Smoking in China seems to be following the same generational shifts that it did in the US. Far fewer 30-somethings smoke than 50-somethings, and 20-somethings that smoke are fewer still.
She's involved with the MIT Media Lab and a breadboard with a few LEDs and a battery is supposed to constitute a project?!!!??! One which she's proud of?
Nicholas Negroponte needs to spend less time on OLPC and get back to teaching.
The system isn't picking something for you... The idea is that by giving you an "original" composition with features similar to music you already like, you will inevitably like the new composition as well.
Which makes Eigenradio an elaborate comment on how mass-market "cultural" products are engineered rather than created.
Seems like the checksum approach could be its own punishment for the software company -- first time a major virus hits and they get spammed by their own anti-theft code, I'd imagine they'd rethink the whole thing. They'd either have to follow-up on all of the bogus reports and try to re-verify all of the valid licenses, or they'd have to ignore all reports until the virus had run its course (probably a month or so because of stragglers).
Not only would it cost them a ton, but they'd alienate huge chunks of their customers.
I think that if, as the original post says, it only phones home when it thinks it's been truly cracked (ie tampered with) that's OK.
I have concerns if:
1. It phones home periodically on legit installations too: if I have a legally licensed copy, then the software co. really has no right to collect data (even if its inadvertent) about how often I'm using their software, what kind of computer I use, etc...
2. It phones home if the software is untampered with but "suspects" that I've used a serial # twice, etc. Why? Because, and maybe this is on the slippery slope, the world needs some form of wiggle-room built-in -- while its technically a violation to install the same package on my home desktop and my home laptop at the same time, in my case its a fact that that package will _never_ be used on both at once -- the software co isn't losing anything. I think we do justice far more damage by not overlooking this. To me this is the same case as with too much surveillance and tracking -- we all commit minor indiscretions sometimes.
I agree -- most non-geeks shouldn't have to worry about such things (and likely don't even know that they might need to).
But DefCon isn't an average-joe situation -- I'm amazed that the attendees at a conference like DefCon wouldn't know better than to wander around a conference filled with other geeks surfing/mailing/etc over WiFi without at least using SSH.
While we're all waiting for plague to descend on good 'ol Darl and his league of flying monkeys (read legal department), what about creating a "clean" kernel that they don't have claims against?
Since 2.2 apparently doesn't infringe, why not create a super 2.2 kernel and swap it in for the (allegedly) infringing newer kernels on as many systems as possible?
Here's what I'm thinking/wondering: 1. How many Linux users actually need/use the components that IBM contributed? 2. How much non-infringing post-2.2 stuff can be back-ported to the 2.2 kernel? 3. If you managed to back-port as much as possible and polish-up a 2.2 kernel as much as it can be polished, will it meet the needs of most users?
I'm curious if the GPL argument holds in this case (though I hope it does).
Here's what I'm thinking: SCO could easily say that they were previously distributing Linux according to the GPL, having included the source (per the GPL) to the improvements they knowingly contributed.
But now, the story is different because they found out (GASP!) that the kernel includes code that belongs to them but that was improperly contributed by somebody else without SCO's permission. Once they learned that this improper code was there, they ceased distributing Linux.
I don't know if this makes sense -- just trying to reason this out.
I look forward to seeing the release version of 7, especially since so many sites are now using WMP.
The beta of 7 also has a really bizarre looking interface (trying to cop other media layers tech-metal-looking skins but failing). Hoping they clean this up as well -- IE5 has a nice interface, so we know they can do it.
Windows Media Player 7 for Mac is still only in beta. I've tried both the beta of 7 and the release version of 6.3, and they're both slow and buggy, and the streams seem to stall ALOT and for long periods (5-20+ seconds on a DSL connection) -- much worse than QT and Real streams from the same sites.
What's being proposed is that 'impossible' packets be rejected out of hand -- this wouldn't restrict any traffic from any users or subnets, because no legitimate user or subnet would be sending traffic with these 'impossible' addresses.
