Of course I'm not being so arrogant as to claim that my website is as valuable to the world as google! What I claim is that the value I provide to the web is a fair exchange for the value that the web provides to *me*. [This is only a specific example; the phrasing is clearer that way than if we start talking about parties 'A' and 'B']. Now, because data costs (almost) nothing to duplicate, the web can easily cope with a large population who haven't contributed (yet) with their own content. The gift of this data is freely given to all humanity. However the moral "price of admission to the web" is just to contribute, if you can. Please remember that there are millions of stakeholders in the web. Don't compare google to an individual author but to the set of all web authors.
P.S. As it happens, my personal web contribution is quite significant (thousands of hours). But even the people who do only put up a few simple pages are still contributing, and they *are* pulling their weight.
P.P.S. "IP" is an enshrinement of the *denial* of reciprocity in law.
There is a fundamental assumption of the web, which google seem to have misunderstood. It is this: "Anything you publish, I can use. In return, anything I publish, you can use".
for example, I make my website accessible to googlebot without restriction (including indexing, caching etc). In return, google is available to me. It's simply about fairness: the "price of entry to the Internet" is that one should contribute one's own material. This is how, for example, people share html layouts. The unfortunate thing is that this combination of reciprocity, fair-use and courtesy is not enshrined in law, and we persist in the ludicrous notion of "intellectual property".
Besides which, if google really want to do (and be seen to do) the right thing, they should offer gaia a blanket license. Fortunately, gaia is free-software, and it will get forked if necessary. It's time google had some stiff competition.
I think that moving the kernel to GPL3 is actually practical.
a)Most contributors can be tracked (we already know where code came from, thanks to SCO)
b)Those that can't be found will probably still hear about the relicensing, and get in touch. I expect most
of the small contributors to be willing to go along with the change.
c)For a very small number, code could be re-written.
I don't think I follow. What your postulating is that some Nasty,Inc could extort Linus' private key from him by this means? I think you're wrong on 2 counts:
i)If Bar,Inc make hardware which contains the binary, then it is Bar,Inc upon whom the obligation of key-distribution rests. Besides which, Linus doesn't tend to sign random binaries with his key! Signing the source would never be a problem.
ii)Linus's main beef is that he doesn't think "Tivoisation" is immoral, so doesn't wish to prevent it. In his view, as long as Tivo share their own source, that's sufficient, and he doesn't want them to be obliged to produce hardware whose software anyone else can modify. I think that's a short-sighted view - I don't care about the TiVo, but I'd be really cross if we get "Trusted(*) PCs" which can only run signed binaries.
(*)Note: this depends on who signed them. I'm all in favour of signed binaries (like rpms) when it is under my control and protects against abuse, but if the BIOS prevented me from running arbitrary code when I *explicitly* want it to, that's Bad.
I agree, Linus has misunderstood GPL3. I don't think I'd call it stupidity though. He originally made 2 arguments; as I understand, the first was a simple error of fact which he no longer pursues:
1)GPL3 makes you give up all your private keys => "Developers would never have any privacy". This isn't true - you only have to give up a private key which you used to sign code, if the hardware will only boot that signed code. This is essentially a ban on hardware which runs open code but cannot be modified.
2)GPL3 will impose conditions on hardware manufacturers, making it less free (in a BSD-sense), and therefore less useful to them. But kernel developers shouldn't try to impose usage restrictions. This seems a fair point - but it underestimates the vital importance of "Free as in Freedom". Stallmann hasn't yet been wrong on this, although sometimes it takes a decade for him to be proved right.
I think GPL3 is inevitable, and even the "open source" people who don't fundamentally care about politics will come to need it. Fortunately, we do have 2 alternatives, should we need them: GNU/BSD (or GNU/Darwin) and GNU/HURD. I hope it doesn't come to that, though!
I don't understand the reason for the new user interface? It can't be due to resource limits, since they already ship GTK (with firefox), and it can't be for reasons of complexity, since most kids pick up computing skills really fast. So why re-write all the apps? I think that it's quite wasteful of programmer effort. Wouldn't it be better to work on reducing the resource requirements(*) of Gnome (refactor it to make some parts of it compile-time options)? Existing minimal linux distros (eg DSL) are very good at providing good programs with low resource requirements; they just need some "tidying" to make them more user-friendly.
* For lower resources still, use the excellent IceWM. But if we already have firefox, then we've already loaded GTK, and may as well use Gnome. Anyonw who wants a demonstration of how fast simpler programs can be, please try Dillo.
