One of the benefits of ZFS is that it does disc scrubbing - i.e. periodically reading the data on the disks to verify consistency, along with marking of failing sectors - and combined with options on how many parity bits for RAIDZ.
Who knows, maybe Sun can sell a bunch of 'Thumper' boxes to Hollyweird for preservation of the digital masters.
Lacie has been around in the Mac world for a long time, which may explain the nice Firewire performance. It may also explain why it "wasn't formatted"... I'm speculating, but it is conceivable that it was Mac-formatted.
The docs that came with the LaCie BigDisk bought by a co-worker specifically stated that it was formatted for the Mac as Mac's were the most common computers with a FW800 interface. My co-worker had no trouble re-formatting the drive for ext3 using the FW400 interface (fastest external interface on a two or so year old e-machines AMD64 laptop - we tested the interface using my LaCie). The drive has not given any troubles despite being subject to bouncing while running, powerdowns while running and other not so gentle treatment.
While FAT32 does have the advantage that just about anything can read/write to it, the 4GB file size limitation (which dates back to Seattle Computer Products' QDOS original FAT implementation) really limits its usefulness for large drives. SOP for me is to re-format FAT drives with the most appropriate file system: UFS for Solaris (fun part was figuring out that I needed to use svcadm to kill vold) or NTFS for Windows XP.
I've been involved in the design and construction of several magnets for NMR use - and the supporting structure is usually considered to be part pf the magnet - including the cryostat used in supercons.
The interesting part of the article was that the cryostat design was reviewed by CERN personnel, so the issue of asymmetric loading on the cryostat was overlooked by more than just Fermilab. Sounds like and "Oh shit - nobody thunk of that" moment.
We don't have mere "correlation" between rising CO2 and rising temperature, there is a known causal link, via the greenhouse effect, which is not difficult to calculate from basic physics.
As someone else pointed out, are we sure that CO2 is the only forcing function?
Unfortunately the models snip
(f) essentially assume that CO2 is the *only* climate forcing agent That's my biggest beef with the global climate change models. One of the results of the three day 'no-fly' over the US in Sep 2001 was that nights were about 1C cooler due to the lack of contrails - clear skies have a lower sky temperature than cloudy skies.
What I'd like to see is some consistent estimates of how much warming is caused by CO2, how much by methane, how much by CFC's, how much by contrails and for good measure, how much by reducing particulates.
The radio frequency energy absorbed by all those water molecules has to go somewhere....and the conversion of RF energy into molecular vibration (heat) is a good candidate for the cause of the excessive path loss at 2.4 GHz compared to path losses at 2.3 GHz or 2.5GHz.
Only problem is that the water resonance is something like 20 GHz. I've an industrial scale (1 MW) microwave oven running in the 900 MHz ISM band and I read about someone cooking hamburgers with a cavity tuned to 144 MHz.
If you read TFA, BlueJ is being developed with the support of Sun Microsystems - 'course Sun's recent deal with MS may cool their ardor for defending their IP rights - OTOH, Sun could dredge up the settlement for their Java suit.
And it differs from 8-way multiprocessor systems how?
Memory access. The cache and four memory banks are tied to the 8 cores on the Niagara via a crossbar switch, so the cache and memory are local. On an 8-way Opteron system, non-local memory access can involve 1 to 3 hops over the HT links and good luck getting adequate memory perfomance on an 8-way Xeon system.
Another difference is power consumption - a Niagara system consumes about the same (or less) power than a single core Opteron. Due to the lower clock speed of the Niagara, the 8 cores on a Niagara will not be 8X the throughput of a single Opteron. And as someone else pointed out, the FP performance of the Niagara 1 sucks - but the Niagara 2 is reported to be about 10X better.
Not sure if you're being sarcastic or ignorant. One of Sun's latest jewels is the 8 core Niagara and it behooves them to come up with ways of keeping all 8 cores going on a processor.
