Somewhere there ought to be a place for HP or Dell or both to get out of the race to the bottom and actually produce high quality hardware with first class customer service, at a price ~30-40% above what the no-service no-name brand of the day charges.
Telephone support I could live without. But both HP and Dell's websites make me ill. There seems to be a problem out there where as soon as marketing gets in charge of a web site they water it down, take all the useful information out, and make it impossible to use to get anything done.
And service manual? Can anyone say service manual? Or a way to easily find and purchase important spare parts like fans? As opposed to "accessories"? Or information about which chipsets are used in a certain model, so you can check hardware compatibility?
Sun used to be pretty good (borderline excellent) in this area, but Oracle is killing that off in a hurry. Oracle's website and visible level of support is far worse than Sun's in any case. Oracle apparently does not want to be a small and medium business market provider of anything.
I use Chrome all the time, but I always go to another browser to print anything. Internet Explorer's printing support isn't all that great (always cutting stuff off on the right instead of scaling for example), but Chrome's (at least on Windows XP) is positively pathetic. It looks like a kindergartner did the kerning. More or less unreadable. I am looking forward to a fix for that.
The telephone company doesn't want to care whether it is a legal download. Common carriers are insulated from liability for the traffic that goes across the network. Even common carriers can refuse to provide service for blatantly illegal activity, and on occasion may be required by the government to do so.
Under normal conditions they just have no incentive to do so, in part because the burden of proof is on them instead of the other way around. Nobody who is largely insulated from legal liability wants to go around provoking lawsuits. To say nothing of the fact that electronic communications providers are generally prohibited from analyzing traffic content for the purpose of determining whether it is legal or not.
The only reason why Comcast won is that the FCC was trying to regulate the Internet under a section of the law that gave them absolutely no power to do so. The FCC is in a position to fix that problem, by regulating Internet access providers under Title II.
Avoiding Title II was criminally irresponsible on the part of the FCC in the first place, they got smacked down for it and now we all suffer the consequences of the FCC's previous decision to operate under a purely fictional legal regime. Wishful thinking doesn't usually carry you very far in the legal system. The FCC has the proper tools, now they just have to use them.
If you actually want to test the limits of the bus, a test with SSDs would be far more realistic. Standards like this are designed to support what people will be doing a decade from now, not so much what they are doing today.
There are other more important improvements in USB 3.0 than raw bandwidth, mostly the transition from a polling based to a DMA/interrupt driven interface. This will help things immensely, even when only a fraction of the rated bandwidth is being used, due to less CPU overhead and power usage mostly. USB 3.0 is more like "USB grows up" than "USB 2.0 this time with feeling".
Ethernet now has 40 Gb/s and 100 Gb/s standards as well. Rather expensive and used only for special purpose applications at this point of course. Data centers, cluster interconnect, and WAN links mostly.
for example, there was a time when a foregone conclusion was one that was so unlikely you may as well not think about it
On the contrary, a foregone conclusion has _always_ been a conclusion that was so definite that it was like a conclusion that was made "before" and is "gone" such that there isn't any chance of any other.
It is your responsibility if you publish video to publish in the ISO/IEC standard.
Maybe if you live in a police state. Everywhere else, perhaps ISO standards would get a little more adoption if they were royalty and patent free. In this case the ISO is just acting as a shill for the MPEG LA.
If you don't care about speed or volume by all means buy an inkjet. I have an relatively inexpensive HP CP1215 color laser printer that will print 100 high quality full color full page photos in about half an hour.
The problem with this has been pointed out a gazillion times: updating the leap second table requires either connectivity or manual intervention, which are not available in all applications where people want to use UTC
Give me a break. A local UTC clock isn't accurate without connectivity or manual intervention the moment an unrecognized leap second passes either.
Using a TAI standard for the system clock has no real downside and lots of upsides. If you don't have an up-to-date leap second table, your idea of TAI might be off, but it won't make any difference as long as you use purpose appropriate timestamp storage formats. The "wall clock" conversion will always read what you expect it to be within the limitations of any astronomically derived time system.
And unlike a typical POSIX time_t implementation, the system will not require all sorts of crazy hacks to allow it to survive (or function correctly during) a leap second while simultaneously pretending it doesn't exist.
Sysadmins have to fix the problems that occur from departure bombs and dead man switches, once you've done that you develop an extreme dislike of them.
