The difference is that with a limited set of operator symbols to choose from, it's very tempting to use one that bares little resemblance to what your operator actually does. There are substantially more plausible function names, so the temptation to use one that is inappropriate is smaller.
What's wrong with a C-style cast? Simply consider it a shorthand notation for a const_cast or reinterpret_cast, depending on situation.
Frankly, there's no better way of truncating the result of a floating point operation than x = (int) (f / 2.3);. and "reinterpret_cast" is just too verbose when what you really want is bits = *(int *).
Re:I call this the LineOfView (as in PoV) Problem
on
Tim Bray Says RELAX
·
· Score: 1
I think your problem is that you're using XML to perform the job of a relational database.
The problem with asking something like this on a site like slashdot is twofold:
1. Armchair lawyers. Sure, I'm one of 'em. I like to think I'm better than most, but it's very hard to tell who's reliable and who isn't. 2. This kind of thing varies from place to place. You need somebody local.
I've never played with proper model rocket engines like that, but I did build a little solid-fuel rocket once. Real home-made stuff: a 100ml aluminium drink can with ground matchheads as the fuel. It didn't go high, maybe only 10 metres or so, but it was carrying a rather heavy payload... it stuck to the stake that I launched it from and took that up with it, probably about a kilo or so.
That was well & truly cool.
(Note to others who might try this: use a thick transparent plastic shield to protect you from shrapnel in case the can explodes...)
I did something similar, although I connected a few machines up in a ring using the RS423 port (Tx on machine A -> Rx on B, Tx on B to Rx on C, and Tx on C to Rx on A). That was cool.:)
Err... OK. I'd like to know the author of that page's testing methodology, 'cause I really can't see how the 'faster_sqrtf' function presented could possibly take 426 cycles to execute: it's a memory reference followed by 3 standard integer operations. On the 1.2GHz Opteron described, it should take 3 cycles (during each of which there are 2 spare execution units for out-of-order execution of other code) plus a memory access if there's a cache miss, which with 133MHz memory should take 9 cycles, for a grand total of 12 cycles. I'm assuming the code is already in the cache, because in any situation where you care that much it will be. It'll also be inlined into the code that's calling it, so there shouldn't be a function call overhead. FSQRT has a latency of 35 cycles + memory access overhead. RSQRTSS has a latency of 3 cycles + memory access overhead. So this approach *should* be the same speed as that, and substantially faster than using C's sqrtf() function.
My guess as to the mistake the author made: the C compiler inlined sqrtf(), but he didn't enable inlining for the code for the other approaches, meaning that there's a function call overhead involved, including a probable pipeline stall for a mispredicted return and a potential pipeline stall for an indirect function call if his compiler was producing position independent call included in the results for the two faster implementations.
One thing I will grant though: the SSE RSQRTSS instrunction (which isn't available on the target architecture of Quake 3) *is* faster than calculating an inverse square root using the method from Quake3. But then you'll need to calculate 1/, which takes an additional 20 cycles. I think the end result will be pretty much neck-and-neck, if tested properly.
And then after something comes up, I'd have a big company telling me to shut down my project (because it may not be possible to revert a big, findamental patch, for example).
I don't think this is avoidable, unfortunately. If you have to remove a fundamental piece of code due to copyright considerations, that's going to effectively mean reverting your codebase to the point it was added and starting again from there. Code added after it was may be a derivitive under copyright law, so you probably can't use a lot of that, either.
Not in the case of the GNU GPL, as far as I understand. I have asked a lawyer about this once (last year I guess).
Copyright laws vary from place to place; most lawyers only consider local issues. You may find that some regions have local laws that allow revocation of a licence even when that licence describes itself as irrevocable (as the GPL does). If you're worried, make sure the lawyer you consult is well versed in international copyright issues. Try to find a copyright specialist who deals worldwide, if you can afford one.
I hereby certify that the work I have submitted to is my own work, which I am entitled to licence under the provisions of , and that I am not aware of any patents or other legal issues that may prevent its use in . I hereby grant a licence to distribute the work under the terms of (attached).
Ahem.