Saddam is evil. No question of that. Likewise, even though his trial doesn't appear to have been strictly fair, he is unambiguously guilty, and, morally, he deserves to die. BUT it doesn't make it right. Killing in cold blood is wrong, even when done by the State. I think he ought to be imprisoned for life, but in a neutral country, so that he cannot be a figurehead for any Iraqi dissidents. Furthermore, I think that Historical Truth is more important than Justice, so all the trials on all the charges must be completed.
Incidentally, the US Government have done plenty of things which, by the standards of the Nuremberg trials, would leave the leaders to face the death penalty.
1)I'd suggest *only* permitting Firefox, and removing the I.E. icons. [The smart students can still run iexplore.exe if they really want it, but it prevents most people getting attacked by IE exploits.]. Then install adblock + filtersetG updater.
2)Trust your students. Most of the problems result from honest mistakes - therefore, you need to make it easy to revert to a previous state, but not to prevent experiments. My view is you should avoid all censorship; if you feel you must impose rules, then punish students for what you see them doing, rather than asking the machine to do it.
3)Have a system where the brighter students may request greater privileges. And/or a few test machines that are less locked down. If some of them want to try Knoppix, then let them. Maybe even have one machine which the students can administrate - you may be pleasantly surprised. On that machine, if they want to install Ubuntu, go for it. You'll have some able assistants in no time!
4)Don't even think of monitoring students with keystroke loggers etc. It's just wrong. It sets a bad example, and will only lead to heartache. And you wouldn't catch the really bad ones anyway.
Under Linux, the easiest thing to do is to encrypt the swapfile. It's trivial - just modify/etc/fstab:/dev/hda6 swap swap defaults,loop=/dev/loop0,encryption=AES256 0 0 and requires no further user-intervention: a new key is automatically generated at boot, and discarded on shutdown. The advantage of this is that *anything* could get swapped out, and would remain there for some time. This is true even if you deleted the original file (or it was in RAM, and never saved, such as, perhaps, a password/phrase). This is going after the low-hanging fruit, but it's a big improvement over nothing, and it does protect against bad luck and trivial forensic analysis of the filesystem.
The grounding issue is a side effect, which isn't really related to "DC vs AC". To be in any danger, you must always be completing a circuit (example: you can safely touch a Van der Graaf generator, provided you stand on an insulator).
Some (but not all) DC circuits "float", i.e. neither end of the supply is earthed. Therefore, touching just one wire (+ earth) can't harm you. On the other hand, many DC circuits earth the negative or intermediate rail.
Most AC (1 or 3-phase) circuits (except for isolation transformers, used on building sites) have the neutral connection grounded. Therefore, touching a live wire will usually hurt you.
This all assumes that you are grounded fairly well, which is usually true: either to earth, or through a metal chassis.
As to whether an AC or DC shock is worse, I can't comment. [But remember, 120V AC is actually nearer 165V peak voltage]
A thought: if MSOffice already runs on MacOS, why can't we make a thin API for Linux and emulate a Mac? Presumably, it would be easier than emulating Win32 - although the Wine Project does this very well. Actually, you can already run Office on Linux, via Wine.
Stallman is, as usual, right - and has been urging this. The trouble is that the demand is insufficient, because there are already free 2d drivers, and the 3d binaries work "well enough" that customers aren't clamouring for alternatives. The Open Graphics project will help a bit.
I agree with you. The thing is, the HDAPS sensor has documentation... I'd love to buy a decent card with open drivers. I'd even contribute $100 to the development fund, it there were a serious attempt to reverse engineer it. However, as I understand it, graphics drivers are very hard to reverse engineer.
The whole point of the internet is that it is a global medium for free speech. Therefore, that which is permitted *anywhere* in the world (be it politics, gambling, or pornography) may be accessed from anywhere else. It's high time our law-makers realised, this, and stopped trying to regulate the net. Free speech is binary: you have it, or you don't - and there is no half-way area. So, someone somewhere doesn't like $X on the web. They don't have to look at it!
That's isn't really a a fair comparison. When Windows removes a Linux bootloader, you are left with a system that can *only* boot Windows. When Linux removes the Windows bootloader, it installs grub/lilo so that you can boot the OS of your choice. Last time I checked, the Windows bootloader wasn't capable of booting Linux.
Yes - but I have a dual monitor setup. 2 x 1600x1200 LCDs, both connected via DVI to the same card. The VGA/VESA drivers make them both display the same thing! So does the xorg "nv" driver.
