Sorry, but I don't agree with you for one instant. Management often do not know what is best for a company. Sometimes, it's up to the poor sods who actually do the work -- you know, the ones who stand there waiting for a bus (which will be late, due to some wanker sitting on his own in a brand new BMW cutting it up on the A52, and you will eventually end up punishing them for it) while you drive past in a car that cost more than their house, welcome to Thatcher's Classless Society -- to take matters into their own hands for the greater good.
There's a simple principle at play here; people who are good at their jobs get promoted, till they are in a job they can't do very well; then they stay there. The solution would be either to demote them to the last job they were any good at, or promote someone over their head; either of which presupposes the existence of a very good manager at a very high level.
That's exactly how MySQL works -- rely on the application layer to do sanity-checking and even (to some extent) result-set filtering (you get the LIKE operator with the % wildcard, but it's nowhere as comprehensive as Perl's regular expressions).
Where [Microsoft] have the biggest problem is with GPL'd stuff, which they can't use at all.
Won't, not can't. Just like most vegetarians could eat meat, they just don't want to.
Microsoft would be welcome to use as much GPL code as they liked, as long as they respected the requirements of the GPL. However, they don't want to do this, because they are in denial. For Microsoft to use GPL code would be to admit that someone else has written better software than they could, and that would raise all manner of awkward questions.
Apache figured out long ago that a separate process for each request is a more logical way of doing it. Anything else implies you don't trust the process scheduler to do its job properly. In which case, the proper response is to replace the system-wide process scheduler with an improved version; not to try bypassing it within your own application. Not only is that downright selfish, but you will not benefit when somebody eventually does improve init.
PostgreSQL is licenced under a BSD-style licence. This means that closed-source forks might be permitted (when the BSD licence was first drawn up, binary-only distribution was inconceivable; binary compatibility even between machines of the same make and base model but configured differently was poor). I say "might be" because the BSD licence seems to allow you to distribute the source code even if you have only received the binary code; which might give you the option to recover the source code using reasonable force. This remains untested TTBOMK.
Microsoft could make some "improvements" to Postgres to make it "run better" under Windows, but keep those improvements Windows-only and binary-only..... and end up with a Postgres version that is subtly incompatible with everyone else's, making it harder to migrate to an alternative OS. Even integrating Postgres support tightly into one of their own mickey-mouse programming languages would end up tying users to that language.
I have heard of at least one company where the IT department were ordered to set up IIS, ASP and MS SQL server for a management-initiated project (one of the bosses had taken some mickey-mouse correspondence course and fancied himself as a designer of database-driven web sites). What they actually did was set up a small test server for Gates's toady; pocket the rest of the money meant for Microsoft; set up Linux, Apache, PHP and MySQL on the outward-facing server; make some flimsy security-related excuses why the boss shouldn't have direct access to the outward-facing server; and translate all the well-meaning-but-terminally-incompetent boss's badly-written ASP code into PHP.
I can already hear the entire MySQL fanbase chanting in unison, "What do you expect? Postgres IS slow!" But on the other hand, Postgres is a proper relational database server, not merely a variable-persistence layer with an SQL-like syntax.
Many applications that were originally developed on Linux tend to run slower on other OSs -- or even on Linux, with file systems other than ext2. This is mainly due to Linux' ext2 file system, with its write-caching policy; which was basically "never, ever commit anything at all to disk until we need the RAM for something else or are about to reboot -- we can serve read requests from the write cache if needs be".
As a result of this caching policy, it's entirely possible to create a file in one application (which, if you've enough free RAM, will merely place it in the write-cache), read it (straight from the write-cache) into another application, and then delete it without it ever being written to disk!
Totally. Heroin withdrawal -- any kind of opiate withdrawal, for that matter, be it morphine, codeine, dihydrocodeine, methadone or any of myriad others developed under the misapprehension that it would be a less-addictive alternative to the last one -- is intensely physically painful, because a heroin habit interferes with the body's pain-regulation mechanism, reducing the production of endorphins (the body's natural painkillers and feelgood chemicals). Nicotine addiction does not compromise the pain-regulatory mechanism.
Nonetheless, the legal status will contribute its own systemic error to the results of any comparison, artificially understating the addictiveness of illegal drugs. There is more to be gained by a medical practitioner "curing" a person of an addiction to an illegal drug than a legal one, and more to be gained by an ex-addict staying (apparently) clean of an illegal drug than a legal one.
The fact that the aggrieved party may have been committing a crime by using the software without authorisation does not alter anything. Two wrongs do not make a right. Deleting files from a user's home directory goes above and beyond reasonable force and is a criminal offence punishable by five years' imprisonment and/or a fine.