What I'm calling 'impossible' packets are packets with addresses that defy the way things are configured. The two examples I've seen in the thread so far are reserved address ranges, and internal address ranges. I'll explain...
1. The IP standard reserves certain address ranges (198.168.1.xxx is one) for use only on internal networks. This means that no packets with source addresses of 198.168.1.xxx should EVER be found out on the internet at large, because nobody on the internet at large legitimately owns these addresses. So if an ISP receives packets from an address in this range, then it is 100% guaranteed to be errant traffic (either something misconfigured or a deliberate attack).
2. ISP's own certain blocks of IP addresses, which they assign to their customers. So... If an ISP owns a certain IP address, then it is 100% guaranteed that packets with that IP source address will only cross their border router coming from inside their network and heading out to the internet -- the border router should never receive a packet with this IP coming from the outside and headed towards the inside.
It sounds like the other thing to watch for would be packets from your own legitimate users that have source addresses that don't fall within your IP range -- this would help to cure the disease, where everything else is treating symptoms.
The patent says that (paraphrased) the link text is the first part, the address is the second part.
In HTML, the address (HREF attribute) comes first, then the link text.
May seem silly, but Motorola patented cell-phones with flip-down microphones. So what did the competition do? They created cell-phones with flip-up earpieces.
I know I'm nit-picking, but maybe it'll work. Any thoughts?
Actually, I've read a few articles that discuss Rambus architecture in detail and why Rambus doesn't live-up to its latency and speed claims.
They were way beyond my technical understanding, but the conclusion was that Rambus RAM in real world situations would be far slower than DDR-SDRAM, and even (sometimes) slower than existing PC133-SDRAM. Sounded convincing.
I'm afraid I don't have links handy, but if you're interested, check-out some of the past/. discussions of Rambus -- this is where I found the links originally.
If these articles are right, then Rambus is charging more for an inferior product, and through this patent maneuvering, are trying to kill-off the competing (and superior and cheaper) product by forcing it's price above Rambus RAM through questionable royalties. Legal or not, it isn't fair or desireable corporate beavior, and it's far from the free-market ideal of best-product-wins.
This is late enough that nobody will read it... but...
I agree that MS online properties don't have the presence necessary to unfairly promote other Microsoft ventures.
I was suggesting that Microsoft has in the past, and will now be able to continue, to use their dominance in certain apps (Office, IE) to give preferred placement to Microsoft's online properties -- whether they're successful or not, it's still an unfair trade practice to use dominance in one market to give unfair advantage to your product in another market.
Two good examples of MS already doing this are: Outlook's automatic setup of new Hotmail accounts, and MSN being the default startpage in IE (side note: in my work copy of IE, MSN was the startpage, and the box in preferences was ghosted, preventing changes -- I had to go to the registry to change it).
Apologies in advance if I've misunderstood your point.;->
This sounds like a complicated (and pixel-centric) way of thinking about using floating dpi-ratings based on actual physical size of the monitor -- text that is 1/4" high is 1/4" high regardless of the monitor, but on a 21" 3072x2304 monitor it is clearer than on a 21" 1024x768 monitor.
As we approach higher pixel densities, it seems like this would be a good shift for interfaces to make, since you can't assume that a 3072x2304 monitor is 72dpi anymore, and having text get smaller is a silly trade-off for clearer images. It's more intuitive anyway, as print material works this way.
Now I'm really straying, but maybe monitor profiles should contain the screen's actual physical size so that the OS can calculate the effective DPI of the display and size everything appropriately. I'd imagine that approaches like these will become increasingly feasible with UI's based on vector engines (like Apple's Quartz)--describe UI elements as vectors, and render them on-the-fly at a consistent physical size based on the monitor's resolution.