I hereby certify that the work I have submitted to [project] is my own work, which I am entitled to licence under the provisions of [licence], and that I am not aware of any patents or other legal issues that may prevent its use in [project]. I hereby grant a licence to distribute the work under the terms of [licence] (attached).
I think the idea is to get contributors to send you something, signed, on a piece of paper. This is what the FSF does, and a few other large projects have followed their lead. In FSF's case, the piece of paper is also a copyright assignment, which you probably don't want as it puts a lot of potential contributors off. What you want to do is check out something like the following with a lawyer:
I hereby certify that the work I have submitted to is my own work, which I am entitled to licence under the provisions of , and that I am not aware of any patents or other legal issues that may prevent its use in . I hereby grant a licence to distribute the work under the terms of (attached).
You possibly also want to include a similarly phrased paragraph to cover future submissions by the same contributor, if you expect any.
What this does is (again, IANAL, so this isn't legal advice, check it with a professional, actual facts may vary from jurisdiction to jurisdiction):
1. Means you've performed "due diligence" before accepting the work. You've got a signed statement from somebody stating that there wouldn't be any issues. If you do have legal expenses insurance (and I'd recommend it; at least where I live it isn't expensive) your insurers will almost certainly want to see something like this before they'll agree to defend you in a court case. In a court case, I think it would be enough to show that you hadn't knowingly infringed any copyrights, which may be enough to prevent any damages being awarded against you. You'd have to cease distribution, of course, but in the end it would probably not actually cost you anything. It's probably not as good in the case of a patent infringement, where I believe strict liability rules apply, but that's substantially less likely to affect you, fortunately.
2. Means you've got a clear, easy to prove licence to distribute, so your contributor can't turn around and sue you. Yes, this is unlikely, but it's great to cover all angles.
A GPG-signed e-mail may be adequate, but check with a lawyer. In my jurisdiction, I believe it would be iff I could prove the key belonged to the person I believe it to, which can be a quite tricky proposition. In yours, it might not be acceptable at all. Check everything. A signed fax may be better than an e-mail. This is the kind of knowledge you pay a lawyer for.
Just my personal opinion, which is all you can really give here: no, it isn't. If anything, I'd say GNOME's strong point is its slavish obedience of the directives of so-called usability experts. The fact that most people seem to struggle to use it should tell us something: a lot of usability experts don't have a clue what they're talking about.
Unix assumes you're on a local system. Go install Plan 9 or something, or wait until someone comes up with a really awesome FUSE hack.
Both GNOME and KDE are doing this the hacky stop-gap way, and the only difference is that KDE folks have a solution that works, kind of. The elegant way would be to allow this stuff to work on any application. I'm not calling the present situation elegant until I can do "cat http://slashdot.org/ ".
You are aware that there's a FUSE/kioslave bridge available, aren't you?
It's not quite as good as what you're looking for, because the kernel doesn't understand URLs and wouldn't know what to do with one, but it certainly enables some useful features.
What specific usability issues would the average user have in Browsing, Emailing and Wordprocessing ? was Re:I hope the Gnome folks read this bit...
There is no easily-discoverable user interface that allows a user to type in the name of a file they wish to open.
Save-file dialogs use a totally different layout to open-file dialogs, requiring the user to learn two different user interfaces where the job can be trivially simplified to just one simple interface.
And "user mode applications" in this sense means applications running without administrative privilges.
Not according to the CNET article that the wikipedia one links to as a source for that statement -- that states that it won't work for programs "even when run as an administrator".
Unsigned drivers are not allowed in the 64 bit version, but are allowed in the 32 bit. The theory there is that all drivers are new for 64 bit, while 32 bit may still have a great deal of legacy drivers.
What about the drivers that are necessary for the programs I cited? Is MS going to be certifying those? Somehow I doubt it.
You've had it explained TWICE now that the higher taxes were absolute figures (even allowing for inflation), not percentages.
OK; you've not explained anything to me before, as that was my first post in this thread. But carry on. Percentages are more meaningful because they allow us to determine the impact on individual people, which is the only real relevance of this, surely? But if you want "absolute figures allowing for inflation"...
