to dismiss the idea out-of-hand. Whatever's keeping people from building weapons is not ignorance.
I think all Open Source users are, to some greater or lesser extent, a bit self-righteous about it. Like "My system is purer than yours". It's the same with the BSDs versus Linux. This attitude can be counter-productive because it creates artificial divisions between people who, if it came down to a battle of "good" versus "evil", should be on the same side. And while the Free / Open Source Community descend into factions and squabble amongst ourselves over trivia, The Enemy is making progress.
It would be nice to see some articles on switching from MySQL to PostgreSQL, covering the how {bridging the dialect gap}, the why {explaining the improvements}, and including a checklist of points to help MySQL users see when its limitations are beginning to bite. And what sort of things can PostgreSQL do for itself, that would need a script to do them in MySQL. I'm sure MySQL is adequate for what most of its users expect {glorified variable persistence, if we're brutally honest}; but there must be a proportion of advanced users who are encountering its limitations and probably don't even realise it. The people who have had projects that started modest and grew slowly but relentlessly, like a family that eventually outgrow their car. These are the people the PostgreSQL folks really need to convince -- which means not getting all snobbish and hostile because of where they've come from, but welcoming them into the community. Try catching flies with honey as opposed to vinegar!
Well, that worked fine for me ..... maybe it's a problem your end? If it's worked before and you've since installed something that broke cropping, you probably need to check your libraries. Check the output of ldd on each of the main ImageMagick executables {probably in /usr/X11/bin/ or /usr/bin, but try something like which convert if you're not sure} and make sure the files referred to are OK. In the worst case, you may have to re-compile ImageMagick from source. If you aren't using your own distribution's specific source package, change the PREFIX to match wherever the old versions were installed, so they will be overwritten by the new ones.
Why would you want to run Windows XP on an Apple computer? You might just as well watch a scratchy old worn copy-of-a-copy VHS cassette on a brand spanking new high-definition digital-ready TV set; or go to a posh restaurant, all done up in a tux and bow tie, and order sausage, egg and chips with brown sauce.
You could probably do it by using elilo to bootstrap ntldr.exe, but this is a gross oversimplification and I have not thought it through fully.
I recently acquired a "broken" HP OfficeJet for nothing. Turned out not to be very broken after all; I can't get any yellow ink to come out for love nor money {and it cost me £50 just to find that out}, but it does mono copies fine {though at £25 for an ink cartridge, I can't imagine I will be making many copies} and it's also supported under GNU/Linux.
You just need to get a scanner which is supported by SANE. That rules out a lot of the povvy cheap Windows-only ones. Write a BASH script to perform full-page scans, translate the format if required and dump the files into a directory. Configure an FTP daemon to see that directory. Easy!
$ convert -crop widthxheight+xoffset+yoffset filename1 filename2
If you want to overwrite the original then use the more dangerous
$ mogrify -crop widthxheight+xoffset+yoffset filename
This book sounds as though it is very necessary.
That's because the PS/2 mouse hardware generates interrupts; the processor has to stop what it's doing in order to see what was trying to interrupt it and decide if or not it's important enough to deal with there and then or it can wait. This fraction of a second is just long enough to relieve whatever condition was causing the processor to foul up. My guess, since it involves reading and writing large quantities of data in a contiguous fashion, is some sort of race hazard where if two logic inputs change state in the wrong order, the output is not what you were expecting. But I Am No Chip Designer.
Ubuntu have source repositories, you know. An Ubuntu source package {and for that matter, a Debian source package, since Ubuntu is based on Debian} consists of three files. The orig.tar.gz file {as supplied by the author}, the diff.gz file {containing various tweaks required to customise the source tree and Makefile to Debian / Ubuntu's way of doing things} and the .dsc file {containing various metadata}. You can use apt-get or aptitude to download the source files, and dpkg to build a .deb package and then install this as normal. I don't think there's a way to bypass the .deb creation stage and go straight to installing the freshly-compiled binaries; but I'm sure that someone will correct me soon enough, if there is one.
What really makes the difference between a good distribution and a great distribution is the size and quality of the package repository. Problems with repositories might be: packages completely absent, straightforward bugs or unmentioned dependencies -- i.e. where a package foo depends on a package bar but this is not included in the dependency metadata, and all the developers happen to be running bar anyway so nobody notices. This is actually the worst of all potential repository issues, as those most likely to be in a position to notice it are inexperienced users {otherwise they probably would have installed the dependency anyway with something else} and so not used to the whole dependency system {it does take a bit of getting your head round at first, though it's obvious once you come to understand it}.
Back in the day, only a minority of people could afford recording equipment. They soon found that they couldn't make enough money out of charging artists to record albums which would then become the artist's property that they could sell to the public, so they came up with another model: get the artist to sign over their rights -- in exchange for a one-off payment -- to the recording company, who would take care of the whole business of selling records and arrange for the artist to be paid a cut from each one sold.