Ok, so CT apparently looks good, and is therefore an asset on LCD's. Any thoughts on whether other groups (Apple, Linux people, Be, Transmeta etc...) will work something like this into non-MS systems? My impression from previous posts is that MS can't claim exclusive use of the technique, only their particular algorithm.
My other question is this: where can we go from here with CRT displays? Are there better algorithms out there than what's currently and commonly used for anti-aliasing (ala Adobe Type Manager, etc)? If so, why haven't we seen them yet, and when can we expect them?
"Byron noted that utilities can promise only 99.9 percent reliability--a figure that translates to about eight hours of blackouts a year--while high-tech firms stand to lose millions of dollars from a blackout lasting just a minute."
I've heard estimates like this before (millions lost per minute). Maybe I'm looking at it the wrong way, but for e-commerce this seems exaggerated -- if I hit a site intending to buy something and it's down, I'll definitely try back in 1,2,5,10 minutes.
If a site is down for 1 or 2 minutes, how would I know the difference between that and an overloaded server or high net traffic anyway?
I can see longer outages causing lost business, but not minutes-long outages.
You're right -- legally, I don't see a difference.
To me, the real difference is that software companies insist (advertise, market) that their software provides a benefit, and therefore makes users pay them for it. But if that benefit is nullified by a quality problem in the software, the company insists that the user has no recourse.
By not charging for their software, GPL authors aren't making promises (implicit or explicit) that the product is stable or useful, and therefore shouldn't be accountable if the software turns out to be unstable or useless.
It seems like the process of rushing (what otherwise would have been) beta software out the door to beat-out the competition for market share might be stunted if software companies were less fortified by EULA's.
Seems like a huge mistake was made in not breaking-off MS's online ventures. MS is obviously staking alot on MSN, MSN-mobile, MSNBC, etc...
If these properties are in the same company as applications, why WON'T they continue to give their sites preferential treatment in built-in channel lists, etc...
Slashdot Mirror
Smoking in China seems to be following the same generational shifts that it did in the US. Far fewer 30-somethings smoke than 50-somethings, and 20-somethings that smoke are fewer still.
She's involved with the MIT Media Lab and a breadboard with a few LEDs and a battery is supposed to constitute a project?!!!??! One which she's proud of?
Nicholas Negroponte needs to spend less time on OLPC and get back to teaching.
The system isn't picking something for you... The idea is that by giving you an "original" composition with features similar to music you already like, you will inevitably like the new composition as well.
Which makes Eigenradio an elaborate comment on how mass-market "cultural" products are engineered rather than created.
All of the 2Wire routers I've seen, which are, I think, distributed by DSL providers, seem to have WEP turned on.
So some provider is doing something right.
This would be more akin to using data about which cars are stolen most as marketing data about which cars are most desirable.
Don't forget:
Cathy
Hagar the Horrible
B.C.
These guys are all still beating the same old tired themes/jokes to death.
Seems like the checksum approach could be its own punishment for the software company -- first time a major virus hits and they get spammed by their own anti-theft code, I'd imagine they'd rethink the whole thing. They'd either have to follow-up on all of the bogus reports and try to re-verify all of the valid licenses, or they'd have to ignore all reports until the virus had run its course (probably a month or so because of stragglers).
Not only would it cost them a ton, but they'd alienate huge chunks of their customers.
I think that if, as the original post says, it only phones home when it thinks it's been truly cracked (ie tampered with) that's OK.
I have concerns if:
1. It phones home periodically on legit installations too: if I have a legally licensed copy, then the software co. really has no right to collect data (even if its inadvertent) about how often I'm using their software, what kind of computer I use, etc...
2. It phones home if the software is untampered with but "suspects" that I've used a serial # twice, etc. Why? Because, and maybe this is on the slippery slope, the world needs some form of wiggle-room built-in -- while its technically a violation to install the same package on my home desktop and my home laptop at the same time, in my case its a fact that that package will _never_ be used on both at once -- the software co isn't losing anything. I think we do justice far more damage by not overlooking this. To me this is the same case as with too much surveillance and tracking -- we all commit minor indiscretions sometimes.