That was what I was thinking -- it's just the same as the War On Some Drugs. Although most recreational drugs theoretically should cost pennies per dose (poppies, cannabis, hallucinogenic mushrooms and cacti, coca and valerian all grow wild, and they're just the ones I can think off off the top of my head), the very fact that they are illegal introduces artificial scarcity and allows dealers to control prices. And such "legal" ways of getting high as there are, are a PITB. There are medicines you can buy from a pharmacy that will get you off your tits (e.g. Paramol {Paracetamol and Dihydrocodeine}; Benylin Chesty Coughs -- two drugs in one really, Original {Diphenhydramine} is a downer, Non-Drowsy {Guaifenesin} is a mild upper; Night Nurse {diphenhydramine, same ingredient as Benylin Original} and the perennial standby, Kaolin and Morphine mixture -- worth faking a tummy ache to be given a dose of) if you take enough of them, and of course there's booze..... but getting p!$$&d really isn't quite the same thing. It's too dirty a "high". There are legal plant extracts but the reason that most of them haven't been banned is that they aren't really much cop (though Sida Cordifolia isn't bad..... name's a bit off-putting if you speak French though).
Most of the crime is created in response to the problem of illegality. Junkies steal to buy heroin because it's sold at vastly inflated prices by dealers, they daren't seek help for fear of dropping their mates in the s#!t, and anyway they're already criminals just for having a toot so what's a bit of thieving between friends? Tobacco is more addictive than heroin (to the extent you can compare an illegal drug with a legal one), yet smokers are generally law-abiding. Apart from the ones who are bleeding the National Health Service dry by buying tobacco abroad..... we should send them to Belgium to get treated if they get cancer..... but I digress.
Microsoft's paid shills -- and a few people so stupid, they even shill for Microsoft without getting paid for it* -- will constantly tell you that "Open Source isn't the answer to everything". That's only half right; computers aren't the answer to everything.
Basically, you need to understand how a process works before you can think about automating it. There's no point having the ability to produce a neat typed document, if you can't construct a proper sentence. And you don't need a computer to know your verbs from your adjectives; in fact, pretty much anything you try to "computerise" learning some of the real fundamentals is going to detract from the meaning. Shooting the adjectives out of a sentence with a laser gun on a BBC Model B (Hmm, is that MODE 2 or MODE 5? That's a user-defined character, I wonder what the numbers would be? What a terrible sound effect, I could do better! Oh, is it my turn?) just isn't as conducive to learning as underlining the adjectives on a Banda'd A4 sheet.
* On mentioning LAMP on another forum, I was immediately reminded that "You can run Apache, PHP and MySQL on Windows". Yes you can. And it's a shitload more effort than just typing apt-get install apache2 php mysql.
Things which don't have evolutionary purposes tend to get diluted and obliterated. The fact that anti-social behaviour has not succumbed to this fate suggests that it may be in some way useful. And I never said evolution was sentient. A population's success is "cared about" implicitly, as a matter of definition; any population which does not succeed, perishes and is forgotten.
Try reading something about evolution that wasn't written by ID advocates, will you?
Actually, natural selection favours altruism. Game theory, as applied to a population of predators who live and hunt as a pack, suggests that the population's survival chances are improved by co-operation within the group. Example: You kill something which is too big for you to eat all by yourself. If you share it with the others, everyone gets fed and the population's overall viability goes up. If not, it goes to waste and depletes the population of prey, which could be deleterious to the group; or worse, it attracts predators.
That being said, anti-social personality disorder probably did once serve some important evolutionary function. Either by providing a self-destruct mechanism to prevent populations from becoming too successful, or by providing a legitimate target within the population.
Any kind of BASIC -- except, possibly, one of the British dialects such as Spectrum BASIC (would need to get rid of LET, though, now we are no longer using single-key entry) or BBC BASIC (one of the later incarnations with ON... PROC) -- would be bad enough, but VB? Why?
Linux isn't "stealing Microsoft's candy". A better analogy would be that Linux is eating fruit that grows on trees everywhere, instead of buying Microsoft's expensive proprietary candy.
Agree with you that closed source will soon be extinct. But it will be the development of a decompiler that puts the nail in the coffin. Whoever spoke with a straight face of abolishing slavery, before the time of James Watt?
Yeah, but before there was binary compatibility, there was ZERO possibility of virus propagation.
If we brought back deliberate binary incompatibility, but in a controlled way (i.e., enforced source compatibility; well-packaged source tarballs build anywhere and ones that don't are badly packaged), and insisted for compiling an application to require a deliberate act on the part of the user, there would be no more malware.
