Chrome (10.7%) + Safari (6.3%) = 17% with approx 7%/mo growth (yes, I know there are others, but the percentages are very small). In about 6 months, Webkit based browsers should surpass Mozilla derived browsers to become the 2nd most popular. Since both Mozilla and Webkit derived browsers will then be near 25%, add in Opera to push the total over 50% and finally IE will be less than half the web traffic.
Besides, any of those IPv4 bandaid type solutions just delay the impetus to move to IPv6. If you are going to make people check and patch their equipment to support future use addresses, you are going to take away resources that could be focused on moving to v6.
Dual stack will allow people to connect to those who can only get a new v6 address (i.e. a v4 peer will be able to connect to a v6 socket). That is a better solution than making the IPv4 situation even more broken than it already is.
Your approach breaks connectivity for more people, but it doesn't reduce support in the slightest. My approach is based upon a dual stack model, and it reduces the problems for the end user, and doesn't delay the implementation of full v6 capabilities at all because v6 is the preferred protocol, as specified in the RFCs. v4 is used as a fallback when v6 isn't available.
If their HW/SW (and their ISP, etc.) supports v6, they'll just connect using v6. However, if their HW/SW doesn't support v6, and there is a formerly reserved v4 address, they can at least attempt the connection (i.e. no v6, potential v4 connection). Without that, they can't get to the destination until they upgrade their HW/SW and verify their ISP provides v6 connectivity (i.e. no v6, no connection)
Support is no harder because you try v6 first. If that fails, they try v4 knowing that some devices won't be able to connect. If the users wants to go through the process of checking a "huge list of incompatible equipment", they can, if not, well, they're no worse off than they will be by assigning v6 addresses after the current v4s run out.
I fail to see how some users not being able to reach them on v4 due to outdated routers is worse than far more users not being able to reach them because they have only a v6 address. Clearly v6 support to the end node is still very limited. Make the previously reserved v4 addresses available.
If it were used in a MacBook or Mac Mini, it probably wouldn't affect any users. To date, Apple hasn't introduced a MB/MBP/mini that uses more than 2 SATA connections. Assuming that remains true for any SB based machines, and that they would use the 2 6Gbps SATA ports, this bug would not affect those machines.
In short, given what's currently known and likely, "you're using it wrong" would be correct. Of course, that's all speculation since Apple hasn't yet released machines using SB.
However, most of the slowdown over time is due to Internet Explorer. The problem was made worse by integrating IE into Windows rather than making it an app like ever other app. Reduce the IE cache size to 10MB and set "empty temporary items on exit" and Windows performance doesn't degrade nearly as badly. It'll still degrade due to installing crapware and every program installing it's own notification/task bar utility that runs at startup, and it may degrade if you don't defragment the HD periodically, but the IE cache is the biggest culprit. Of course, if you only use IE for Windows Update, then IE's cache isn't as much of a factor.
I can't imagine a hardware bug that would manifest only as degraded performance after extended uptime. Anything of that nature could probably be worked around with a software fix that periodically reset the controller. Therefore, I think it's safe to assume it's literally the SATA logic degrading with age, which would require a chip level change.
And what do you call the people in the FBI's Counterintelligence Division? What do you think they do? The CIA is not chartered to operate counterintelligence operation in the US, that's the FBI's job.
How do you divide space? In the vacuum of space, what constitutes a clear, measurable boundary?
Geographical boundaries are often based upon some physical feature, a body of water, river, mountain, etc. Some are arbitrary, such as a line of latitude or longitude, but it's something that's pretty easy to identify and measure. As far as we know, you don't have such neat boundaries in interstellar space.
Even the concept of "gravitationally bound" isn't sufficient, as stars and planets have objects gravitationally bound to them, and clusters of stars are gravitationally bound to each other, and the stars and clusters are bound to the "galaxy", and clusters of galaxies are gravitationally bound to one another. Where does one level end and the next level begin?
With the possible exception of the PS3, when is the last time Sony marketed a product that you might actually want to buy (for it's features and/or price)? Ok, maybe some movies they market.
Sony has made it very easy to not buy their products. It's a great example of how to destroy a huge company:
1. Stop making great products that people want to buy.
Check out the history section in the above link for the awesome story of how the guy who discovered the link between H. Pylori and ulcers got people to take his results seriously. Its one of my favorites.
So you're saying Africa is to blame?;) It is an interesting history.
Slashdot Mirror
You don't follow the math correctly.
Chrome (10.7%) + Safari (6.3%) = 17% with approx 7%/mo growth (yes, I know there are others, but the percentages are very small). In about 6 months, Webkit based browsers should surpass Mozilla derived browsers to become the 2nd most popular. Since both Mozilla and Webkit derived browsers will then be near 25%, add in Opera to push the total over 50% and finally IE will be less than half the web traffic.
Besides, any of those IPv4 bandaid type solutions just delay the impetus to move to IPv6. If you are going to make people check and patch their equipment to support future use addresses, you are going to take away resources that could be focused on moving to v6.