I agree -- most non-geeks shouldn't have to worry about such things (and likely don't even know that they might need to).
But DefCon isn't an average-joe situation -- I'm amazed that the attendees at a conference like DefCon wouldn't know better than to wander around a conference filled with other geeks surfing/mailing/etc over WiFi without at least using SSH.
While we're all waiting for plague to descend on good 'ol Darl and his league of flying monkeys (read legal department), what about creating a "clean" kernel that they don't have claims against?
Since 2.2 apparently doesn't infringe, why not create a super 2.2 kernel and swap it in for the (allegedly) infringing newer kernels on as many systems as possible?
Here's what I'm thinking/wondering:
1. How many Linux users actually need/use the components that IBM contributed?
2. How much non-infringing post-2.2 stuff can be back-ported to the 2.2 kernel?
3. If you managed to back-port as much as possible and polish-up a 2.2 kernel as much as it can be polished, will it meet the needs of most users?
I'm curious if the GPL argument holds in this case (though I hope it does).
Here's what I'm thinking: SCO could easily say that they were previously distributing Linux according to the GPL, having included the source (per the GPL) to the improvements they knowingly contributed.
But now, the story is different because they found out (GASP!) that the kernel includes code that belongs to them but that was improperly contributed by somebody else without SCO's permission. Once they learned that this improper code was there, they ceased distributing Linux.
I don't know if this makes sense -- just trying to reason this out.
I look forward to seeing the release version of 7, especially since so many sites are now using WMP. The beta of 7 also has a really bizarre looking interface (trying to cop other media layers tech-metal-looking skins but failing). Hoping they clean this up as well -- IE5 has a nice interface, so we know they can do it.
Windows Media Player 7 for Mac is still only in beta. I've tried both the beta of 7 and the release version of 6.3, and they're both slow and buggy, and the streams seem to stall ALOT and for long periods (5-20+ seconds on a DSL connection) -- much worse than QT and Real streams from the same sites.
What's being proposed is that 'impossible' packets be rejected out of hand -- this wouldn't restrict any traffic from any users or subnets, because no legitimate user or subnet would be sending traffic with these 'impossible' addresses.
What I'm calling 'impossible' packets are packets with addresses that defy the way things are configured. The two examples I've seen in the thread so far are reserved address ranges, and internal address ranges. I'll explain...
1. The IP standard reserves certain address ranges (198.168.1.xxx is one) for use only on internal networks. This means that no packets with source addresses of 198.168.1.xxx should EVER be found out on the internet at large, because nobody on the internet at large legitimately owns these addresses. So if an ISP receives packets from an address in this range, then it is 100% guaranteed to be errant traffic (either something misconfigured or a deliberate attack).
2. ISP's own certain blocks of IP addresses, which they assign to their customers. So... If an ISP owns a certain IP address, then it is 100% guaranteed that packets with that IP source address will only cross their border router coming from inside their network and heading out to the internet -- the border router should never receive a packet with this IP coming from the outside and headed towards the inside.
It sounds like the other thing to watch for would be packets from your own legitimate users that have source addresses that don't fall within your IP range -- this would help to cure the disease, where everything else is treating symptoms.
The patent says that (paraphrased) the link text is the first part, the address is the second part.
In HTML, the address (HREF attribute) comes first, then the link text.
May seem silly, but Motorola patented cell-phones with flip-down microphones. So what did the competition do? They created cell-phones with flip-up earpieces.
I know I'm nit-picking, but maybe it'll work. Any thoughts?
Actually, I've read a few articles that discuss Rambus architecture in detail and why Rambus doesn't live-up to its latency and speed claims.
They were way beyond my technical understanding, but the conclusion was that Rambus RAM in real world situations would be far slower than DDR-SDRAM, and even (sometimes) slower than existing PC133-SDRAM. Sounded convincing.