Indeed. Nvidia's drivers are a royal pain to install, but at least they are pretty good when running. That said, as soon as anyone makes a card with proper GPL drivers that can do dual-head,DVI,3200x1600, I shall switch. I like 3D accel, but I'd be quite happy with an unaccelerated frame-buffer - just as long as is stable. The OpenGraphics project is trying....
PHP protects you by default from any SQL injection attack by using magic_quotes.
You can write things like this:
$sql="SELECT value FROM REGISTRY WHERE name='$name'"; perfectly safely.
Even better, it saves you having to bother with string escaping, and if you make a mistake, it is guaranteed to fail safe. The only downside is that you must remember to use stripslashes() if your user-data does NOT go via the database; however the worst that can happen here is some harmless user-facing ugliness like this:
Hello again, Mr O\'Neil Which is ugly to read, but will never create an attack risk.
You do still have to be careful if you data uses the DB *twice*, eg:
User Input ->PHP -> Database -> PHP(*) -> Database -> PHP -> User output In this case, the PHP script marked (*) must re-do the addslashes().
Yes - but you could reasonably hope to destroy the flash device - either a high current, or just by crunching it between a pair of pliers! Flash disks could be designed with a self-destruct circuit on the chip. But the best bet: encrypt the disks, use a temporary key, stored in RAM; when the system powers off, the key is lost.
Wouldn't it be easier to use a flash memory chip? It's unlikely that more than a few GB would be needed. And destroying a flash chip is much easier. Or, just encrypt the data with the key in RAM. (Linux can already do this with swap - it's completely transparent to the user, and the key only lasts as long as the system remains running).
I do see your point. However, in the common case, where (almost) all data goes into a database, magic_quotes is really useful. I'd much rather risk a superfluous backslash than an SQL injection attack any day! And although the code I write gets a massive amount of use, and so far hasn't suffered even an unncessary backslash on the production server, I know that I do make errors, and I'd prefer them to be harmless.
Of course, if you don't like magic_quotes, turn it off.
But how do you guarantee your DB is safe against programmer errors? SQL injection is so easy to do, so nasty in scope, and so easy to spot (especially with an open source PHP app) that you have to assume it will get exploited if you make a single trivial error.
Yes, you do have to be careful with unusual data flow. However, it's no more work to remember to call stripslashes() occasionally than it would be to remember to call addslashes(); the benefit is that the failure mode is far less severe.
Lastly, remember that insertion into a database and subsequent retrieval already performs a stripslashes() implicitly.
Enlighten me as to why magic_quotes is a bad thing? It seems to me to be a very good thing, relied on not as a "safety net" but as a deliberate feature. After all, it just saves you a whole lot of addslashes(), and has the added advantage of failsafe, if you do do something daft. In my view, removing the feature is insane; fortunately you can easily replicate it.
P.S. just because I like magic_quotes doesn't mean I also approve of register globals. That was a mistake!
But he didn't do anything harmful. Hacking (in the sense of "just taking a look around, out of curiosity" may be illegal, but it shouldn't be. If anything, it's a good thing, because it keeps admins on their toes, and slightly less vulnerable to computer criminals.
BTW, people in the UK *are* judging the US on its actions. We've seen quite enough of 'the war on terror' to realise that the USA is not to be trusted with legal due process.
When I first bought a thinkpad (the 560) in 1997, it was the best machine on the market. The attention to detail was stunning (better than Apple). It was clear that the ThinkPads were the best, and everyone knew it. I've had 3 more since then - and they have all been excellent machines - and the A22p (which I bought 5 years ago) has a better keyboard/mouse/display than any laptop I have ever seen since!
BUT...I feel that IBM/Lenovo is going downmarket - and I wish they wouldn't. They seem to have lost confidence in their own designs, and are competing more on price, and less on quality. Examples of lost confidence include the addition of a Windows Key (making the fabulous keyboard slightly less expert-friendly), re-arranging the mouse buttons from 2+1 to 3 in a row (to accommodate a touchpad [it takes about an hour to learn to love a trackpoint, after which you'd never go back]), and the replacement of the velvet-black finish by the tougher, but less comfortable ABS. Also, IBM's Linux support seems to be wavering - the newer thinkpads don't have any Free 3D drivers, and there have been a few more hardware issues of late.
Please, Leneovo, put the price *up* by 20% and give us something special.
