We have about 15 of the DP 2Ghz G5 systems for our viz lab and classes. Under normal operation they are very quiet (even quieter than my FP iMac sometimes:-) Once they are placed under load, or the side panel is removed, the fans do ramp up but I have never heard them going at full blast. In any case, Apple's idea was to radically over-fan the G5 so that each fan would normally never have to spin very fast or get very loud. It's a great design and, yes, I'll bet you see the PC case manufacturers do something similar in the future (although since PC stuff is so price sensitive and has to deal with such a variety of 'innards', they may not be nearly as elaborate).
The posted article, while interesting, is wrong about the volt being based on the Kilogram. Since about 1990, the volt is defined to be the voltage applied to a Josephson junction that produces a frequency of 483,597.9 GHz. This new standard was implemented in order to get away from relying on 'artifact' standards (such as the Kg cylinder). One quick source page on Josephson junctions (which completely revolutionized the field of Metrology back when I was a calibration tech in the AF) is:
http://hyperphysics.phy-astr.gsu.edu/hbase/solids/ squid.html
If I recall correctly, the eventual goal of the international standards organization was to find ways to define everything in terms of frequency/time since we can measure time so accurately/precisely.
I find it especially ironic how so many slashdotters are talking about how silly it is for Apple to put lights and things in the case when I see all of these case mod posts and topics. If LEDs and lamps aren't cool than why are there rows of illuminated cases, fans, power supplies, cables, etc at Fry's, thinkgeek, etc?
In any case (no pun intended;-), I think that this has more to do with an adaptive tablet than with a system case. For example, picture a Wacom tablet that uses inkwell and that has an interface that changes based on the application you are using. Since it would not be an LCD panel, it would be much more rugged.
Ah, but the difference is that Unix boxes don't change the ownership of files outside of the user's $HOME when that user logs in. OSX does (all of the files/dirs that are set to 'unknown' group in / are set to be owned by that user when the user logs in). Again, this is to support the legacy OS9 environment. This is more than just the preferences since the question remains: who should own, say, the OS9 'System Folder'? The console GUI user? or the remote GUI user? or just root/admin?
One thing that needs to be kept in mind in all of this is that much more than the GUI is involved. In case you had not noticed, OS X changes the ownership of several files and directories to match whomever is the current console user. For example, the OS 9 Applications directory and the/Documents folder. THese changes are needed to support the classic environment. Thus, if you have two folks logged in via a GUI who should get ownership of these files/directories? One solution would be to limit classic to just the 'console' user, but that would probably mean changing large portions of Aqua/Quartz.
In addition,Aqua would have to be changed to understand how to lock USB and FireWire resources so that remote and console users don't crash into each other.
In my opinion, however, the biggest issue that would kill having an X-style remote GUI is philosophical: nearly all of the software vendors and general Mac users assume that the console user is the only user and having the ability to have multiple GUI logins would really confuse the heck out of them.
The first thing that pooped into my mind when reading the article was that this could put a whole new spin on games like Myth. I mean if you could use a somewhat simplified version of the control nodes and behaviors to generate an army-- and your opponent (or the computer) did the same-- you could create some awesome battles. Even in the case of individual battles, the idea of better ai for each character is pretty compelling;-) Oh, and for more fun, throw in some genetic/adaptive algorithms and watch as your characters get better and better...
Just to let you know, I got a copy of the OS X for Unix Geeks book and it is great. IT even has man pages for some of the undocumented Apple CL tools! Definitely worth getting if you want details about OS X.
(PS: I'm writing this while in the 'Mac Hacking' session of the O'Reilly OS X conference... This has been a really good conference and I'm quite glad I came to it;-)
Actually, I think that the real real answer is to just fully publish the file formats and specifications and let us create our own toolkits. I mean, c'mon, there are many creative folks in both the OSX and Linux camps who would love to tackle writing a toolset for editing NWN modules. Once we know the module format, this becomes much easier. Im fact, I sucpect that reverse-engineered toolsets (for linux at least) will start popping up soon after the game is released.