I was wondering how weak of a grip you have on reality, the US pipe size standards go back for decades and it isn't likely that they will change soon (threads for pipe are in threads per inch). While it is possible that new construction could use metric plumbing sizes, the lack of selection for fixtures and problems with getting replacemnt parts would deter most buyers. Other standard dimensions are 4 by 8 feet for plywood and drywall sheets, screws are typically SAE sizes (an example is the 19 inch relay rack with 10-32 screws).
For similar reasons, I don't expect the US to move away from the "letter" size paer sheet (8.5 by 11 inches or 215.9 by 279.4mm).
Then there's the really odd standard RR gauge for most of the world - 4'8.5" or 1435mm - it would be really beneficial to go to a wider gauge, but the cost of conversion would far outweigh the benefits. Only thing odder than that is the definition of HO scale - 3.5mm to the foot.
No, the people who build the atomic bombs were giants, the people who decided to use them were midgets (actually, they were not, but that it another discussion).
The decision to use nuclear weapons on Japan was ultimatelly Harry Truman's and it was the correct choice. The real atrocity committed by an Allied politician of that era was Chamberlain's appeasement of Hitler which led to the European theater of WW2 - the Asian theater started in 1931 when Japan invaded China.
FWIW, Japan came very close to provoking a full scale chemical and biological war with the US - fortunately for Japan, the Japanese sunmarine carrying the jars of bubonic plague infested fleas was sunk en route to the Mariannas.
Virtually all of the titles to real estate in the US and Canada are made in imperial/customary units and would be an enormous umdertaking to convert to metric. For a more complete explanantion of the problem, see Andro Linklater's book, Measuring America, which gives some very good insights on why the US did not go metric - even though the metric system was Thomas Jefferson's idea (and Jefferson had a much better idea for the unit of length than what the French Academy of Science came up with).
There are two things about the metric system that I find are highly overrated.
#1 is the Celsius scale for temperature - 100C is the temperature at which the vapor pressure of water is 101.6kp - would make more sense to express temp in K or better yet, electron volts.
#2 is fixing the gram to be one cc of water at God knows what temperature and pressure - for almost all cases, I'm going to look up the density in a table and it really isn't that much different in dealing with lbm/cu.ft versus tonnes/m^3.
The comment about pipe sizes is a good one - transferring to metric will be a very long process.
Good point wrt Cairo - although I would go a bit further to say that Cairo was the plan to kill off OS/2.
As for GUI's, I'm much less frustrated using the Unix based GUI's (including CDE) than Windows GUI's. As an example, I was backing up a directory on my wife's old laptop with an early version of XP installed - with the back-up functionality not present in that version of XP, I figured that most of that could be accomplished by copying the directories to the external drive. After a few minutes it came across a file that was in use by an application - rather than reporting it couldn't copy that particular file (and it gave just the file name without the path) and move on, the copy function just simply quit.
Anyway, you've given me another reason to avoid Vista...
Wasn't that originally promised in Cairo? Come on, this is getting to be even worse than Duke Nukem Forever. What's even worse is the "abscence of drive letters has been a feature of Unix since the early 1970's.
As for "file locations" being managed by applications - mmm, no thanks, I rather group files by projects which can involve many applications. What I'd really love is a return to the functionality present in Word for DOS, where the application would look in the current working directory for project specific configuration files.
One of the problems with maintaining backwards compatiblity in Windoze is that the whole mess evolved by acreting guano rather than having a clear path for upgrades. That is define where the various modules will interface with each other before you start coding.
Am I missing something here? Why doesn't python count?
One not so trivial reason is that the examples in textbooks are in BASIC not Python.
OTOH, I did have a flashback while entering in the HTML code for my reply of using the early "dot" commands for text formatters. Probably no-one under the age of 40 has even been exposed to those, but they were good for getting an insight on mark-up.
Good to see that you weren't as humor impaired as the moderator who mod'ed me "Troll".
The latest news listed 8 of the 9 newshounds that had their phone records "pretexted", none of which worked for either the Inquirer or Rogister - oh well...
Did have a wicked thought about dear departed Carly - picturing her getting autographed the same way you did by "Hedgehog". "nother wicked thought was getting Tadpole to come up with a dual PA-RISC or Itanic version of the Bullfrog -um- laptop for knocking a few heads at HP silly.