I'm afraid I don't have links handy, but if you're interested, check-out some of the past /. discussions of Rambus -- this is where I found the links originally.
If these articles are right, then Rambus is charging more for an inferior product, and through this patent maneuvering, are trying to kill-off the competing (and superior and cheaper) product by forcing it's price above Rambus RAM through questionable royalties. Legal or not, it isn't fair or desireable corporate beavior, and it's far from the free-market ideal of best-product-wins.
And that IS horrible.
This is late enough that nobody will read it... but...
I agree that MS online properties don't have the presence necessary to unfairly promote other Microsoft ventures.
I was suggesting that Microsoft has in the past, and will now be able to continue, to use their dominance in certain apps (Office, IE) to give preferred placement to Microsoft's online properties -- whether they're successful or not, it's still an unfair trade practice to use dominance in one market to give unfair advantage to your product in another market.
Two good examples of MS already doing this are: Outlook's automatic setup of new Hotmail accounts, and MSN being the default startpage in IE (side note: in my work copy of IE, MSN was the startpage, and the box in preferences was ghosted, preventing changes -- I had to go to the registry to change it).
Apologies in advance if I've misunderstood your point. ;->
This sounds like a complicated (and pixel-centric) way of thinking about using floating dpi-ratings based on actual physical size of the monitor -- text that is 1/4" high is 1/4" high regardless of the monitor, but on a 21" 3072x2304 monitor it is clearer than on a 21" 1024x768 monitor.
As we approach higher pixel densities, it seems like this would be a good shift for interfaces to make, since you can't assume that a 3072x2304 monitor is 72dpi anymore, and having text get smaller is a silly trade-off for clearer images. It's more intuitive anyway, as print material works this way.
Now I'm really straying, but maybe monitor profiles should contain the screen's actual physical size so that the OS can calculate the effective DPI of the display and size everything appropriately. I'd imagine that approaches like these will become increasingly feasible with UI's based on vector engines (like Apple's Quartz)--describe UI elements as vectors, and render them on-the-fly at a consistent physical size based on the monitor's resolution.
blah, blah, blah... Sorry.
Ok, so CT apparently looks good, and is therefore an asset on LCD's. Any thoughts on whether other groups (Apple, Linux people, Be, Transmeta etc...) will work something like this into non-MS systems? My impression from previous posts is that MS can't claim exclusive use of the technique, only their particular algorithm.
My other question is this: where can we go from here with CRT displays? Are there better algorithms out there than what's currently and commonly used for anti-aliasing (ala Adobe Type Manager, etc)? If so, why haven't we seen them yet, and when can we expect them?
I've heard estimates like this before (millions lost per minute). Maybe I'm looking at it the wrong way, but for e-commerce this seems exaggerated -- if I hit a site intending to buy something and it's down, I'll definitely try back in 1,2,5,10 minutes.
If a site is down for 1 or 2 minutes, how would I know the difference between that and an overloaded server or high net traffic anyway?
I can see longer outages causing lost business, but not minutes-long outages.
You're right -- legally, I don't see a difference.
To me, the real difference is that software companies insist (advertise, market) that their software provides a benefit, and therefore makes users pay them for it. But if that benefit is nullified by a quality problem in the software, the company insists that the user has no recourse.
By not charging for their software, GPL authors aren't making promises (implicit or explicit) that the product is stable or useful, and therefore shouldn't be accountable if the software turns out to be unstable or useless.
It seems like the process of rushing (what otherwise would have been) beta software out the door to beat-out the competition for market share might be stunted if software companies were less fortified by EULA's.
Seems like a huge mistake was made in not breaking-off MS's online ventures. MS is obviously staking alot on MSN, MSN-mobile, MSNBC, etc...
If these properties are in the same company as applications, why WON'T they continue to give their sites preferential treatment in built-in channel lists, etc...