And do you really believe that tax revenue in 1976 was HIGHER than it is today?
In inflation-adjusted terms, yes. In 1976, total managed expenditure in the UK was £300bn (approx) whereas in 2005 it was £500bn (in 2005-adjusted pounds). In terms of percentages of national income (which, for a variety of reasons, is the only meaningful measure) it was 50% in 1976 but only 42% in 2005. Its lowest point over the last 40 years was in 2001, where it dropped to 38%.
FACT: The UK government is collecting MORE in taxes today, even when taking into account inflation, than it was ten years ago.
Wrong. In 1996, tax income was £286.4bn; in 2005 (latest year available) it was &451.4bn. £286.4bn in 1996 is worth roughly £504bn in 2006.
FACT: The UK government has never collected less taxes in any year than it did in the prior year, since the 1980s.
Incorrect. The new year tax revenue was less than the previous year's adjusted for inflation in 1990, 1992, 2001 and 2002.
FACT: The entire rationale behind lowering income tax percentages was to encourage people to earn more and therefore pay more in taxes over-all.
Yes; lower tax rates are generally held to create a better economic climate, which does generally, in the end, result in higher tax revenues. Your point?
FACT: There has only ever been one decrease in VAT (fuel tax).
Actually, there were two. Fuel was reclassified from 17.5% to 8% in 1994 and then reduced further to 5% in 1997. There's only been one increase in the last 25 years (all the data I have available) from 15% to 17.5% in 1991. I believe this is the only increase there has ever been. The reduced rate has also been extended to a number of other products beside fuel, now.
They're constitutional monarchys, which is substantially different from an absolute monarchy, which is apparently what you were talking about (i.e., you implied that the existence of a monarch has a substantial effect on the power structure of the nation, which is generally not the case in a constitutional monarchy). Also note that Britain is also a constituational monarchy, yet you classified it as a democracy in your original post.
Tax revenue is a perfectly good measure of bureaucracy. A huge government made up of many millions of bureaucrats requires far more than a small government. If you disagree, I suggest you come up with another, more reliable, objective metric. Oh, but you can't, so you just chose to pretend you don't understand the point.
Large tax may also indicate the presence of substantial social security programs, military spending or any of a number of other factors. A much more useful indicator is number of civil servants per capita, but even then you'd expect to see the figures skewed higher in smaller countries.
Almost every democratic country in Europe, if not all of them, have much higher taxes than, say, ten years ago, even disregarding inflation
Interesting that you think that. The UK, for instance, has decreased income tax from 20%,25% and 40% to 10%,22% and 40%. If you look back farther, you see tax rates were even higher. Thirty years ago, the lowest rate didn't even exist -- income tax varied between 35% and 83%. I don't believe the UK is an exception.
My latest mac is about 1 or 2 years old. I got it from the apple store pre-loaded with OSX 10.4. I have yet to re-install it. Has run perfect just as expected this whole time.
To be fair to MS, my current PC is one I've been using for about 18 months; I installed XP+SP2 on it when I got it and haven't had to reinstall since. If I hadn't bought a second hand PC, XP SP2 would've been preloaded.
And it's not that I don't play around with strange software. I have a load of shit installed here, much of which has its own unsigned kernel-mode drivers associated (e.g. colinux, openvpn, truecrypt,...). I have effectively zero trouble with the machine. Other than its tiny hard disk, the geriatric processor and the ATI Rage3D video card, but I don't think the OS can fix those problems for me.
In the mid 90s, Microsoft rapidly built upon its past success with MS-DOS to establish Windows as a vast empire...just as Apple used the success of the Apple II as a stepping stone to launch the Mac in the mid 80s.
You're timings somewhat out there. By 1994, ten years after the Mac was launched, Windows 3.1 was almost ubiquitous, and NT 3.5 was already doing rather well.
From 2002-2006, Microsoft dropped features, changed plans, and started over several times in protracted efforts to ship Longhorn...just as Apple had fumbled around with Copland ten years earlier.
