Non-executable pages have been around for a *long* time, and they've been around in the Intel architecture since the 386 at least (I'm too lazy to check and see if they existed for the 286)... Microsoft simply didn't *use* the feature.
That's simply not true.
Non-executable *segments* have been available since the 286. Non-executable *pages* have only been introduced by the latest revisions of AMD's processors, and by a revision of Intel's that I believe is not available yet (?).
This means that you can have NX protection, but only by not using the flat memory model, or by using a hack involving duplicating your address space (meaning you only get half as much of it). I think MS decided that neither of these were acceptable, and I think they were right.
Backwards compatibility with the NX feature would have been impossible. The problem is that it was simply impossible to implement^{*} before now. The APIs that applications are now required to use (i.e. allocating memory using the Virtual* functions and passing the PAGE_EXECUCTE protection flag) have been documented since the first version of NT was released (when was that? ten years ago?). It's just that people could get away without using them.
*: Yes I'm aware that there are kludges that can be used to implement an approximation of it. They either (1) require you to not use a flat address space, an important feature of NT, or (2) substantially reduce the amount of memory space available to processes. Neither of these is an acceptable cost, IMO.
While I don't entirely agree with the GP article's point, we're not talking about Win3.11 apps here. We're talking about Windows XP apps not working on XP service pack 2.
Now, I hardly see "millions" of users complaining. At a guess, less than 0.5% of software will have an issue, and odd things at that.
The biggest thing is NX. But that'll only affect people with 64 bit processors (a tiny majority at the moment). The DCOM and RPC changes only affect people using DCOM or RPC for applications that are not core windows apps (a tiny proportion, almost all custom developed and therefore supported by the developer), and can probably be fixed with an easy-to-install registry update, which a competent developer ought to be able to produce in just a few hours. Firewall related problems can be easily fixed by the user.
Why does everything have to be compared to MP3s? Why couldn't it have been 'Huffman coding used in ZIP files' or '[...] used by GZip' or '[...] used by the huffyuv lossless video codec' or any of about 5 million other applications that use huffman coding. Most of which are a lot more specific than MP3 which also uses a cocktail of other techniques to achieve compression and is, above all else, lossy, which huffman coding isn't.
To be transmitted across the Internet, this message was broken down into bits, like MP3s are.
Read the bill. It doesn't "nuke" the DMCA. It just amends it slightly to permit you to break copy protection for purposes which are otherwise legal. And requires copy protected CDs to be labelled accurately and prominently. All the rest of the DMCA provisions aren't touched by it.
Did you read the comment you were responding to? The poster was pointing out that the situation with web sites is no different in reality to the situation with newspapers or magazines. The article (yes, I did read it) was implying that it is more difficult to track the number of readers of a web site than it is a magazine. This is, of course, bollocks.
There are paradoxes and plotholes large enough to dump an entire series in, to never be seen again (Enterprise would be my favorite here...) and if I see one more pointless time travel...
I think Enterprise was actually designed to be cut off from the rest of the canon. The entire 'temporal cold war' thing from series 1 was a pretty strong hint at this: what you were seeing was the result of somebody using time travel to try to prevent the formation of the federation. Then they went and had that borg episode, and everyone realised that this wasn't actually supposed to have happened before the rest of trek, but after it (in the perception of somebody who had travelled back in time).
Any future series will be free to accept or ignore Enterprise as its authors see fit.
Whenever a scriptwriter suggest timetravel as a plot device, fire the bastard
Time travel can produce interesting plots. It just has to be co-ordinated at the entire series level, in order to (a) prevent inconsistencies from appearing (there are _way_ too many inconsistencies in trek already) and (b) allow for the most cool effect it can have: the fact that there have been little things throughout the entire series telling you it's going to happen.
JMS understands this. He did use time travel to good effect in B5. Doing it in trek would be harder, because of the larger amount of backstory he'd have to agree with, but he might still be able to do it. And if he does, I'm sure it'll work.
These guys are out exploring space and the best they can do for excitement is do fantasy games indoors? Losers.
Hell, no. They play poker too.
Look, the reality of it is, exploring space is _dull_. Even if you're on a ship that appears in its own Trek show, you only get 24 interesting things to do per year (probably as few as 20 by the time you've considered double-length episodes). You've got to fill up the two weeks between each such incedent doing something.
Yeah, I read Joel's article too. However, I don't think XHTML, CSS2, or transparent PNGs are a threat in this regard.