Hey! Thanks! I actually have a copy of GNO (that I actually purchased along with lots of the Orca stuff) and used to run it on my gs. It was pretty cool having a Unixy system for it. I had never even thought to look for the web site (I recognize Devin Reade's name from the old comp.sys.apple2.gno usenet group;-)
Hmmm. I have an Apple//gs sitting in my basement that has a 4Mb memory card, 40Mb HD and a 7Mhz accelerator card (and the old Orca C compilers). This might be an interesting project for it...
Here's a thought: The original revision A iMacs had most of the 'guts' (CPU, ROMs, etc) on a daughtercard that was connected to the motherboard through a PCI-type interface. What if Mr. Drew has just bought up a bunch of old revA iMacs (or just the obsolete daughtercards) and has created some glue hardware that connects them to the PCI bus? The emulation code would need to make the PC hardware and memory look like Mac stuff to OSX. However, as you would be running on real Apple hardware, there should be no problem getting OSX to install/run.
Hmmmmm...
Most video rental places also rent console games and already have limited space for the current systems (PS/PS2/N64/SNES). How willing are they going to be to dedicate _more_ space for these new systems? I've read that game publishers have to pay some chains for shelf space. Perhaps the same may soon hold true for rental shops...
Most of the commercial software companies maintain departments in which basic research into things such as GUI implementations (ala Apple's Human Interface Group) and usability (ala Microsoft's various usability groups) and other areas (ala IBM sponsorship of University projects). It is the work of these groups that has defined much of how GUIs and applications work today.
What sort of research is Red Hat involved in? That is, what kinds of projects are Red Hat's R&D folks involved in? Along those lines, do you see (or hope to see in the future) Red Hat as having a leading role in one of these areas?
Thanks!
Re:What do you think of bastille?
on
Ask Robert Young
·
· Score: 1
Sorry, but I disagree strongly with you that cray is just 'rehashing old technologies'. As a primarily Cray shop (we have a T3E-900 and, after this weekend's upgrade, have the first SV1e) we are quite involved with the Cray User's Group (CUG) meetings and such where Cray's plans for the future (SV2 and beyond) are discussed.
There are _many_ things going on behind the scenes at Cray that show that Cray is once again trying to push the supercomputing envelope as far as they can. One way to look at the SV2 is as a T3E with large vector units in each CPU (no e-registers) and a nearly flat (shared) memory space across all processors. Thus, no need for mixed mode (MPI and OpenMP) programming like on IBM SP-like architectures.
Rather than look at this from the one-time activation viewpoint, look at it from the perspective of what Microsoft wants to do with.Net... That is, since they are going to the subscription model for Office, Windows, etc. then it makes sense why they are creating a unique 'fingerprint' of your system. What better way to track the software you are renting from them?
So, I suggest that Microsoft does not care whether or not the XP scheme is cracked. They just want to get unique fingerprints for all these systems on file for when.Net is rolled out.
And then there is motion sickness ...
on
3D GUI Project
·
· Score: 1
One of the toys that we have around here is a big honkin' stereo display (Pyramid systems ImmersaDesk) which we use during tours of the center (along with researchers who use it for visualizations). One thing that I've noticed is that many folks are quite suceptible to getting motion sickness after using the display. The problem seems to be that even though the scene may seem to be 3D, your inner-ear is telling your brain that it is not.
Just think about how many folks you know that get nauseous after playing Doom/Quake/whatever and imagine how they would handle a 3D display...
In sum, I think that it is an interesting project. However, for most folks, the disorientation caused by the stereo images would outweigh any perceived navigational benefits.