Slashdot Mirror
Of course I'm not being so arrogant as to claim that my website is as valuable to the world as google! What I claim is that the value I provide to the web is a fair exchange for the value that the web provides to *me*. [This is only a specific example; the phrasing is clearer that way than if we start talking about parties 'A' and 'B'].
Now, because data costs (almost) nothing to duplicate, the web can easily cope with a large population who haven't contributed (yet) with their own content. The gift of this data is freely given to all humanity. However the moral "price of admission to the web" is just to contribute, if you can. Please remember that there are millions of stakeholders in the web. Don't compare google to an individual author but to the set of all web authors.
P.S. As it happens, my personal web contribution is quite significant (thousands of hours). But even the people who do only put up a few simple pages are still contributing, and they *are* pulling their weight.
P.P.S. "IP" is an enshrinement of the *denial* of reciprocity in law.
There is a fundamental assumption of the web, which google seem to have misunderstood. It is this:
"Anything you publish, I can use. In return, anything I publish, you can use".
for example, I make my website accessible to googlebot without restriction (including indexing, caching etc). In return, google is available to me. It's simply about fairness: the "price of entry to the Internet" is that one should contribute one's own material.
This is how, for example, people share html layouts. The unfortunate thing is that this combination of reciprocity, fair-use and courtesy is not enshrined in law, and we persist in the ludicrous notion of "intellectual property".
Besides which, if google really want to do (and be seen to do) the right thing, they should offer gaia a blanket license. Fortunately, gaia is free-software, and it will get forked if necessary. It's time google had some stiff competition.
I think that moving the kernel to GPL3 is actually practical.
a)Most contributors can be tracked (we already know where code came from, thanks to SCO)
b)Those that can't be found will probably still hear about the relicensing, and get in touch. I expect most
of the small contributors to be willing to go along with the change.
c)For a very small number, code could be re-written.
I don't think I follow. What your postulating is that some Nasty,Inc could extort Linus' private key from him by this means? I think you're wrong on 2 counts:
i)If Bar,Inc make hardware which contains the binary, then it is Bar,Inc upon whom the obligation of key-distribution rests.
Besides which, Linus doesn't tend to sign random binaries with his key! Signing the source would never be a problem.
ii)Linus's main beef is that he doesn't think "Tivoisation" is immoral, so doesn't wish to prevent it. In his view, as long as Tivo share their own source, that's sufficient, and he doesn't want them to be obliged to produce hardware whose software anyone else can modify. I think that's a short-sighted view - I don't care about the TiVo, but I'd be really cross if we get "Trusted(*) PCs" which can only run signed binaries.
(*)Note: this depends on who signed them. I'm all in favour of signed binaries (like rpms) when it is under my control and protects against abuse, but if the BIOS prevented me from running arbitrary code when I *explicitly* want it to, that's Bad.
I agree, Linus has misunderstood GPL3. I don't think I'd call it stupidity though.
He originally made 2 arguments; as I understand, the first was a simple error of fact which he no longer pursues:
1)GPL3 makes you give up all your private keys => "Developers would never have any privacy".
This isn't true - you only have to give up a private key which you used to sign code, if the hardware will only boot that signed code. This is essentially a ban on hardware which runs open code but cannot be modified.
2)GPL3 will impose conditions on hardware manufacturers, making it less free (in a BSD-sense), and therefore less useful to them. But kernel developers shouldn't try to impose usage restrictions.
This seems a fair point - but it underestimates the vital importance of "Free as in Freedom". Stallmann hasn't yet been wrong on this, although sometimes it takes a decade for him to be proved right.
I think GPL3 is inevitable, and even the "open source" people who don't fundamentally care about politics will come to need it. Fortunately, we do have 2 alternatives, should we need them: GNU/BSD (or GNU/Darwin) and GNU/HURD. I hope it doesn't come to that, though!
I don't understand the reason for the new user interface? It can't be due to resource limits, since they already ship GTK (with firefox), and it can't be for reasons of complexity, since most kids pick up computing skills really fast. So why re-write all the apps? I think that it's quite wasteful of programmer effort. Wouldn't it be better to work on reducing the resource requirements(*) of Gnome (refactor it to make some parts of it compile-time options)? Existing minimal linux distros (eg DSL) are very good at providing good programs with low resource requirements; they just need some "tidying" to make them more user-friendly.
* For lower resources still, use the excellent IceWM. But if we already have firefox, then we've already loaded GTK, and may as well use Gnome. Anyonw who wants a demonstration of how fast simpler programs can be, please try Dillo.