If he's right, what would be a threat is fixing all the bugs in the jscript implementation, providing mechanisms that allow better user interfaces to be developed (e.g. adding anything similar to XUL), or anything along those lines. Microsoft aren't against browser-based applications. But they want you to use ActiveX (or any other technology nobody else has) to achieve it.
I.e. "threatening to fire somebody" is illegal in the EU. You may just do it or leave it, but it is explicit illegal to put it under any condition..
Do you have a source for that? I was under the impression that it was actually hard to fire somebody _without_ threatening them first (e.g., issuing a warning phrased like "if you do this again we'll fire you" or similar) under EU regulations.
I don't know about the US, but in the UK, and I believe most other countries, you _must_ threaten to sue before doing so. It's usually called "issuing a notice of intent to commence procedings" or some other such name, but really it just says "pay us some reasonable compensation for this now, or we'll sue you."
Note, however, that as Linux does not have the legal status of an individual (which Microsoft does, being an incorporated company), Linux does not benefit from such legal protection as Microsoft does in almost all jurisdictions.
You can say whatever you like about Linux, and there's not a lot anyone can do about it.
Thunderbird would start to hang after a few (e.g. 10-ish) minutes of inactivity, and the only solution would be to quit and re-start
One thing that bugs me about both TBird and mozilla's mail client is that there is no way of forcing a disconnect/reconnect to an IMAP server, so if your server hangs, the mail client stops being able to get mail. Very annoying.
I think XP beats 2K; I can't think of much that 2K does better. Yes, it doesn't require product activation, and doesn't have an annoying UI to switch off, but those are pretty minor, all things considered. And XP home's fast user switching is worth it, IMO. Win2K's runas facility was a step in the right direction, but there's enough that doesn't work properly with it (e.g. you can't open an explorer window as administrator while logged in as an ordinary user, which would be really useful) that FUS fixes to make XP a clear winner, IMO.
And XP seems to have faster startup times, too, although that could be the hardware I have them installed on.
Does RH9 do prelinking? If it doesn't, it might be worth your while setting it up. Windows has a huge advantage over Linux without prelinking in loading applications that use a lot of dynamic linking, because Windows has always had prelinking (of a kind, not as good as the Linux implementation, but it works).
Huh? For me, Word 97: 6 seconds; OO.o 1.1.0 34 seconds to splash page, over a _minute_ until I could edit the document. Both of them loading off the same network filesystem (100MBit network).
I love it, but they really need to work on that startup speed.
And no, I didn't have any MSOffice components already running. And, to make things even more interesting, I never installed on this machine in the first place, so it can't be preloading anything, and that 6 seconds _includes_ the time it took me to dismiss the message box warning me that I don't have VBA installed.
In professional publishing, that's the way it still works. Authors and editors use word processors -- only at the very last stage does a "typesetter" do the TeX markup or whatever.
However, is there some reason they couldn't fire this at foot/knee level and still get effective results?
Yes. In order to be effective, the electrical discharge needs to cross a major nerve centre, e.g. the spine. Typical strategy for using a tazer to disable someone is to try to hit their back with it. Chest will do, but is more dangerous. If you hit somebody in the leg with it, it will make that leg go numb, but they'll be able to limp around on the other one still, I reckon.
In crowd control, self defense, and effecting arrests, what you're trying to do is win WITHOUT fighting. What this means in practice is bringing a big enough threat to bear that even the dumbest cop (or mob) realizes you'll win if you use it.
Yes, I knew sooner or later somebody would come up with a way to justify tactical nuclear weapons as a crowd control measures... there's a missing factor here. The crowd has to believe that you're likely to use the weapon.
If I was in a situation where I was in a mob the police were trying to control, and the _only_ weapon they had available was a gun, and I figured the rest of the crowd would follow my lead, I wouldn't be intimidated by it, because I know that they're trained only to use them when they believe there is no other option. Just make sure you're not threatening anybody's life, and you'll be fine.
Of course, in reality, the police also have big sticks.
Slashdot Mirror
Non-executable pages have been around for a *long* time, and they've been around in the Intel architecture since the 386 at least (I'm too lazy to check and see if they existed for the 286)... Microsoft simply didn't *use* the feature.
That's simply not true.
Non-executable *segments* have been available since the 286. Non-executable *pages* have only been introduced by the latest revisions of AMD's processors, and by a revision of Intel's that I believe is not available yet (?).