You're ignoring the fact that MS dropped just as many features and plans from the long-anticipated release of NT 5.0 (renamed Windows 2000, perhaps to distance themselves from the lists of features that had been widely reported but which didn't actually make it), which was expected to have features like Remote Desktop and Fast User Switching that were eventually deferred to XP. And that the Object File System that was originally planned for NT 4.0 has yet to be implemented. And...
Essentially, MS has a long history of promising features that never make it. Using a subset of that to support some kind of analogy is selective blindness.
1) Integration of Internet Explorer. Microsoft claims that IE and Windows are inextricably linked together, and at least for Windows 2000 and newer this seems to be true. For instance, if you type a URL into the address bar of the Windows Explorer, it will show you web pages. IMHO a stupid design, the web browser should be an application, not a fixed part of the GUI.
You do know that one of the major new features of Vista is that it has separated IE to be just an application, not integrated into the shell, right?
2) The RPC service being responsible for things a "remote procedure call service" has no business handling.
An RPC system is designed to enable procedure calls to be made between two different processes. Whether they are on the same machine or not is irrelevant, it doesn't matter, the issues to be dealt with are pretty much the same, so why not use the same system? It reduces complexity by eliminating unnecessary code.
If you don't want remote RPC to be possible, firewall your RPC server.
A worldwide known symbol for turning off computers gets used to sleep machines.
My experience is that the symbol (a red digit one circumscribed by a zero) is more frequently used for 'standby' mode (e.g. on televisions, DVD players and other similar home equipment) than for a complete power-off (as it often is on PCs). Besides, you're much more likely to want standby than power off in most circumstances.
Ctrl+Esc has been an instinctive reaction for me to wanting to start a new program since I switched to NT4 in 97. I didn't get a keyboard with a windows key on it until some time around 2001. Even now, I occasionally use machines without them, so learning to use it instinctively is something I hesitate over. It'll be hard to unlearn this behaviour.
Slashdot Mirror
The difference is that with a limited set of operator symbols to choose from, it's very tempting to use one that bares little resemblance to what your operator actually does. There are substantially more plausible function names, so the temptation to use one that is inappropriate is smaller.
What's wrong with a C-style cast? Simply consider it a shorthand notation for a const_cast or reinterpret_cast, depending on situation.
Frankly, there's no better way of truncating the result of a floating point operation than x = (int) (f / 2.3);. and "reinterpret_cast" is just too verbose when what you really want is bits = *(int *).
I think your problem is that you're using XML to perform the job of a relational database.
Not all tasks can be solved with the same tools.
The problem with asking something like this on a site like slashdot is twofold:
1. Armchair lawyers. Sure, I'm one of 'em. I like to think I'm better than most, but it's very hard to tell who's reliable and who isn't.
2. This kind of thing varies from place to place. You need somebody local.
I've never played with proper model rocket engines like that, but I did build a little solid-fuel rocket once. Real home-made stuff: a 100ml aluminium drink can with ground matchheads as the fuel. It didn't go high, maybe only 10 metres or so, but it was carrying a rather heavy payload... it stuck to the stake that I launched it from and took that up with it, probably about a kilo or so.
That was well & truly cool.
(Note to others who might try this: use a thick transparent plastic shield to protect you from shrapnel in case the can explodes...)
I've built myself a similar project using a cheap FPGA, but that's substantially cooler. :)
I did something similar, although I connected a few machines up in a ring using the RS423 port (Tx on machine A -> Rx on B, Tx on B to Rx on C, and Tx on C to Rx on A). That was cool. :)
Err... OK. I'd like to know the author of that page's testing methodology, 'cause I really can't see how the 'faster_sqrtf' function presented could possibly take 426 cycles to execute: it's a memory reference followed by 3 standard integer operations. On the 1.2GHz Opteron described, it should take 3 cycles (during each of which there are 2 spare execution units for out-of-order execution of other code) plus a memory access if there's a cache miss, which with 133MHz memory should take 9 cycles, for a grand total of 12 cycles. I'm assuming the code is already in the cache, because in any situation where you care that much it will be. It'll also be inlined into the code that's calling it, so there shouldn't be a function call overhead. FSQRT has a latency of 35 cycles + memory access overhead. RSQRTSS has a latency of 3 cycles + memory access overhead. So this approach *should* be the same speed as that, and substantially faster than using C's sqrtf() function.