I am personally aware of one circumstance where this happened, and all I can say is that a sysadmin who does something like this deserves three to five years in the nearest federal penitentiary, if not more.
Because color laser 'photo quality" prints look like modern inkjet prints set to "fast draft"?
My experience is more the opposite. A decent color laser printer produces photo quality prints at about eight pages per minute. A typical inkjet printer produces comparable quality prints about 100 times that slow. Anything an inkjet produces in a reasonable amount of time makes dot matrix printers look like an improvement. Ugliest prints on the planet, ink spatter and all.
It is the way it is now for historical reasons. The debate is about how to change it. The drift is relatively slow, and it would be inconvenient to wait until the clock was off by an entire day (noon at "midnight" would be half way through the cycle).
However, there are serious advocates of the idea that we should drop leap seconds from civil time and then have a "leap hour" to adjust every thousand years or so. Then local noon would only be off at most one hour from what it is already off now.
It seems to be more likely though that computers will be changed to track TAI, an atomic time standard that doesn't have leap seconds, and then use some sort of leap second table to convert to "wall clock" or civil time, much as UTC is converted to local time with time zone adjustments now.
Getting rid of leap seconds in the representation would be a mistake in the long run. A much superior fix would be to have computers keep track of TAI internally and then convert to UTC with a leap second table, much the same way we convert to local time with a time zone table.
Cluster software should be running off of a leap second free distributed clock, and TAI or the equivalent is the best one we have, handily provided (within a constant) on a world wide basis courtesy of the GPS system.
POSIX time_t values as it stands today isn't much of a clock at all, more like a short hand for encoding a wall time. We should have a new interface for a reference clock that doesn't need unpredictable adjustments every couple of years, _by definition_.
Not tested in federal court, no, but rather before an administrative law judge who works for the USITC. However, the company whose products are blocked can appeal any adverse decision to the Court of Appeals for the Federal Circuit, which will do a more thorough job.
Intellectual property rights are exclusive rights. That's the way the law has designed them -- it's not a matter of Oracle being evil. Those IPRs entitle a right holder to enforce exclusivity.
In the case of many kinds of patents, and software patents in general, those exclusive rights cause far more harm than good. Congress should rewrite the whole system from the ground up.
(1) Patents should be restricted to a handful of fields. (2) Patent licensing should only be allowed to recover documented, non-overlapping R&D expenses associated with that patent plus perhaps a 100% gross margin. (3) Patent licensing should be mandatory at government regulated rates in line with that. (4) When the licensing recovery is reached, the patent should auto-expire
With rules like that, patents might actually promote the progress of the science and the useful arts, rather than grossly hindering it.
And a recursive DNS server patched to magically convert results from the internet from IPv6 to a special IPv4 address that the gateway can use to figure out where the IPv4 client really wanted to go?
It is much more likely that the gateway will do this for you, along with a bunch of other ugly v4v6 NAT and "transparent" proxying. It would probably have to strip DNSSEC of course. And the gateway will have to make up lots of synthetic IPv4 addresses for just the reason you describe, at least for any protocol it does not know how to rewrite.
As I understand it, Internet Explorer on XP will never support SNI, no matter what version you are running. That means SNI is impractical for most sites until the day comes the operators are willing to tell all remaining XP users to use a different browser. Three years maybe.
You're wrong on several counts, within 2-3 years your ISP will most likely switch you to IPv6.
It is _much_ more likely that most ISPs will deploy carrier grade NAT long before they start shipping new routers to every single one of their customers. That means two layers of NAT, one at the customer premises (as we have now) and one at the ISP.
Only people who get new routers or flash their current one (assuming it is new enough) are likely to be able to use IPv6 without a tunnel.
With IPv6, all new owners can talk to the old owners.
Strictly speaking this is not true. IPv6 is like a new parallel overlay network. No bidirectional communication occurs between IPv4 only hosts and IPv6 only hosts except through NAT (or application level proxies). Anyone who thinks NAT is going anyway any time soon is deluded.
Slashdot Mirror
They must be doing something right. The colocation center where we have our Sun boxes hosted is about 80% populated with Dell rackmounts.
Dell takes parts and puts them in a box.
Somewhere there ought to be a place for HP or Dell or both to get out of the race to the bottom and actually produce high quality hardware with first class customer service, at a price ~30-40% above what the no-service no-name brand of the day charges.