Saddam is evil. No question of that. Likewise, even though his trial doesn't appear to have been strictly fair, he is unambiguously guilty, and, morally, he deserves to die. BUT it doesn't make it right. Killing in cold blood is wrong, even when done by the State. I think he ought to be imprisoned for life, but in a neutral country, so that he cannot be a figurehead for any Iraqi dissidents. Furthermore, I think that Historical Truth is more important than Justice, so all the trials on all the charges must be completed.
Incidentally, the US Government have done plenty of things which, by the standards of the Nuremberg trials, would leave the leaders to face the death penalty.
Mostly, I agree. However...
1)I'd suggest *only* permitting Firefox, and removing the I.E. icons. [The smart students can still run iexplore.exe if they really want it, but it prevents most people getting attacked by IE exploits.]. Then install adblock + filtersetG updater.
2)Trust your students. Most of the problems result from honest mistakes - therefore, you need to make it easy to revert to a previous state, but not to prevent experiments. My view is you should avoid all censorship; if you feel you must impose rules, then punish students for what you see them doing, rather than asking the machine to do it.
3)Have a system where the brighter students may request greater privileges. And/or a few test machines that are less locked down. If some of them want to try Knoppix, then let them. Maybe even have one machine which the students can administrate - you may be pleasantly surprised. On that machine, if they want to install Ubuntu, go for it. You'll have some able assistants in no time!
4)Don't even think of monitoring students with keystroke loggers etc. It's just wrong. It sets a bad example, and will only lead to heartache. And you wouldn't catch the really bad ones anyway.
Under Linux, the easiest thing to do is to encrypt the swapfile. It's trivial - just modify /etc/fstab: /dev/hda6 swap swap defaults,loop=/dev/loop0,encryption=AES256 0 0
and requires no further user-intervention: a new key is automatically generated at boot, and discarded on shutdown.
The advantage of this is that *anything* could get swapped out, and would remain there for some time. This is true even if you deleted the original file (or it was in RAM, and never saved, such as, perhaps, a password/phrase). This is going after the low-hanging fruit, but it's a big improvement over nothing, and it does protect against bad luck and trivial forensic analysis of the filesystem.
The grounding issue is a side effect, which isn't really related to "DC vs AC". To be in any danger, you must always be completing a circuit (example: you can safely touch a Van der Graaf generator, provided you stand on an insulator).
Some (but not all) DC circuits "float", i.e. neither end of the supply is earthed. Therefore, touching just one wire (+ earth) can't harm you. On the other hand, many DC circuits earth the negative or intermediate rail.
Most AC (1 or 3-phase) circuits (except for isolation transformers, used on building sites) have the neutral connection grounded. Therefore, touching a live wire will usually hurt you.
This all assumes that you are grounded fairly well, which is usually true: either to earth, or through a metal chassis.
As to whether an AC or DC shock is worse, I can't comment. [But remember, 120V AC is actually nearer 165V peak voltage]
A thought: if MSOffice already runs on MacOS, why can't we make a thin API for Linux and emulate a Mac? Presumably, it would be easier than emulating Win32 - although the Wine Project does this very well. Actually, you can already run Office on Linux, via Wine.
Stallman is, as usual, right - and has been urging this. The trouble is that the demand is insufficient, because there are already free 2d drivers, and the 3d binaries work "well enough" that customers aren't clamouring for alternatives. The Open Graphics project will help a bit.
I agree with you. The thing is, the HDAPS sensor has documentation...
I'd love to buy a decent card with open drivers. I'd even contribute $100 to the development fund, it there were a serious attempt to reverse engineer it. However, as I understand it, graphics drivers are very hard to reverse engineer.
The whole point of the internet is that it is a global medium for free speech. Therefore, that which is permitted *anywhere* in the world (be it politics, gambling, or pornography) may be accessed from anywhere else. It's high time our law-makers realised, this, and stopped trying to regulate the net. Free speech is binary: you have it, or you don't - and there is no half-way area. So, someone somewhere doesn't like $X on the web. They don't have to look at it!
That's isn't really a a fair comparison. When Windows removes a Linux bootloader, you are left with a system that can *only* boot Windows. When Linux removes the Windows bootloader, it installs grub/lilo so that you can boot the OS of your choice. Last time I checked, the Windows bootloader wasn't capable of booting Linux.
Yes - but I have a dual monitor setup. 2 x 1600x1200 LCDs, both connected via DVI to the same card.