It's one thing to talk of "choice", but the whole point is that there are some things you don't have a choice over. Using filament bulbs is poisoning babies. Anyone with a brain ought to be able to see that. Banning filament bulbs -- or even taxing them heavily, something like about 5 GB pence a watt -- would be a great first step. The market will adapt, sooner or later. There will be some kind of retrofit bulb available for the weirdy screw-fitting ones they have in fridges, sewing machines, ovens and the like.
Wasn't it Australia where they read out the weather forecast in Fahrenheit one night, then switched to Celsius the next morning, and now nobody can even understand it when "bloody septics" talk about temperature? Sometimes a big, hard push is what it takes.
Please, don't use the "it's slow" argument. Anything worth doing is worth waiting for.
Suppose you had some feature in your kernel that would only allow it to run binaries compiled on that machine -- say, a signing key embedded in the compiler and the corresponding key embedded in the kernel. Suppose also that the compiler is on a filing system that isn't normally mounted, requiring you to do something deliberate to enable it; or it will only accept source code that you have signed. The important thing in either case is a deliberate act on your part.
Malware simply wouldn't have a chance to spread in such an environment. It couldn't spread as a compiled binary because binaries compiled on other machines won't run on your machine. It would have to spread as source code, thus exposing it to people on The Right Side Of The Fence. And it couldn't get itself compiled by accident, because enabling the compiler requires a deliberate act. Its propagation would stop with the first person who didn't just keep blindly clicking on "OK" without reading what the warnings said.
Yeah, but the type of person who's building XOrg or Gnome to run an older system most probably either has a few days to spare, or can use another, faster machine for the compilation. I'm talking about the user with a nice modern system, the sort we're trying to tempt away from Windows Vista.
Good point, though, that BSD ports are patched -- so, for that matter, are Gentoo's ebuilds and Debian's source packages (that's why they come in three files); and even in the LFS documentation they suggest certain options to `configure`.
There is a unified package format, and that is the source code tarball.
People need to be disabused of the notion that there is anything bad about compiling on your own machine. Gentoo and FreeBSD prove that is not the case.
Slashdot Mirror
Sorry, but I don't agree with you for one instant. Management often do not know what is best for a company. Sometimes, it's up to the poor sods who actually do the work -- you know, the ones who stand there waiting for a bus (which will be late, due to some wanker sitting on his own in a brand new BMW cutting it up on the A52, and you will eventually end up punishing them for it) while you drive past in a car that cost more than their house, welcome to Thatcher's Classless Society -- to take matters into their own hands for the greater good.
There's a simple principle at play here; people who are good at their jobs get promoted, till they are in a job they can't do very well; then they stay there. The solution would be either to demote them to the last job they were any good at, or promote someone over their head; either of which presupposes the existence of a very good manager at a very high level.
Look. Get this into your head:
Microsoft REALLY ARE that bad!
That's exactly how MySQL works -- rely on the application layer to do sanity-checking and even (to some extent) result-set filtering (you get the LIKE operator with the % wildcard, but it's nowhere as comprehensive as Perl's regular expressions).
Microsoft would be welcome to use as much GPL code as they liked, as long as they respected the requirements of the GPL. However, they don't want to do this, because they are in denial. For Microsoft to use GPL code would be to admit that someone else has written better software than they could, and that would raise all manner of awkward questions.
Apache figured out long ago that a separate process for each request is a more logical way of doing it. Anything else implies you don't trust the process scheduler to do its job properly. In which case, the proper response is to replace the system-wide process scheduler with an improved version; not to try bypassing it within your own application. Not only is that downright selfish, but you will not benefit when somebody eventually does improve init.
PostgreSQL is licenced under a BSD-style licence. This means that closed-source forks might be permitted (when the BSD licence was first drawn up, binary-only distribution was inconceivable; binary compatibility even between machines of the same make and base model but configured differently was poor). I say "might be" because the BSD licence seems to allow you to distribute the source code even if you have only received the binary code; which might give you the option to recover the source code using reasonable force. This remains untested TTBOMK.
..... and end up with a Postgres version that is subtly incompatible with everyone else's, making it harder to migrate to an alternative OS. Even integrating Postgres support tightly into one of their own mickey-mouse programming languages would end up tying users to that language.