Telephone support I could live without. But both HP and Dell's websites make me ill. There seems to be a problem out there where as soon as marketing gets in charge of a web site they water it down, take all the useful information out, and make it impossible to use to get anything done.
And service manual? Can anyone say service manual? Or a way to easily find and purchase important spare parts like fans? As opposed to "accessories"? Or information about which chipsets are used in a certain model, so you can check hardware compatibility?
Sun used to be pretty good (borderline excellent) in this area, but Oracle is killing that off in a hurry. Oracle's website and visible level of support is far worse than Sun's in any case. Oracle apparently does not want to be a small and medium business market provider of anything.
What's going to happen is a merger between Dell and HP
I suspect such a merger might have a difficult time getting approved due to anti-trust considerations.
I use Chrome all the time, but I always go to another browser to print anything. Internet Explorer's printing support isn't all that great (always cutting stuff off on the right instead of scaling for example), but Chrome's (at least on Windows XP) is positively pathetic. It looks like a kindergartner did the kerning. More or less unreadable. I am looking forward to a fix for that.
The telephone company doesn't want to care whether it is a legal download. Common carriers are insulated from liability for the traffic that goes across the network. Even common carriers can refuse to provide service for blatantly illegal activity, and on occasion may be required by the government to do so.
Under normal conditions they just have no incentive to do so, in part because the burden of proof is on them instead of the other way around. Nobody who is largely insulated from legal liability wants to go around provoking lawsuits. To say nothing of the fact that electronic communications providers are generally prohibited from analyzing traffic content for the purpose of determining whether it is legal or not.
Comcast won in the end in case you forgot
The only reason why Comcast won is that the FCC was trying to regulate the Internet under a section of the law that gave them absolutely no power to do so. The FCC is in a position to fix that problem, by regulating Internet access providers under Title II.
Avoiding Title II was criminally irresponsible on the part of the FCC in the first place, they got smacked down for it and now we all suffer the consequences of the FCC's previous decision to operate under a purely fictional legal regime. Wishful thinking doesn't usually carry you very far in the legal system. The FCC has the proper tools, now they just have to use them.
If you actually want to test the limits of the bus, a test with SSDs would be far more realistic. Standards like this are designed to support what people will be doing a decade from now, not so much what they are doing today.
There are other more important improvements in USB 3.0 than raw bandwidth, mostly the transition from a polling based to a DMA/interrupt driven interface. This will help things immensely, even when only a fraction of the rated bandwidth is being used, due to less CPU overhead and power usage mostly. USB 3.0 is more like "USB grows up" than "USB 2.0 this time with feeling".
networking is starting to hit 10Gb/s.
Ethernet now has 40 Gb/s and 100 Gb/s standards as well. Rather expensive and used only for special purpose applications at this point of course. Data centers, cluster interconnect, and WAN links mostly.
So, each USB iteration offers the smallest possible increments in speed?
In some cases, the "smallest possible" increment is more than a doubling. Check this or this out for example.
Not that it justifies the use of the term out of context of course.
for example, there was a time when a foregone conclusion was one that was so unlikely you may as well not think about it
On the contrary, a foregone conclusion has _always_ been a conclusion that was so definite that it was like a conclusion that was made "before" and is "gone" such that there isn't any chance of any other.
See http://www.phrases.org.uk/meanings/141000.html
It is your responsibility if you publish video to publish in the ISO/IEC standard.
Maybe if you live in a police state. Everywhere else, perhaps ISO standards would get a little more adoption if they were royalty and patent free. In this case the ISO is just acting as a shill for the MPEG LA.
If you don't care about speed or volume by all means buy an inkjet. I have an relatively inexpensive HP CP1215 color laser printer that will print 100 high quality full color full page photos in about half an hour.
The problem with this has been pointed out a gazillion times: updating the leap second table requires either connectivity or manual intervention, which are not available in all applications where people want to use UTC
Give me a break. A local UTC clock isn't accurate without connectivity or manual intervention the moment an unrecognized leap second passes either.
Using a TAI standard for the system clock has no real downside and lots of upsides. If you don't have an up-to-date leap second table, your idea of TAI might be off, but it won't make any difference as long as you use purpose appropriate timestamp storage formats. The "wall clock" conversion will always read what you expect it to be within the limitations of any astronomically derived time system.
And unlike a typical POSIX time_t implementation, the system will not require all sorts of crazy hacks to allow it to survive (or function correctly during) a leap second while simultaneously pretending it doesn't exist.