Bah! Living in Fairbanks, Alaska we get to see the aurora quite often (right now we have less then 11 hours of daylight and are losing about 7 minutes per day). There's been times I've seen it driving home from work. Of course, they look the best when it is about -40 or so and the light drowns out the stars... way cool;-)
For a forcast, see the UAF Geophysical Institute's Aurora Forcast page.
We used to have a Cray Y-MP where I work... One little detail that has not yet been mentioned is that a Y-MP is water cooled. So, anyone without their own heat exchange unit and/or chilled water suppy need not apply;-) Also, the thing used about $20,000 of electricity every month. Yipes!
Hmmm. It may turn out that insisting on going to the District court is exactly the wrong thing for Microsoft. That is, the District Court judges are not stupid. They, I'm sure, have been following this case and are aware of:
How they are perceived by the media (I.e. willing to roll over for Microsoft)
How Microsoft wormed out of the previous rulings.
Thus, they may see this as a means of telling Microsoft "Look, you guys made us look silly. We will make sure that there are no misunderstandings this time." I have a feeling that they are taking this case en banc just to prove that their decision will be unambiguous and apolitical. Kinda puts a different spin on things, eh?;-)
Actually, the SNIA system is a little bit of both. A cluster 'node' will be a shared memory system with 2 to (IIRC) 32 IA64 CPUs. These nodes can then be hooked together in distributed memory clusters. IBM, Compaq and Sun are using similar techniques with their upcoming systems (we are getting another system to go with our T3E and J90, so we all got to sit though several days of NDA vendor talks. Was pretty interesting). They all want to get the benefits of large numbers of processors without the hassles of all those CPUs talking to the same memory pool.
Nice interview. It looks like clustering technology is, indeed, advancing by leaps and bounds (In fact, if you substituted IA64 processors for the Alphas, you'd have just about described SGI's SNIA architecture;-)... Anyway, Greg mentioned that when a node goes down, the job running on that node is restarted. What about for scheduled maintenance? Does the system software have any checkpoint/restart features? That is, the ability to write out a jobs state to disk and then restart it later (on the same or other nodes). On our 272-node T3E checkpoint/restart is really vital as we have many users who run 128node (or larger) jobs that take from 8 to 16 hours. They'd scream bloody murder if we had to restart their jobs from the beginning all the time.
One other detail that should be mentioned is that Sauron was familiar with the Elves, Dwarves, Men, etc. But did not have any real knowledge of Hobbits. Hobbits came on to the scene _after_ the Rings had been created and given and were, thus, largly unaffected by their power. The best example is, of course, Gollum. He, unlike the men, possessed the Ring for many years, yet never fully turned to a wraith. Yes, he was corrupted, but he was still less affected than the humans were. I often think that the Hobbits were created specifically by one of the Valar (or, possibly, by Illuvetor Himself) in order to counter the Ring's power (similarly to the way in which the dwarves were created). Their origins were always shrouded in a mystery that even Gandalf could not penetrate.
Slashdot Mirror
We have about 15 of the DP 2Ghz G5 systems for our viz lab and classes. Under normal operation they are very quiet (even quieter than my FP iMac sometimes :-) Once they are placed under load, or the side panel is removed, the fans do ramp up but I have never heard them going at full blast. In any case, Apple's idea was to radically over-fan the G5 so that each fan would normally never have to spin very fast or get very loud. It's a great design and, yes, I'll bet you see the PC case manufacturers do something similar in the future (although since PC stuff is so price sensitive and has to deal with such a variety of 'innards', they may not be nearly as elaborate).
Uh, Wait just a minute... How can the Beatles be suing Apple if they are giving away an iPod with each Beatle sold? ;-)
The posted article, while interesting, is wrong about the volt being based on the Kilogram. Since about 1990, the volt is defined to be the voltage applied to a Josephson junction that produces a frequency of 483,597.9 GHz. This new standard was implemented in order to get away from relying on 'artifact' standards (such as the Kg cylinder). One quick source page on Josephson junctions (which completely revolutionized the field of Metrology back when I was a calibration tech in the AF) is:/ squid.html
http://hyperphysics.phy-astr.gsu.edu/hbase/solids
If I recall correctly, the eventual goal of the international standards organization was to find ways to define everything in terms of frequency/time since we can measure time so accurately/precisely.