My guess as to the mistake the author made: the C compiler inlined sqrtf(), but he didn't enable inlining for the code for the other approaches, meaning that there's a function call overhead involved, including a probable pipeline stall for a mispredicted return and a potential pipeline stall for an indirect function call if his compiler was producing position independent call included in the results for the two faster implementations.
One thing I will grant though: the SSE RSQRTSS instrunction (which isn't available on the target architecture of Quake 3) *is* faster than calculating an inverse square root using the method from Quake3. But then you'll need to calculate 1/, which takes an additional 20 cycles. I think the end result will be pretty much neck-and-neck, if tested properly.
And then after something comes up, I'd have a big company telling me to shut down my project (because it may not be possible to revert a big, findamental patch, for example).
I don't think this is avoidable, unfortunately. If you have to remove a fundamental piece of code due to copyright considerations, that's going to effectively mean reverting your codebase to the point it was added and starting again from there. Code added after it was may be a derivitive under copyright law, so you probably can't use a lot of that, either.
Not in the case of the GNU GPL, as far as I understand. I have asked a lawyer about this once (last year I guess).
Copyright laws vary from place to place; most lawyers only consider local issues. You may find that some regions have local laws that allow revocation of a licence even when that licence describes itself as irrevocable (as the GPL does). If you're worried, make sure the lawyer you consult is well versed in international copyright issues. Try to find a copyright specialist who deals worldwide, if you can afford one.
I hereby certify that the work I have submitted to is my own work, which I am entitled to licence under the provisions of , and that I am not aware of any patents or other legal issues that may prevent its use in . I hereby grant a licence to distribute the work under the terms of (attached).
Ahem.
I hereby certify that the work I have submitted to [project] is my own work, which I am entitled to licence under the provisions of [licence], and that I am not aware of any patents or other legal issues that may prevent its use in [project]. I hereby grant a licence to distribute the work under the terms of [licence] (attached).
I think the idea is to get contributors to send you something, signed, on a piece of paper. This is what the FSF does, and a few other large projects have followed their lead. In FSF's case, the piece of paper is also a copyright assignment, which you probably don't want as it puts a lot of potential contributors off. What you want to do is check out something like the following with a lawyer:
I hereby certify that the work I have submitted to is my own work, which I am entitled to licence under the provisions of , and that I am not aware of any patents or other legal issues that may prevent its use in . I hereby grant a licence to distribute the work under the terms of (attached).
You possibly also want to include a similarly phrased paragraph to cover future submissions by the same contributor, if you expect any.
What this does is (again, IANAL, so this isn't legal advice, check it with a professional, actual facts may vary from jurisdiction to jurisdiction):
1. Means you've performed "due diligence" before accepting the work. You've got a signed statement from somebody stating that there wouldn't be any issues. If you do have legal expenses insurance (and I'd recommend it; at least where I live it isn't expensive) your insurers will almost certainly want to see something like this before they'll agree to defend you in a court case. In a court case, I think it would be enough to show that you hadn't knowingly infringed any copyrights, which may be enough to prevent any damages being awarded against you. You'd have to cease distribution, of course, but in the end it would probably not actually cost you anything. It's probably not as good in the case of a patent infringement, where I believe strict liability rules apply, but that's substantially less likely to affect you, fortunately.
2. Means you've got a clear, easy to prove licence to distribute, so your contributor can't turn around and sue you. Yes, this is unlikely, but it's great to cover all angles.
A GPG-signed e-mail may be adequate, but check with a lawyer. In my jurisdiction, I believe it would be iff I could prove the key belonged to the person I believe it to, which can be a quite tricky proposition. In yours, it might not be acceptable at all. Check everything. A signed fax may be better than an e-mail. This is the kind of knowledge you pay a lawyer for.
What's the logic of going with a version that is so far behind? I know that you don't go bleeding edge with such a project but 9.3 is ancient.
Huh? My desktop machine is on 9.1. I installed it from the latest available version substantially less than 3 years ago.
And yes, I am rather annoyed that they've stopped issuing updates for it.
