Show how XPSP2 IE6 installs spyware without user action, without the user being given the option and a security warning. Imagine Gator informing the user, in detail, how to allow a Firefox extension being installed to gain great functionality. (Or, of course, just getting the user to install local software, which then plugs in the extension.) And add the hopefully enhanced least user privilege support in Vista. (hopefully keeping more people from using a real full Administrator as the normal use account)
Sorry. The point is that both parties are using highly directional antennas. And, well, a single access point wouldn't be able to serve much bandwidth with that many users anyway. (As the contention management protocols are not really designed for 1000+ users, you will get even less than your theoretical slice.)
I think your gene example is a very bad example of an obvious patent. It's NOT obvious what the function is. (If it was, or it was already published elsewhere, the patent should of course be void.)
If the patent should cover tests, or just artificial use of the gene in other organisms may be worth some discussion, but there are many inventions that are mostly based on the fact that someone has tested a lot of stuff to find what works and what doesn't. The fact that a lot of people might have hit an optimum alloy composition by accident is not prior art for a patent describing a specific alloy, including its wonderful properties.
It makes no sense to use the same reusable design for manned vs. unmanned, at all. (Just as any cargo-with-safe-reentry should be quite different from just a cargo payload up in the sky.)
I think this is an important point to make, and this is what makes this stuff so hard. Is simply providing a router with filtering and logging capabilities a part of "aiding the Chinese government"? Should they actually spend money on the (somewhat) futile attempt to sell one set of firmware on the Chinese routers, and another set on the domestic ones, where there may be legitimate uses (filtering and control of the use in a corporate network, for example).
If a search engine actively filters content, it's one thing. If the chinese proxies mangle the results to exclude content, they are (obviously) not responsible. Likewise, if it's just a matter of using the general flexible Cisco features for a repressionist purpose, I wouldn't blame Cisco. Drawing the line here can get quite tricky, of course, as a new general feature could be introduced after requests from "an important customer"...
Good point, I should have remembered that. But, hey, you don't read articles, nor man pages, before posting comments.
This discrepancy could of course be able to create other problems, especially if the system actually handles the leap second "properly", not just resetting the clock.
Remember that a lot of systems use "seconds since certain point in time" (like January 1, 1970, GMT, you UNIX-based bastards). There are functions in the C runtime library to convert from those to normal calendar dates. I find it kind of unsatisfactory that you'll have to add in magic numbers for each of the leap seconds as they are added. Of course, having a leap hour for some coders in a few centuries won't be too nice, but they will probably be able to declare that it's coming several years in advance, or ditch it altogether. One way or another, I would like to keep the calendar definition and conversion between different types simple. Avoiding leap seconds is one tiny step along that road.
That assumes that the Windows kernel (not other ring 0 services and subsystems, like Win32K) is a problem for Windows. I wouldn't say so. It's clearly not UNIX (the number of jokes regarding Cutley and VMS => WNT is enough to tell you that), but it clearly has some nice pieces of design. And both the Xbox and Windows XP embedded tell us you with a varying level of difficulty can turn Windows, with the NT kernel, into a quite different beast, with as much or as little of Win32 as you like.
Imagine that you are able to extract 1/4 of the data from your file system. But, hey, it's no problem, loads of your friends also run Linux, so you can just take the missing bits from those. The only problem is that you used a very odd patch to 2.0, along with odd software package for almost everything else.
Then, of course, it will matter a little how you fragment those around.
No, you won't be able to make a "good" dinosaur clone. A single nucleotide at a certain position MAY affect only behavior, or skin color, or some other property that we can't know the correct answer for. If we would have to "fill in" large parts from other reptiles/birds, then we maybe, MAYBE, would get a living creature, but we wouldn't be able to know how similar it was to the real thing.
Well, in one of the blog posts mentioning the RSOD, it was hinted that the color might be different depending on some assessment of if the problem was recurring or not, and things like that.
Good point. On the other hand, it's not like it's anything new. You could get Windows 95 to show BMP icons, there's the preview view of XP and just about any exectable has its icon extracted from the content.