I find it especially ironic how so many slashdotters are talking about how silly it is for Apple to put lights and things in the case when I see all of these case mod posts and topics. If LEDs and lamps aren't cool than why are there rows of illuminated cases, fans, power supplies, cables, etc at Fry's, thinkgeek, etc? ;-), I think that this has more to do with an adaptive tablet than with a system case. For example, picture a Wacom tablet that uses inkwell and that has an interface that changes based on the application you are using. Since it would not be an LCD panel, it would be much more rugged.
In any case (no pun intended
Ah, but the difference is that Unix boxes don't change the ownership of files outside of the user's $HOME when that user logs in. OSX does (all of the files/dirs that are set to 'unknown' group in / are set to be owned by that user when the user logs in). Again, this is to support the legacy OS9 environment. This is more than just the preferences since the question remains: who should own, say, the OS9 'System Folder'? The console GUI user? or the remote GUI user? or just root/admin?
One thing that needs to be kept in mind in all of this is that much more than the GUI is involved. In case you had not noticed, OS X changes the ownership of several files and directories to match whomever is the current console user. For example, the OS 9 Applications directory and the /Documents folder. THese changes are needed to support the classic environment. Thus, if you have two folks logged in via a GUI who should get ownership of these files/directories? One solution would be to limit classic to just the 'console' user, but that would probably mean changing large portions of Aqua/Quartz.
In addition,Aqua would have to be changed to understand how to lock USB and FireWire resources so that remote and console users don't crash into each other.
In my opinion, however, the biggest issue that would kill having an X-style remote GUI is philosophical: nearly all of the software vendors and general Mac users assume that the console user is the only user and having the ability to have multiple GUI logins would really confuse the heck out of them.
The first thing that pooped into my mind when reading the article was that this could put a whole new spin on games like Myth. I mean if you could use a somewhat simplified version of the control nodes and behaviors to generate an army-- and your opponent (or the computer) did the same-- you could create some awesome battles. Even in the case of individual battles, the idea of better ai for each character is pretty compelling ;-) Oh, and for more fun, throw in some genetic/adaptive algorithms and watch as your characters get better and better...
Just to let you know, I got a copy of the OS X for Unix Geeks book and it is great. IT even has man pages for some of the undocumented Apple CL tools! Definitely worth getting if you want details about OS X.
;-)
(PS: I'm writing this while in the 'Mac Hacking' session of the O'Reilly OS X conference... This has been a really good conference and I'm quite glad I came to it
Actually, I think that the real real answer is to just fully publish the file formats and specifications and let us create our own toolkits. I mean, c'mon, there are many creative folks in both the OSX and Linux camps who would love to tackle writing a toolset for editing NWN modules. Once we know the module format, this becomes much easier. Im fact, I sucpect that reverse-engineered toolsets (for linux at least) will start popping up soon after the game is released.
Hey! Thanks! I actually have a copy of GNO (that I actually purchased along with lots of the Orca stuff) and used to run it on my gs. It was pretty cool having a Unixy system for it. I had never even thought to look for the web site (I recognize Devin Reade's name from the old comp.sys.apple2.gno usenet group ;-)
Hmmm. I have an Apple //gs sitting in my basement that has a 4Mb memory card, 40Mb HD and a 7Mhz accelerator card (and the old Orca C compilers). This might be an interesting project for it...
Here's a thought: The original revision A iMacs had most of the 'guts' (CPU, ROMs, etc) on a daughtercard that was connected to the motherboard through a PCI-type interface. What if Mr. Drew has just bought up a bunch of old revA iMacs (or just the obsolete daughtercards) and has created some glue hardware that connects them to the PCI bus? The emulation code would need to make the PC hardware and memory look like Mac stuff to OSX. However, as you would be running on real Apple hardware, there should be no problem getting OSX to install/run.