This means that you can have NX protection, but only by not using the flat memory model, or by using a hack involving duplicating your address space (meaning you only get half as much of it). I think MS decided that neither of these were acceptable, and I think they were right.
Backwards compatibility with the NX feature would have been impossible. The problem is that it was simply impossible to implement^{*} before now. The APIs that applications are now required to use (i.e. allocating memory using the Virtual* functions and passing the PAGE_EXECUCTE protection flag) have been documented since the first version of NT was released (when was that? ten years ago?). It's just that people could get away without using them.
*: Yes I'm aware that there are kludges that can be used to implement an approximation of it. They either (1) require you to not use a flat address space, an important feature of NT, or (2) substantially reduce the amount of memory space available to processes. Neither of these is an acceptable cost, IMO.
While I don't entirely agree with the GP article's point, we're not talking about Win3.11 apps here. We're talking about Windows XP apps not working on XP service pack 2.
Now, I hardly see "millions" of users complaining. At a guess, less than 0.5% of software will have an issue, and odd things at that.
The biggest thing is NX. But that'll only affect people with 64 bit processors (a tiny majority at the moment). The DCOM and RPC changes only affect people using DCOM or RPC for applications that are not core windows apps (a tiny proportion, almost all custom developed and therefore supported by the developer), and can probably be fixed with an easy-to-install registry update, which a competent developer ought to be able to produce in just a few hours. Firewall related problems can be easily fixed by the user.
[...] Huffman coding used in MP3s [...]
Why does everything have to be compared to MP3s? Why couldn't it have been 'Huffman coding used in ZIP files' or '[...] used by GZip' or '[...] used by the huffyuv lossless video codec' or any of about 5 million other applications that use huffman coding. Most of which are a lot more specific than MP3 which also uses a cocktail of other techniques to achieve compression and is, above all else, lossy, which huffman coding isn't.
To be transmitted across the Internet, this message was broken down into bits, like MP3s are.
Read the bill. It doesn't "nuke" the DMCA. It just amends it slightly to permit you to break copy protection for purposes which are otherwise legal. And requires copy protected CDs to be labelled accurately and prominently. All the rest of the DMCA provisions aren't touched by it.
Did you read the comment you were responding to? The poster was pointing out that the situation with web sites is no different in reality to the situation with newspapers or magazines. The article (yes, I did read it) was implying that it is more difficult to track the number of readers of a web site than it is a magazine. This is, of course, bollocks.
There are paradoxes and plotholes large enough to dump an entire series in, to never be seen again (Enterprise would be my favorite here...) and if I see one more pointless time travel...
I think Enterprise was actually designed to be cut off from the rest of the canon. The entire 'temporal cold war' thing from series 1 was a pretty strong hint at this: what you were seeing was the result of somebody using time travel to try to prevent the formation of the federation. Then they went and had that borg episode, and everyone realised that this wasn't actually supposed to have happened before the rest of trek, but after it (in the perception of somebody who had travelled back in time).
Any future series will be free to accept or ignore Enterprise as its authors see fit.
Whenever a scriptwriter suggest timetravel as a plot device, fire the bastard
Time travel can produce interesting plots. It just has to be co-ordinated at the entire series level, in order to (a) prevent inconsistencies from appearing (there are _way_ too many inconsistencies in trek already) and (b) allow for the most cool effect it can have: the fact that there have been little things throughout the entire series telling you it's going to happen.
JMS understands this. He did use time travel to good effect in B5. Doing it in trek would be harder, because of the larger amount of backstory he'd have to agree with, but he might still be able to do it. And if he does, I'm sure it'll work.
These guys are out exploring space and the best they can do for excitement is do fantasy games indoors? Losers.
Hell, no. They play poker too.
Look, the reality of it is, exploring space is _dull_. Even if you're on a ship that appears in its own Trek show, you only get 24 interesting things to do per year (probably as few as 20 by the time you've considered double-length episodes). You've got to fill up the two weeks between each such incedent doing something.
From experience, I can tell you that you don't have to kill your firewall to run windows update.
I think the poster interpreted the instruction to "quit all running applications" a little too seriously.
Yeah, I read Joel's article too. However, I don't think XHTML, CSS2, or transparent PNGs are a threat in this regard.
If he's right, what would be a threat is fixing all the bugs in the jscript implementation, providing mechanisms that allow better user interfaces to be developed (e.g. adding anything similar to XUL), or anything along those lines. Microsoft aren't against browser-based applications. But they want you to use ActiveX (or any other technology nobody else has) to achieve it.