Dual stack will allow people to connect to those who can only get a new v6 address (i.e. a v4 peer will be able to connect to a v6 socket). That is a better solution than making the IPv4 situation even more broken than it already is.
Your approach breaks connectivity for more people, but it doesn't reduce support in the slightest. My approach is based upon a dual stack model, and it reduces the problems for the end user, and doesn't delay the implementation of full v6 capabilities at all because v6 is the preferred protocol, as specified in the RFCs. v4 is used as a fallback when v6 isn't available.
If their HW/SW (and their ISP, etc.) supports v6, they'll just connect using v6. However, if their HW/SW doesn't support v6, and there is a formerly reserved v4 address, they can at least attempt the connection (i.e. no v6, potential v4 connection). Without that, they can't get to the destination until they upgrade their HW/SW and verify their ISP provides v6 connectivity (i.e. no v6, no connection)
Support is no harder because you try v6 first. If that fails, they try v4 knowing that some devices won't be able to connect. If the users wants to go through the process of checking a "huge list of incompatible equipment", they can, if not, well, they're no worse off than they will be by assigning v6 addresses after the current v4s run out.
I fail to see how some users not being able to reach them on v4 due to outdated routers is worse than far more users not being able to reach them because they have only a v6 address. Clearly v6 support to the end node is still very limited. Make the previously reserved v4 addresses available.
You do NOT talk about the final five!
Final five? I thought is was the Final Four, and that it was usually in March.
If it were used in a MacBook or Mac Mini, it probably wouldn't affect any users. To date, Apple hasn't introduced a MB/MBP/mini that uses more than 2 SATA connections. Assuming that remains true for any SB based machines, and that they would use the 2 6Gbps SATA ports, this bug would not affect those machines.
In short, given what's currently known and likely, "you're using it wrong" would be correct. Of course, that's all speculation since Apple hasn't yet released machines using SB.
Funny. Actually, hilarious. +1
However, most of the slowdown over time is due to Internet Explorer. The problem was made worse by integrating IE into Windows rather than making it an app like ever other app. Reduce the IE cache size to 10MB and set "empty temporary items on exit" and Windows performance doesn't degrade nearly as badly. It'll still degrade due to installing crapware and every program installing it's own notification/task bar utility that runs at startup, and it may degrade if you don't defragment the HD periodically, but the IE cache is the biggest culprit. Of course, if you only use IE for Windows Update, then IE's cache isn't as much of a factor.
I can't imagine a hardware bug that would manifest only as degraded performance after extended uptime. Anything of that nature could probably be worked around with a software fix that periodically reset the controller. Therefore, I think it's safe to assume it's literally the SATA logic degrading with age, which would require a chip level change.
...it looks like they learned from the Pentium FP fiasco and are handling this one correctly.
When you think about it, Chrome/chromium is a silly name for a browser. Chromium is brittle. Who wants a brittle browser?
With Graphene, I can scribble it on Scotch Tape and get a Nobel Prize; can I do that with Molybdenite?
Yes. RTFA. They don't have a Nobel prize yet, but they did use scotch tape to pull the layer of MoS2 from a rock
Investigation != illegal abuse of power.
You don'1^2=1; (-1)^2=1; 1^2=(-1)^2; 1=-1; 2=0.
There, FTFY. If you're gonna have a effective, defective proof, you can't leave in the obvious errors.
Sounds promising:
Advantages:
Disadvantages:
While Score < 4
mod +1, Informative.
Wend
And what do you call the people in the FBI's Counterintelligence Division? What do you think they do? The CIA is not chartered to operate counterintelligence operation in the US, that's the FBI's job.
Who would have ever guessed that a group of spies would break the law or lie?
Please provide an example of how it should work.
To the idiot who moderated this offtopic, try again. The comment is exactly on topic.
The article was published over 2 months ago and it was well publicized then. Why is it showing up on /. now?
...they're running IPv4. They need update to IPv6, then they won't have those problems.
How do you divide space? In the vacuum of space, what constitutes a clear, measurable boundary?
Geographical boundaries are often based upon some physical feature, a body of water, river, mountain, etc. Some are arbitrary, such as a line of latitude or longitude, but it's something that's pretty easy to identify and measure. As far as we know, you don't have such neat boundaries in interstellar space.
Even the concept of "gravitationally bound" isn't sufficient, as stars and planets have objects gravitationally bound to them, and clusters of stars are gravitationally bound to each other, and the stars and clusters are bound to the "galaxy", and clusters of galaxies are gravitationally bound to one another. Where does one level end and the next level begin?
With the possible exception of the PS3, when is the last time Sony marketed a product that you might actually want to buy (for it's features and/or price)? Ok, maybe some movies they market.
Sony has made it very easy to not buy their products. It's a great example of how to destroy a huge company:
Check out the history section in the above link for the awesome story of how the guy who discovered the link between H. Pylori and ulcers got people to take his results seriously. Its one of my favorites.
So you're saying Africa is to blame? ;) It is an interesting history.