Sysadmins have to fix the problems that occur from departure bombs and dead man switches, once you've done that you develop an extreme dislike of them.
I am personally aware of one circumstance where this happened, and all I can say is that a sysadmin who does something like this deserves three to five years in the nearest federal penitentiary, if not more.
Because color laser 'photo quality" prints look like modern inkjet prints set to "fast draft"?
My experience is more the opposite. A decent color laser printer produces photo quality prints at about eight pages per minute. A typical inkjet printer produces comparable quality prints about 100 times that slow. Anything an inkjet produces in a reasonable amount of time makes dot matrix printers look like an improvement. Ugliest prints on the planet, ink spatter and all.
It is the way it is now for historical reasons. The debate is about how to change it. The drift is relatively slow, and it would be inconvenient to wait until the clock was off by an entire day (noon at "midnight" would be half way through the cycle).
However, there are serious advocates of the idea that we should drop leap seconds from civil time and then have a "leap hour" to adjust every thousand years or so. Then local noon would only be off at most one hour from what it is already off now.
It seems to be more likely though that computers will be changed to track TAI, an atomic time standard that doesn't have leap seconds, and then use some sort of leap second table to convert to "wall clock" or civil time, much as UTC is converted to local time with time zone adjustments now.
Getting rid of leap seconds in the representation would be a mistake in the long run. A much superior fix would be to have computers keep track of TAI internally and then convert to UTC with a leap second table, much the same way we convert to local time with a time zone table.
Cluster software should be running off of a leap second free distributed clock, and TAI or the equivalent is the best one we have, handily provided (within a constant) on a world wide basis courtesy of the GPS system.
POSIX time_t values as it stands today isn't much of a clock at all, more like a short hand for encoding a wall time. We should have a new interface for a reference clock that doesn't need unpredictable adjustments every couple of years, _by definition_.
Finally, the PTO doesn't cost taxpayers a dime
No, it is the patents they issue that cost taxpayers far more dearly. We would probably be better off if the PTO was a smoking hole in the ground.
Not tested in federal court, no, but rather before an administrative law judge who works for the USITC. However, the company whose products are blocked can appeal any adverse decision to the Court of Appeals for the Federal Circuit, which will do a more thorough job.
Intellectual property rights are exclusive rights. That's the way the law has designed them -- it's not a matter of Oracle being evil. Those IPRs entitle a right holder to enforce exclusivity.
In the case of many kinds of patents, and software patents in general, those exclusive rights cause far more harm than good. Congress should rewrite the whole system from the ground up.
(1) Patents should be restricted to a handful of fields.
(2) Patent licensing should only be allowed to recover documented, non-overlapping R&D expenses associated with that patent plus perhaps a 100% gross margin.
(3) Patent licensing should be mandatory at government regulated rates in line with that.
(4) When the licensing recovery is reached, the patent should auto-expire
With rules like that, patents might actually promote the progress of the science and the useful arts, rather than grossly hindering it.
And a recursive DNS server patched to magically convert results from the internet from IPv6 to a special IPv4 address that the gateway can use to figure out where the IPv4 client really wanted to go?
It is much more likely that the gateway will do this for you, along with a bunch of other ugly v4v6 NAT and "transparent" proxying. It would probably have to strip DNSSEC of course. And the gateway will have to make up lots of synthetic IPv4 addresses for just the reason you describe, at least for any protocol it does not know how to rewrite.
As I understand it, Internet Explorer on XP will never support SNI, no matter what version you are running. That means SNI is impractical for most sites until the day comes the operators are willing to tell all remaining XP users to use a different browser. Three years maybe.
You're wrong on several counts, within 2-3 years your ISP will most likely switch you to IPv6.
It is _much_ more likely that most ISPs will deploy carrier grade NAT long before they start shipping new routers to every single one of their customers. That means two layers of NAT, one at the customer premises (as we have now) and one at the ISP.
Only people who get new routers or flash their current one (assuming it is new enough) are likely to be able to use IPv6 without a tunnel.
With IPv6, all new owners can talk to the old owners.
Strictly speaking this is not true. IPv6 is like a new parallel overlay network. No bidirectional communication occurs between IPv4 only hosts and IPv6 only hosts except through NAT (or application level proxies). Anyone who thinks NAT is going anyway any time soon is deluded.