In soviet russia... I'm sure you can figure the rest out for yourself.
I think IE 4 was a much better browser than Netscape 4. IE 3 was the last version that Netscape could win against, I think. IE 3 was horrible.
Actually...
I.e. "threatening to fire somebody" is illegal in the EU. You may just do it or leave it, but it is explicit illegal to put it under any condition..
Do you have a source for that? I was under the impression that it was actually hard to fire somebody _without_ threatening them first (e.g., issuing a warning phrased like "if you do this again we'll fire you" or similar) under EU regulations.
I don't know about the US, but in the UK, and I believe most other countries, you _must_ threaten to sue before doing so. It's usually called "issuing a notice of intent to commence procedings" or some other such name, but really it just says "pay us some reasonable compensation for this now, or we'll sue you."
Note, however, that as Linux does not have the legal status of an individual (which Microsoft does, being an incorporated company), Linux does not benefit from such legal protection as Microsoft does in almost all jurisdictions.
You can say whatever you like about Linux, and there's not a lot anyone can do about it.
Thunderbird would start to hang after a few (e.g. 10-ish) minutes of inactivity, and the only solution would be to quit and re-start
One thing that bugs me about both TBird and mozilla's mail client is that there is no way of forcing a disconnect/reconnect to an IMAP server, so if your server hangs, the mail client stops being able to get mail. Very annoying.
I think XP beats 2K; I can't think of much that 2K does better. Yes, it doesn't require product activation, and doesn't have an annoying UI to switch off, but those are pretty minor, all things considered. And XP home's fast user switching is worth it, IMO. Win2K's runas facility was a step in the right direction, but there's enough that doesn't work properly with it (e.g. you can't open an explorer window as administrator while logged in as an ordinary user, which would be really useful) that FUS fixes to make XP a clear winner, IMO.
And XP seems to have faster startup times, too, although that could be the hardware I have them installed on.
Does RH9 do prelinking? If it doesn't, it might be worth your while setting it up. Windows has a huge advantage over Linux without prelinking in loading applications that use a lot of dynamic linking, because Windows has always had prelinking (of a kind, not as good as the Linux implementation, but it works).
Huh? For me, Word 97: 6 seconds; OO.o 1.1.0 34 seconds to splash page, over a _minute_ until I could edit the document. Both of them loading off the same network filesystem (100MBit network).
I love it, but they really need to work on that startup speed.
And no, I didn't have any MSOffice components already running. And, to make things even more interesting, I never installed on this machine in the first place, so it can't be preloading anything, and that 6 seconds _includes_ the time it took me to dismiss the message box warning me that I don't have VBA installed.
In professional publishing, that's the way it still works. Authors and editors use word processors -- only at the very last stage does a "typesetter" do the TeX markup or whatever.
And they use underline for highlighting.
The only reason I stopped using WP5.1 was that under the NT-based Windowses the virtual DOS machine sucks up 100% of the CPU while not doing anything.
This isn't a problem I've had with NT. Maybe its a bug in wordperfect?
Incorrect firing stance for a large weapon (knees locked, stance too narrow)
As this weapon is (presumably) recoilless, I don't suspect the stance is too important. But I take the rest of your points.
However, is there some reason they couldn't fire this at foot/knee level and still get effective results?
Yes. In order to be effective, the electrical discharge needs to cross a major nerve centre, e.g. the spine. Typical strategy for using a tazer to disable someone is to try to hit their back with it. Chest will do, but is more dangerous. If you hit somebody in the leg with it, it will make that leg go numb, but they'll be able to limp around on the other one still, I reckon.
In crowd control, self defense, and effecting arrests, what you're trying to do is win WITHOUT fighting. What this means in practice is bringing a big enough threat to bear that even the dumbest cop (or mob) realizes you'll win if you use it.
Yes, I knew sooner or later somebody would come up with a way to justify tactical nuclear weapons as a crowd control measures... there's a missing factor here. The crowd has to believe that you're likely to use the weapon.
If I was in a situation where I was in a mob the police were trying to control, and the _only_ weapon they had available was a gun, and I figured the rest of the crowd would follow my lead, I wouldn't be intimidated by it, because I know that they're trained only to use them when they believe there is no other option. Just make sure you're not threatening anybody's life, and you'll be fine.
Of course, in reality, the police also have big sticks.