The VGA/VESA drivers make them both display the same thing! So does the xorg "nv" driver.
Indeed. Nvidia's drivers are a royal pain to install, but at least they are pretty good when running. That said, as soon as anyone makes a
card with proper GPL drivers that can do dual-head,DVI,3200x1600, I shall switch. I like 3D accel, but I'd be quite happy with an unaccelerated frame-buffer - just as long as is stable. The OpenGraphics project is trying....
PHP protects you by default from any SQL injection attack by using magic_quotes.
You can write things like this:
$sql="SELECT value FROM REGISTRY WHERE name='$name'";
perfectly safely.
Even better, it saves you having to bother with string escaping, and if you make a mistake, it is guaranteed to fail safe. The only downside is that you must remember to use stripslashes() if your user-data does NOT go via the database; however the worst that can happen here is some harmless user-facing ugliness like this:
Hello again, Mr O\'Neil
Which is ugly to read, but will never create an attack risk.
You do still have to be careful if you data uses the DB *twice*, eg:
User Input ->PHP -> Database -> PHP(*) -> Database -> PHP -> User output
In this case, the PHP script marked (*) must re-do the addslashes().
Yes - but you could reasonably hope to destroy the flash device - either a high current, or just by crunching it between a pair of pliers!
Flash disks could be designed with a self-destruct circuit on the chip. But the best bet: encrypt the disks, use a temporary key, stored in RAM; when the system powers off, the key is lost.
Wouldn't it be easier to use a flash memory chip? It's unlikely that more than a few GB would be needed. And destroying a flash chip is much easier.
Or, just encrypt the data with the key in RAM. (Linux can already do this with swap - it's completely transparent to the user, and the key only lasts as long as the system remains running).
I do see your point. However, in the common case, where (almost) all data goes into a database, magic_quotes is really useful.
I'd much rather risk a superfluous backslash than an SQL injection attack any day! And although the code I write gets a massive amount of use, and so far
hasn't suffered even an unncessary backslash on the production server, I know that I do make errors, and I'd prefer them to be harmless.
Of course, if you don't like magic_quotes, turn it off.
But how do you guarantee your DB is safe against programmer errors? SQL injection is so easy to do, so nasty in scope, and so easy to spot (especially with an open source PHP app) that you have to assume it will get exploited if you make a single trivial error.
Yes, you do have to be careful with unusual data flow. However, it's no more work to remember to call stripslashes() occasionally than it would be to remember to call addslashes(); the benefit is that the failure mode is far less severe.
Lastly, remember that insertion into a database and subsequent retrieval already performs a stripslashes() implicitly.
Enlighten me as to why magic_quotes is a bad thing? It seems to me to be a very good thing, relied on not as a "safety net" but as a deliberate feature. After all, it just saves you a whole lot of addslashes(), and has the added advantage of failsafe, if you do do something daft. In my view, removing the feature is insane; fortunately you can easily replicate it.
P.S. just because I like magic_quotes doesn't mean I also approve of register globals. That was a mistake!
But he didn't do anything harmful. Hacking (in the sense of "just taking a look around, out of curiosity" may be illegal, but it shouldn't be. If anything, it's a good thing, because it keeps admins on their toes, and slightly less vulnerable to computer criminals.
BTW, people in the UK *are* judging the US on its actions. We've seen quite enough of 'the war on terror' to realise that the USA is not to be trusted with legal due process.
When I first bought a thinkpad (the 560) in 1997, it was the best machine on the market. The attention to detail was stunning (better than Apple). It was clear that the ThinkPads were the best, and everyone knew it. I've had 3 more since then - and they have all been excellent machines - and the A22p (which I bought 5 years ago) has a better keyboard/mouse/display than any laptop I have ever seen since!
BUT...I feel that IBM/Lenovo is going downmarket - and I wish they wouldn't. They seem to have lost confidence in their own designs, and are competing more on price, and less on quality. Examples of lost confidence include the addition of a Windows Key (making the fabulous keyboard slightly less expert-friendly), re-arranging the mouse buttons from 2+1 to 3 in a row (to accommodate a touchpad [it takes about an hour to learn to love a trackpoint, after which you'd never go back]), and the replacement of the velvet-black finish by the tougher, but less comfortable ABS. Also, IBM's Linux support seems to be wavering - the newer thinkpads don't have any Free 3D drivers, and there have been a few more hardware issues of late.
Please, Leneovo, put the price *up* by 20% and give us something special.
Or, just use mplayer. It handles *everything*