Microsoft could make some "improvements" to Postgres to make it "run better" under Windows, but keep those improvements Windows-only and binary-only
I have heard of at least one company where the IT department were ordered to set up IIS, ASP and MS SQL server for a management-initiated project (one of the bosses had taken some mickey-mouse correspondence course and fancied himself as a designer of database-driven web sites). What they actually did was set up a small test server for Gates's toady; pocket the rest of the money meant for Microsoft; set up Linux, Apache, PHP and MySQL on the outward-facing server; make some flimsy security-related excuses why the boss shouldn't have direct access to the outward-facing server; and translate all the well-meaning-but-terminally-incompetent boss's badly-written ASP code into PHP.
.....
It makes me think of this song
I can already hear the entire MySQL fanbase chanting in unison, "What do you expect? Postgres IS slow!" But on the other hand, Postgres is a proper relational database server, not merely a variable-persistence layer with an SQL-like syntax.
Many applications that were originally developed on Linux tend to run slower on other OSs -- or even on Linux, with file systems other than ext2. This is mainly due to Linux' ext2 file system, with its write-caching policy; which was basically "never, ever commit anything at all to disk until we need the RAM for something else or are about to reboot -- we can serve read requests from the write cache if needs be".
As a result of this caching policy, it's entirely possible to create a file in one application (which, if you've enough free RAM, will merely place it in the write-cache), read it (straight from the write-cache) into another application, and then delete it without it ever being written to disk!
Totally. Heroin withdrawal -- any kind of opiate withdrawal, for that matter, be it morphine, codeine, dihydrocodeine, methadone or any of myriad others developed under the misapprehension that it would be a less-addictive alternative to the last one -- is intensely physically painful, because a heroin habit interferes with the body's pain-regulation mechanism, reducing the production of endorphins (the body's natural painkillers and feelgood chemicals). Nicotine addiction does not compromise the pain-regulatory mechanism.
Nonetheless, the legal status will contribute its own systemic error to the results of any comparison, artificially understating the addictiveness of illegal drugs. There is more to be gained by a medical practitioner "curing" a person of an addiction to an illegal drug than a legal one, and more to be gained by an ex-addict staying (apparently) clean of an illegal drug than a legal one.
This is illegal in the UK. It quite clearly falls under Section 3 of the Misuse of Computers Act 1990.
The fact that the aggrieved party may have been committing a crime by using the software without authorisation does not alter anything. Two wrongs do not make a right. Deleting files from a user's home directory goes above and beyond reasonable force and is a criminal offence punishable by five years' imprisonment and/or a fine.
That was what I was thinking -- it's just the same as the War On Some Drugs. Although most recreational drugs theoretically should cost pennies per dose (poppies, cannabis, hallucinogenic mushrooms and cacti, coca and valerian all grow wild, and they're just the ones I can think off off the top of my head), the very fact that they are illegal introduces artificial scarcity and allows dealers to control prices. And such "legal" ways of getting high as there are, are a PITB. There are medicines you can buy from a pharmacy that will get you off your tits (e.g. Paramol {Paracetamol and Dihydrocodeine}; Benylin Chesty Coughs -- two drugs in one really, Original {Diphenhydramine} is a downer, Non-Drowsy {Guaifenesin} is a mild upper; Night Nurse {diphenhydramine, same ingredient as Benylin Original} and the perennial standby, Kaolin and Morphine mixture -- worth faking a tummy ache to be given a dose of) if you take enough of them, and of course there's booze ..... but getting p!$$&d really isn't quite the same thing. It's too dirty a "high". There are legal plant extracts but the reason that most of them haven't been banned is that they aren't really much cop (though Sida Cordifolia isn't bad ..... name's a bit off-putting if you speak French though).
..... we should send them to Belgium to get treated if they get cancer ..... but I digress.
Most of the crime is created in response to the problem of illegality. Junkies steal to buy heroin because it's sold at vastly inflated prices by dealers, they daren't seek help for fear of dropping their mates in the s#!t, and anyway they're already criminals just for having a toot so what's a bit of thieving between friends? Tobacco is more addictive than heroin (to the extent you can compare an illegal drug with a legal one), yet smokers are generally law-abiding. Apart from the ones who are bleeding the National Health Service dry by buying tobacco abroad
Exactly.
Microsoft's paid shills -- and a few people so stupid, they even shill for Microsoft without getting paid for it* -- will constantly tell you that "Open Source isn't the answer to everything". That's only half right; computers aren't the answer to everything.
Basically, you need to understand how a process works before you can think about automating it. There's no point having the ability to produce a neat typed document, if you can't construct a proper sentence. And you don't need a computer to know your verbs from your adjectives; in fact, pretty much anything you try to "computerise" learning some of the real fundamentals is going to detract from the meaning. Shooting the adjectives out of a sentence with a laser gun on a BBC Model B (Hmm, is that MODE 2 or MODE 5? That's a user-defined character, I wonder what the numbers would be? What a terrible sound effect, I could do better! Oh, is it my turn?) just isn't as conducive to learning as underlining the adjectives on a Banda'd A4 sheet.