The way cdrecord accesses CD-RW drives in GNU/Linux is different from all other applications. Another one of those "I'll do it my way" ideas by Jörg Schilling.
The 'scg' library (which cdrecord uses to access the CD drives) predates Linux by several years and Linux is only one of the several OS's supported by cdrecord. FWIW, Linux has its share of "I'll do it my way" ideas by Linus and the other developers.
Well put - it is his code and it is his decision to choose what license he wants to use.
I've read a lot of criticism about the way device names are handled in Joerg's 'scg' library, but Joerg does have a valid point in that the naming is consistent across platforms (and predates Linux by several years) - which is more than can be said about device naming/enumeration under Linux.
As for "Free Software" - the BSD folks have a point in that the various flavors of the GPL are more restrictive than the BSD license (which was adapted from the license terms for SPICE chosen by D.O. Pederson).
Hey, speaking of Sun and different lines of chips, what about the clockless chip I used to hear about? Is that still in development, or has it been canned?
The US-IIIi is supposed to have some asynchronous (i.e. clockless) logic in it.
Slashdot Mirror
Who knows, maybe Sun can sell a bunch of 'Thumper' boxes to Hollyweird for preservation of the digital masters.
The docs that came with the LaCie BigDisk bought by a co-worker specifically stated that it was formatted for the Mac as Mac's were the most common computers with a FW800 interface. My co-worker had no trouble re-formatting the drive for ext3 using the FW400 interface (fastest external interface on a two or so year old e-machines AMD64 laptop - we tested the interface using my LaCie). The drive has not given any troubles despite being subject to bouncing while running, powerdowns while running and other not so gentle treatment.While FAT32 does have the advantage that just about anything can read/write to it, the 4GB file size limitation (which dates back to Seattle Computer Products' QDOS original FAT implementation) really limits its usefulness for large drives. SOP for me is to re-format FAT drives with the most appropriate file system: UFS for Solaris (fun part was figuring out that I needed to use svcadm to kill vold) or NTFS for Windows XP.
I wasn't impressed by the review either.
The interesting part of the article was that the cryostat design was reviewed by CERN personnel, so the issue of asymmetric loading on the cryostat was overlooked by more than just Fermilab. Sounds like and "Oh shit - nobody thunk of that" moment.
As someone else pointed out, are we sure that CO2 is the only forcing function?snip
(f) essentially assume that CO2 is the *only* climate forcing agent
That's my biggest beef with the global climate change models. One of the results of the three day 'no-fly' over the US in Sep 2001 was that nights were about 1C cooler due to the lack of contrails - clear skies have a lower sky temperature than cloudy skies.
What I'd like to see is some consistent estimates of how much warming is caused by CO2, how much by methane, how much by CFC's, how much by contrails and for good measure, how much by reducing particulates.
Only problem is that the water resonance is something like 20 GHz. I've an industrial scale (1 MW) microwave oven running in the 900 MHz ISM band and I read about someone cooking hamburgers with a cavity tuned to 144 MHz.
If you read TFA, BlueJ is being developed with the support of Sun Microsystems - 'course Sun's recent deal with MS may cool their ardor for defending their IP rights - OTOH, Sun could dredge up the settlement for their Java suit.
Memory access. The cache and four memory banks are tied to the 8 cores on the Niagara via a crossbar switch, so the cache and memory are local. On an 8-way Opteron system, non-local memory access can involve 1 to 3 hops over the HT links and good luck getting adequate memory perfomance on an 8-way Xeon system.
Another difference is power consumption - a Niagara system consumes about the same (or less) power than a single core Opteron. Due to the lower clock speed of the Niagara, the 8 cores on a Niagara will not be 8X the throughput of a single Opteron. And as someone else pointed out, the FP performance of the Niagara 1 sucks - but the Niagara 2 is reported to be about 10X better.
Not sure if you're being sarcastic or ignorant. One of Sun's latest jewels is the 8 core Niagara and it behooves them to come up with ways of keeping all 8 cores going on a processor.