In most of these cases, I think that it's due to one thing:
There is at least one thread or operation stuck in kernel space, handling the memory of this user process. It happens to me quite a lot when doing video capture with crappy drivers, but also in other situations where I can second-guess that it's all due to some weird I/O timeout never happening.
Of course, with a system profilerating in multi-threading (compared to UNIX where many apps still really are single-threaded, without any obvious flaws) and asynchronous I/O like Windows, it's not obvious how to solve the "pending kernel operation" state.
So, anyone would be more than welcome to fill in how their OS of choice handles killing a process that's waiting for a system call to finish.
3. Regarding the transaction logs, you gotta remember that SQL Server (and most databases, even if some popular ones for web applications are a bit more lax here) goes to quite great lengths, at least theoretically, to never commit a transaction until the log has been properly flushed. In this specific situation, battery-backed RAM may be a quite different thing than normal volatile RAM. Properly used, a quite limited amount of non-volatile fast solid state storage could also be used for journaling of file system operations. We could basically have a much more lazy attitude towards write-behind caching and other kinds of out-of-order writing, as we won't lose the data when/if we lose power.
Some people might. I predict that the/. crowd will for a quite long time find use for more data than you can fit into a keychain device, if only for maintaining our egos.
Your math is flawed. A 1 MHz clock means 20,000 cycles waiting for a 20 ms seek. Not, it would be in the order of 15,000,000 cycles.
So, your relative numbers are right, but we should remember that, even if an Apple II, you couldn't even come close to treating random harddrive access as "almost-RAM".
The problems of RAM latency (and the need for larger and larger caches to try to hide it) is more of a qualitative change, because in a flat-memory architecture, you can far too easily as a high-level coder get the impression that all memory accesses are created equal. They are not.
Don't bother reading it. It's basically a crusade against general relativity based on points like "precession of Mercury is not explained but is described", like you could somehow have a theory that really "exaplined" anything at all.
There is also some glorious flaws in logic, like that a photon travelling a longer distance (in non-expanding space) would be reddened and that the seemingly red sun when it dawns would be caused by a longer distance between the Sun and Earth. It's not like a longer distance through the atmosphere, due to the low angle, would enhance the effects that all day long makes the sky blue.
The observation that gravity couldn't be transmitted through vacuum, as vacuum doesn't contain anything, and therefore space couldn't be curved (because the curvation would be "something") is also... interesting.
Now, there may be flaws in GR, but we won't expose them by reading Newton in manuscript, like these guys do.
Slashdot Mirror
Show how XPSP2 IE6 installs spyware without user action, without the user being given the option and a security warning. Imagine Gator informing the user, in detail, how to allow a Firefox extension being installed to gain great functionality. (Or, of course, just getting the user to install local software, which then plugs in the extension.) And add the hopefully enhanced least user privilege support in Vista. (hopefully keeping more people from using a real full Administrator as the normal use account)
Hopefully, most of those are using IE quirks mode, as they shouldn't know about DOCTYPE. Just a slight hope.
Sorry. The point is that both parties are using highly directional antennas. And, well, a single access point wouldn't be able to serve much bandwidth with that many users anyway. (As the contention management protocols are not really designed for 1000+ users, you will get even less than your theoretical slice.)
You don't have to sniff to find his head lying outside San Francisco, even though it may help. Then it should be easy to steal him.
Yeah. The question is if any hairstyling has ever been applied to billg.
If the patent should cover tests, or just artificial use of the gene in other organisms may be worth some discussion, but there are many inventions that are mostly based on the fact that someone has tested a lot of stuff to find what works and what doesn't. The fact that a lot of people might have hit an optimum alloy composition by accident is not prior art for a patent describing a specific alloy, including its wonderful properties.
There sure is "prior" for it. If it would also be considered "art", well, maybe...
It makes no sense to use the same reusable design for manned vs. unmanned, at all. (Just as any cargo-with-safe-reentry should be quite different from just a cargo payload up in the sky.)
Hey, they could take $20,000 to go up, $180,000 to go back down. On the other hand, that might attract the suicidal demographic, of several kinds.
If a search engine actively filters content, it's one thing. If the chinese proxies mangle the results to exclude content, they are (obviously) not responsible. Likewise, if it's just a matter of using the general flexible Cisco features for a repressionist purpose, I wouldn't blame Cisco. Drawing the line here can get quite tricky, of course, as a new general feature could be introduced after requests from "an important customer"...