I thought usability was Gnome's strong point?
Just my personal opinion, which is all you can really give here: no, it isn't. If anything, I'd say GNOME's strong point is its slavish obedience of the directives of so-called usability experts. The fact that most people seem to struggle to use it should tell us something: a lot of usability experts don't have a clue what they're talking about.
Unix assumes you're on a local system. Go install Plan 9 or something, or wait until someone comes up with a really awesome FUSE hack.
Both GNOME and KDE are doing this the hacky stop-gap way, and the only difference is that KDE folks have a solution that works, kind of. The elegant way would be to allow this stuff to work on any application. I'm not calling the present situation elegant until I can do "cat http://slashdot.org/ ".
You are aware that there's a FUSE/kioslave bridge available, aren't you?
It's not quite as good as what you're looking for, because the kernel doesn't understand URLs and wouldn't know what to do with one, but it certainly enables some useful features.
What specific usability issues would the average user have in Browsing, Emailing and Wordprocessing ? was Re:I hope the Gnome folks read this bit ...
There is no easily-discoverable user interface that allows a user to type in the name of a file they wish to open.
Save-file dialogs use a totally different layout to open-file dialogs, requiring the user to learn two different user interfaces where the job can be trivially simplified to just one simple interface.
Then Gnome shouldn't be the default desktop for most Linux distributions, as most of them do have that as a goal.
Step #1 of picking a name for any OSS is looking for it on Google.
Step #1 of naming *anything* is looking it up on Google.
And "user mode applications" in this sense means applications running without administrative privilges.
Not according to the CNET article that the wikipedia one links to as a source for that statement -- that states that it won't work for programs "even when run as an administrator".
Unsigned drivers are not allowed in the 64 bit version, but are allowed in the 32 bit. The theory there is that all drivers are new for 64 bit, while 32 bit may still have a great deal of legacy drivers.
What about the drivers that are necessary for the programs I cited? Is MS going to be certifying those? Somehow I doubt it.
You've had it explained TWICE now that the higher taxes were absolute figures (even allowing for inflation), not percentages.
OK; you've not explained anything to me before, as that was my first post in this thread. But carry on. Percentages are more meaningful because they allow us to determine the impact on individual people, which is the only real relevance of this, surely? But if you want "absolute figures allowing for inflation"...
And do you really believe that tax revenue in 1976 was HIGHER than it is today?
In inflation-adjusted terms, yes. In 1976, total managed expenditure in the UK was £300bn (approx) whereas in 2005 it was £500bn (in 2005-adjusted pounds). In terms of percentages of national income (which, for a variety of reasons, is the only meaningful measure) it was 50% in 1976 but only 42% in 2005. Its lowest point over the last 40 years was in 2001, where it dropped to 38%.
FACT: The UK government is collecting MORE in taxes today, even when taking into account inflation, than it was ten years ago.
Wrong. In 1996, tax income was £286.4bn; in 2005 (latest year available) it was &451.4bn. £286.4bn in 1996 is worth roughly £504bn in 2006.
FACT: The UK government has never collected less taxes in any year than it did in the prior year, since the 1980s.
Incorrect. The new year tax revenue was less than the previous year's adjusted for inflation in 1990, 1992, 2001 and 2002.
FACT: The entire rationale behind lowering income tax percentages was to encourage people to earn more and therefore pay more in taxes over-all.
Yes; lower tax rates are generally held to create a better economic climate, which does generally, in the end, result in higher tax revenues. Your point?
FACT: There has only ever been one decrease in VAT (fuel tax).
Actually, there were two. Fuel was reclassified from 17.5% to 8% in 1994 and then reduced further to 5% in 1997. There's only been one increase in the last 25 years (all the data I have available) from 15% to 17.5% in 1991. I believe this is the only increase there has ever been. The reduced rate has also been extended to a number of other products beside fuel, now.
So you admit they're monarchies
They're constitutional monarchys, which is substantially different from an absolute monarchy, which is apparently what you were talking about (i.e., you implied that the existence of a monarch has a substantial effect on the power structure of the nation, which is generally not the case in a constitutional monarchy). Also note that Britain is also a constituational monarchy, yet you classified it as a democracy in your original post.