* On mentioning LAMP on another forum, I was immediately reminded that "You can run Apache, PHP and MySQL on Windows". Yes you can. And it's a shitload more effort than just typing apt-get install apache2 php mysql.
Things which don't have evolutionary purposes tend to get diluted and obliterated. The fact that anti-social behaviour has not succumbed to this fate suggests that it may be in some way useful. And I never said evolution was sentient. A population's success is "cared about" implicitly, as a matter of definition; any population which does not succeed, perishes and is forgotten.
Try reading something about evolution that wasn't written by ID advocates, will you?
Actually, natural selection favours altruism. Game theory, as applied to a population of predators who live and hunt as a pack, suggests that the population's survival chances are improved by co-operation within the group. Example: You kill something which is too big for you to eat all by yourself. If you share it with the others, everyone gets fed and the population's overall viability goes up. If not, it goes to waste and depletes the population of prey, which could be deleterious to the group; or worse, it attracts predators.
That being said, anti-social personality disorder probably did once serve some important evolutionary function. Either by providing a self-destruct mechanism to prevent populations from becoming too successful, or by providing a legitimate target within the population.
Why don't you just get an all-regions DVD player? But note that you'll need a TV that can handle 625x50, and it must have RGB inputs.
Any kind of BASIC -- except, possibly, one of the British dialects such as Spectrum BASIC (would need to get rid of LET, though, now we are no longer using single-key entry) or BBC BASIC (one of the later incarnations with ON ... PROC) -- would be bad enough, but VB? Why?
Linux isn't "stealing Microsoft's candy". A better analogy would be that Linux is eating fruit that grows on trees everywhere, instead of buying Microsoft's expensive proprietary candy.
Agree with you that closed source will soon be extinct. But it will be the development of a decompiler that puts the nail in the coffin. Whoever spoke with a straight face of abolishing slavery, before the time of James Watt?
Yeah, but before there was binary compatibility, there was ZERO possibility of virus propagation.
If we brought back deliberate binary incompatibility, but in a controlled way (i.e., enforced source compatibility; well-packaged source tarballs build anywhere and ones that don't are badly packaged), and insisted for compiling an application to require a deliberate act on the part of the user, there would be no more malware.
By the way, are Aussie light bulbs push-and-twist like ours, or are they screw-fit?
I can honestly say I'm in favour of this.
It's one thing to talk of "choice", but the whole point is that there are some things you don't have a choice over. Using filament bulbs is poisoning babies. Anyone with a brain ought to be able to see that. Banning filament bulbs -- or even taxing them heavily, something like about 5 GB pence a watt -- would be a great first step. The market will adapt, sooner or later. There will be some kind of retrofit bulb available for the weirdy screw-fitting ones they have in fridges, sewing machines, ovens and the like.
Wasn't it Australia where they read out the weather forecast in Fahrenheit one night, then switched to Celsius the next morning, and now nobody can even understand it when "bloody septics" talk about temperature? Sometimes a big, hard push is what it takes.
Please, don't use the "it's slow" argument. Anything worth doing is worth waiting for.
Suppose you had some feature in your kernel that would only allow it to run binaries compiled on that machine -- say, a signing key embedded in the compiler and the corresponding key embedded in the kernel. Suppose also that the compiler is on a filing system that isn't normally mounted, requiring you to do something deliberate to enable it; or it will only accept source code that you have signed. The important thing in either case is a deliberate act on your part.
Malware simply wouldn't have a chance to spread in such an environment. It couldn't spread as a compiled binary because binaries compiled on other machines won't run on your machine. It would have to spread as source code, thus exposing it to people on The Right Side Of The Fence. And it couldn't get itself compiled by accident, because enabling the compiler requires a deliberate act. Its propagation would stop with the first person who didn't just keep blindly clicking on "OK" without reading what the warnings said.
Yeah, but the type of person who's building XOrg or Gnome to run an older system most probably either has a few days to spare, or can use another, faster machine for the compilation. I'm talking about the user with a nice modern system, the sort we're trying to tempt away from Windows Vista.
Good point, though, that BSD ports are patched -- so, for that matter, are Gentoo's ebuilds and Debian's source packages (that's why they come in three files); and even in the LFS documentation they suggest certain options to `configure`.
There is a unified package format, and that is the source code tarball.
People need to be disabused of the notion that there is anything bad about compiling on your own machine. Gentoo and FreeBSD prove that is not the case.