For similar reasons, I don't expect the US to move away from the "letter" size paer sheet (8.5 by 11 inches or 215.9 by 279.4mm).
Then there's the really odd standard RR gauge for most of the world - 4'8.5" or 1435mm - it would be really beneficial to go to a wider gauge, but the cost of conversion would far outweigh the benefits. Only thing odder than that is the definition of HO scale - 3.5mm to the foot.
The decision to use nuclear weapons on Japan was ultimatelly Harry Truman's and it was the correct choice. The real atrocity committed by an Allied politician of that era was Chamberlain's appeasement of Hitler which led to the European theater of WW2 - the Asian theater started in 1931 when Japan invaded China.
FWIW, Japan came very close to provoking a full scale chemical and biological war with the US - fortunately for Japan, the Japanese sunmarine carrying the jars of bubonic plague infested fleas was sunk en route to the Mariannas.
There are two things about the metric system that I find are highly overrated.
#1 is the Celsius scale for temperature - 100C is the temperature at which the vapor pressure of water is 101.6kp - would make more sense to express temp in K or better yet, electron volts.
#2 is fixing the gram to be one cc of water at God knows what temperature and pressure - for almost all cases, I'm going to look up the density in a table and it really isn't that much different in dealing with lbm/cu.ft versus tonnes/m^3.
The comment about pipe sizes is a good one - transferring to metric will be a very long process.
This is not a specification; this is a DNA sequence.Outrageously funny and to the point.
As for GUI's, I'm much less frustrated using the Unix based GUI's (including CDE) than Windows GUI's. As an example, I was backing up a directory on my wife's old laptop with an early version of XP installed - with the back-up functionality not present in that version of XP, I figured that most of that could be accomplished by copying the directories to the external drive. After a few minutes it came across a file that was in use by an application - rather than reporting it couldn't copy that particular file (and it gave just the file name without the path) and move on, the copy function just simply quit.
Anyway, you've given me another reason to avoid Vista...
As for "file locations" being managed by applications - mmm, no thanks, I rather group files by projects which can involve many applications. What I'd really love is a return to the functionality present in Word for DOS, where the application would look in the current working directory for project specific configuration files.
For OOo, Solaris is more of a primary platform than Mac OSX...
One of the problems with maintaining backwards compatiblity in Windoze is that the whole mess evolved by acreting guano rather than having a clear path for upgrades. That is define where the various modules will interface with each other before you start coding.
Israel getting tense, wants one in self-defense
In this case, Arco should be the responsible party for repaying the money lost - should give thme incentive to improve security.
One not so trivial reason is that the examples in textbooks are in BASIC not Python.
OTOH, I did have a flashback while entering in the HTML code for my reply of using the early "dot" commands for text formatters. Probably no-one under the age of 40 has even been exposed to those, but they were good for getting an insight on mark-up.
The latest news listed 8 of the 9 newshounds that had their phone records "pretexted", none of which worked for either the Inquirer or Rogister - oh well...
Did have a wicked thought about dear departed Carly - picturing her getting autographed the same way you did by "Hedgehog". "nother wicked thought was getting Tadpole to come up with a dual PA-RISC or Itanic version of the Bullfrog -um- laptop for knocking a few heads at HP silly.
Question is: Will you feel worse if HP didn't make an attempt to get at your phone/email records?
The 'scg' library (which cdrecord uses to access the CD drives) predates Linux by several years and Linux is only one of the several OS's supported by cdrecord. FWIW, Linux has its share of "I'll do it my way" ideas by Linus and the other developers.
I've read a lot of criticism about the way device names are handled in Joerg's 'scg' library, but Joerg does have a valid point in that the naming is consistent across platforms (and predates Linux by several years) - which is more than can be said about device naming/enumeration under Linux.
As for "Free Software" - the BSD folks have a point in that the various flavors of the GPL are more restrictive than the BSD license (which was adapted from the license terms for SPICE chosen by D.O. Pederson).
The US-IIIi is supposed to have some asynchronous (i.e. clockless) logic in it.