This discrepancy could of course be able to create other problems, especially if the system actually handles the leap second "properly", not just resetting the clock.
Remember that a lot of systems use "seconds since certain point in time" (like January 1, 1970, GMT, you UNIX-based bastards). There are functions in the C runtime library to convert from those to normal calendar dates. I find it kind of unsatisfactory that you'll have to add in magic numbers for each of the leap seconds as they are added. Of course, having a leap hour for some coders in a few centuries won't be too nice, but they will probably be able to declare that it's coming several years in advance, or ditch it altogether. One way or another, I would like to keep the calendar definition and conversion between different types simple. Avoiding leap seconds is one tiny step along that road.
That assumes that the Windows kernel (not other ring 0 services and subsystems, like Win32K) is a problem for Windows. I wouldn't say so. It's clearly not UNIX (the number of jokes regarding Cutley and VMS => WNT is enough to tell you that), but it clearly has some nice pieces of design. And both the Xbox and Windows XP embedded tell us you with a varying level of difficulty can turn Windows, with the NT kernel, into a quite different beast, with as much or as little of Win32 as you like.
Then, of course, it will matter a little how you fragment those around.
No, you won't be able to make a "good" dinosaur clone. A single nucleotide at a certain position MAY affect only behavior, or skin color, or some other property that we can't know the correct answer for. If we would have to "fill in" large parts from other reptiles/birds, then we maybe, MAYBE, would get a living creature, but we wouldn't be able to know how similar it was to the real thing.
Well, in one of the blog posts mentioning the RSOD, it was hinted that the color might be different depending on some assessment of if the problem was recurring or not, and things like that.
RTFA, they are killing the processes after boot-up. For winlogon, this means they can't log off properly.
It should be no surprise that networking can get quite strange without DNS Client and DHCP Client (among others)...
Good point. On the other hand, it's not like it's anything new. You could get Windows 95 to show BMP icons, there's the preview view of XP and just about any exectable has its icon extracted from the content.
Of course, with a system profilerating in multi-threading (compared to UNIX where many apps still really are single-threaded, without any obvious flaws) and asynchronous I/O like Windows, it's not obvious how to solve the "pending kernel operation" state.
So, anyone would be more than welcome to fill in how their OS of choice handles killing a process that's waiting for a system call to finish.
3. Regarding the transaction logs, you gotta remember that SQL Server (and most databases, even if some popular ones for web applications are a bit more lax here) goes to quite great lengths, at least theoretically, to never commit a transaction until the log has been properly flushed. In this specific situation, battery-backed RAM may be a quite different thing than normal volatile RAM. Properly used, a quite limited amount of non-volatile fast solid state storage could also be used for journaling of file system operations. We could basically have a much more lazy attitude towards write-behind caching and other kinds of out-of-order writing, as we won't lose the data when/if we lose power.
Some people might. I predict that the /. crowd will for a quite long time find use for more data than you can fit into a keychain device, if only for maintaining our egos.
So, your relative numbers are right, but we should remember that, even if an Apple II, you couldn't even come close to treating random harddrive access as "almost-RAM".
The problems of RAM latency (and the need for larger and larger caches to try to hide it) is more of a qualitative change, because in a flat-memory architecture, you can far too easily as a high-level coder get the impression that all memory accesses are created equal. They are not.
The roadmaps for next year show both dual-core and (I think) EM64T support in the laptop line.
Well, the load problems would be alleviated by dual-core CPUs! Total parallelism, since the ad has nothing to do with the game, at all...
There is also some glorious flaws in logic, like that a photon travelling a longer distance (in non-expanding space) would be reddened and that the seemingly red sun when it dawns would be caused by a longer distance between the Sun and Earth. It's not like a longer distance through the atmosphere, due to the low angle, would enhance the effects that all day long makes the sky blue.
The observation that gravity couldn't be transmitted through vacuum, as vacuum doesn't contain anything, and therefore space couldn't be curved (because the curvation would be "something") is also... interesting.
Now, there may be flaws in GR, but we won't expose them by reading Newton in manuscript, like these guys do.