It's this exclusivity that's the beginning and end of the problem -- the fact that once an artist is having their recordings distributed by one label, no other label can distribute their recordings.
Most other things are available from several sources. For instance, I can buy a loaf of bread baked by Sunblest, Kingsmill or Warburtons; or I can buy flour and yeast and bake my own. There is, in principle, no artificial barrier to a new player entering the market; if their product offers value, as judged by those who buy it, then it will be successful. "Value" is of course a nebulous concept, and so it should be; but in this case it is likely to mean a loaf of bread that tastes better, or costs less than what is already available.
The same holds true for things such as standardised industrial components. If I'm buying M4 x 20 steel bolts with a raised Posidriv head, or 4.7k ohm 0.25 watt resistors, or 80 gsm A4 paper in packs of 500 sheets, or 15mm. end-feed plumbers' tee-pieces, I still have a choice of suppliers; but there is no subjective assessment of value, and anyone's part will do the same job as anyone else's.
I think there is a market about to emerge for a new way of running a record label. The steps would be as follows:
There's no reason in principle why an artist shouldn't have the same album available on different labels, possibly even selling at different RRPs, at the same time; they would be competing with one another strictly on the basis of the services they offered {studio facilities, in-house producers and session musicians, pressing capacity and so forth}. Some labels would specialise in quick turnaround at a slight premium. Supermarkets probably would have their own record pressing operations, but no studio facilities -- they would work strictly from masters.
What is really required is for the law to state unequivocally, once and for all that, as long as you own a recording on a legitimately-acquired medium sanctioned by the copyright holder, your "fair dealing" rights include the right to make an unlimited number of copies of that recording in alternative formats for your own use, and to perform any necessary step in the process: copyright would not be infringed unless you used a copy you had made in some way that you would not be permitted to use the original.
What would be even better would be a ban on DRM systems that prevent absolutely the exercise of Fair Dealing rights and/or copying under Special Licence {e.g., I have permission from the band Ocean Colour Scene to make copies of any of their work for my own use; any DRM system that does not take this into account, perhaps by requiring a password to enable copying, should be illegal}.
At the end of the day, you're damned if you do try to imitate Windows and you're damned if you don't. If you try to be like Windows, that can mean copying inherently broken behaviour -- and will lose you friends in the "keep it pure" camp. If you don't try to be like Windows, somebody will complain that Linux is "too hard" {i.e. "not like Windows"}. If you try to make an application finely customisable, you end up driving people away because it's "hard to use"; if you don't include options to change things, you end up accused of "dumbing down".
I can only really attribute the "problem" to Microsoft's dominance in the marketplace combined with the popular mindset, which deems that "ignorance is bliss" and eschews learning to do something very hard, very well in favour of instant gratification with a half-arsed job.
That's why I think it's important for distributions to specialise. At the moment we have Ubuntu and Mandriva for people who want everything easy; Slackware and Debian for server administrators who feel the need to ride the metal; and Red Hat and SUSE for people who would rather pay someone else to do the donkey work. Not to mention hundreds if not thousands of less well-known distributions, catering to niche markets {self-booting mini-CDs, distributions tailored for antique hardware, retro gaming kits, movies on a self-booting CD, Linux on a USB stick and so forth}. One distribution simply can't be all things to all people.
One thing I would like to see would be a GUI front-end to the configure, make, make install process. It's distribution-agnostic, sometimes even architecture-agnostic. Now that processor power is so cheap, the only compelling reason not to compile locally has been mitigated. A graphical front-end would look a little bit like a Windows InstallShield installer. What puts people off source tarballs isn't so much the idea of compilation {though that's where they will inevitably transfer the blame}, as the thought of unresolved dependencies breaking the process. There's no reason why a properly-put-together automake/autoconf package should not be able to detect everything it needs at the configure stage. Linux allows you to mix and match libraries to an extent; so if a particular application imperatively requires newer libraries than are already installed, that need not be a problem. The installer should be able to determine for itself whether it's possible to download and install its own dependencies, and proceed automatically if it is safe to do so.
Of course, probably before the GUI source installer goes mainstream, we will need a reliable developer tool for creating self-installing packages; analysing libraries and creating a dependency database. Although this sounds like a huge effort, it probably will be more likely to succeed than any attempt towards achieving cross-distribution binary compatibility; binaries were never really meant to be compatible, source was always meant to be compatible.
That's hardly surprising in a metropolitan area. Whenever one train has to be re-platformed, the train that was going to use the platform that train is now using also has to be re-platformed, and so on, until there is a big enough time window to get one train out of the way before the xext comes in.
:)
It's worse in the South because there are two electrification systems in use; the old Southern Electric, third-rail DC system and the modern, overhead AC system. Not all vehicles are dual-powered, and neither are all tracks, so re-platforming options are limited. Of course, since {as every Londoner knows} there is no electricity in the North of England, this will not be a problem in Leeds
Further proof that the US patent office needs closing down NOW.
But what if someone wants to use \0 in a string?