Tax revenue is a perfectly good measure of bureaucracy. A huge government made up of many millions of bureaucrats requires far more than a small government. If you disagree, I suggest you come up with another, more reliable, objective metric. Oh, but you can't, so you just chose to pretend you don't understand the point.
Large tax may also indicate the presence of substantial social security programs, military spending or any of a number of other factors. A much more useful indicator is number of civil servants per capita, but even then you'd expect to see the figures skewed higher in smaller countries.
Almost every democratic country in Europe, if not all of them, have much higher taxes than, say, ten years ago, even disregarding inflation
Interesting that you think that. The UK, for instance, has decreased income tax from 20%,25% and 40% to 10%,22% and 40%. If you look back farther, you see tax rates were even higher. Thirty years ago, the lowest rate didn't even exist -- income tax varied between 35% and 83%. I don't believe the UK is an exception.
My latest mac is about 1 or 2 years old. I got it from the apple store pre-loaded with OSX 10.4. I have yet to re-install it. Has run perfect just as expected this whole time.
...). I have effectively zero trouble with the machine. Other than its tiny hard disk, the geriatric processor and the ATI Rage3D video card, but I don't think the OS can fix those problems for me.
To be fair to MS, my current PC is one I've been using for about 18 months; I installed XP+SP2 on it when I got it and haven't had to reinstall since. If I hadn't bought a second hand PC, XP SP2 would've been preloaded.
And it's not that I don't play around with strange software. I have a load of shit installed here, much of which has its own unsigned kernel-mode drivers associated (e.g. colinux, openvpn, truecrypt,
In the mid 90s, Microsoft rapidly built upon its past success with MS-DOS to establish Windows as a vast empire ...just as Apple used the success of the Apple II as a stepping stone to launch the Mac in the mid 80s.
...just as Apple had fumbled around with Copland ten years earlier.
You're timings somewhat out there. By 1994, ten years after the Mac was launched, Windows 3.1 was almost ubiquitous, and NT 3.5 was already doing rather well.
From 2002-2006, Microsoft dropped features, changed plans, and started over several times in protracted efforts to ship Longhorn
You're ignoring the fact that MS dropped just as many features and plans from the long-anticipated release of NT 5.0 (renamed Windows 2000, perhaps to distance themselves from the lists of features that had been widely reported but which didn't actually make it), which was expected to have features like Remote Desktop and Fast User Switching that were eventually deferred to XP. And that the Object File System that was originally planned for NT 4.0 has yet to be implemented. And...
Essentially, MS has a long history of promising features that never make it. Using a subset of that to support some kind of analogy is selective blindness.
1) Integration of Internet Explorer.
Microsoft claims that IE and Windows are inextricably linked together, and at least for Windows 2000 and newer this seems to be true. For instance, if you type a URL into the address bar of the Windows Explorer, it will show you web pages. IMHO a stupid design, the web browser should be an application, not a fixed part of the GUI.
You do know that one of the major new features of Vista is that it has separated IE to be just an application, not integrated into the shell, right?
2) The RPC service being responsible for things a "remote procedure call service" has no business handling.
An RPC system is designed to enable procedure calls to be made between two different processes. Whether they are on the same machine or not is irrelevant, it doesn't matter, the issues to be dealt with are pretty much the same, so why not use the same system? It reduces complexity by eliminating unnecessary code.
If you don't want remote RPC to be possible, firewall your RPC server.
A worldwide known symbol for turning off computers gets used to sleep machines.
My experience is that the symbol (a red digit one circumscribed by a zero) is more frequently used for 'standby' mode (e.g. on televisions, DVD players and other similar home equipment) than for a complete power-off (as it often is on PCs). Besides, you're much more likely to want standby than power off in most circumstances.
Ctrl+Esc has been an instinctive reaction for me to wanting to start a new program since I switched to NT4 in 97. I didn't get a keyboard with a windows key on it until some time around 2001. Even now, I occasionally use machines without them, so learning to use it instinctively is something I hesitate over. It'll be hard to unlearn this behaviour.