Hmmmmm...
Most video rental places also rent console games and already have limited space for the current systems (PS/PS2/N64/SNES). How willing are they going to be to dedicate _more_ space for these new systems? I've read that game publishers have to pay some chains for shelf space. Perhaps the same may soon hold true for rental shops...
Most of the commercial software companies maintain departments in which basic research into things such as GUI implementations (ala Apple's Human Interface Group) and usability (ala Microsoft's various usability groups) and other areas (ala IBM sponsorship of University projects). It is the work of these groups that has defined much of how GUIs and applications work today.
What sort of research is Red Hat involved in? That is, what kinds of projects are Red Hat's R&D folks involved in? Along those lines, do you see (or hope to see in the future) Red Hat as having a leading role in one of these areas?
Thanks!
Personally, I think he's a pretty cool guy ;-)
Sorry, but I disagree strongly with you that cray is just 'rehashing old technologies'. As a primarily Cray shop (we have a T3E-900 and, after this weekend's upgrade, have the first SV1e) we are quite involved with the Cray User's Group (CUG) meetings and such where Cray's plans for the future (SV2 and beyond) are discussed.
There are _many_ things going on behind the scenes at Cray that show that Cray is once again trying to push the supercomputing envelope as far as they can. One way to look at the SV2 is as a T3E with large vector units in each CPU (no e-registers) and a nearly flat (shared) memory space across all processors. Thus, no need for mixed mode (MPI and OpenMP) programming like on IBM SP-like architectures.
Rather than look at this from the one-time activation viewpoint, look at it from the perspective of what Microsoft wants to do with .Net... That is, since they are going to the subscription model for Office, Windows, etc. then it makes sense why they are creating a unique 'fingerprint' of your system. What better way to track the software you are renting from them?
.Net is rolled out.
So, I suggest that Microsoft does not care whether or not the XP scheme is cracked. They just want to get unique fingerprints for all these systems on file for when
One of the toys that we have around here is a big honkin' stereo display (Pyramid systems ImmersaDesk) which we use during tours of the center (along with researchers who use it for visualizations). One thing that I've noticed is that many folks are quite suceptible to getting motion sickness after using the display. The problem seems to be that even though the scene may seem to be 3D, your inner-ear is telling your brain that it is not.
Just think about how many folks you know that get nauseous after playing Doom/Quake/whatever and imagine how they would handle a 3D display...
In sum, I think that it is an interesting project. However, for most folks, the disorientation caused by the stereo images would outweigh any perceived navigational benefits.
What I'd like to know is: when will the royalties for the use of my name start rolling in ? ;-)
Regards,
Derek Bastille
Bah! Living in Fairbanks, Alaska we get to see the aurora quite often (right now we have less then 11 hours of daylight and are losing about 7 minutes per day). There's been times I've seen it driving home from work. Of course, they look the best when it is about -40 or so and the light drowns out the stars... way cool ;-)
For a forcast, see the UAF Geophysical Institute's Aurora Forcast page.
We used to have a Cray Y-MP where I work... One little detail that has not yet been mentioned is that a Y-MP is water cooled. So, anyone without their own heat exchange unit and/or chilled water suppy need not apply ;-) Also, the thing used about $20,000 of electricity every month. Yipes!
Thus, they may see this as a means of telling Microsoft "Look, you guys made us look silly. We will make sure that there are no misunderstandings this time." I have a feeling that they are taking this case en banc just to prove that their decision will be unambiguous and apolitical. Kinda puts a different spin on things, eh?
Actually, the SNIA system is a little bit of both. A cluster 'node' will be a shared memory system with 2 to (IIRC) 32 IA64 CPUs. These nodes can then be hooked together in distributed memory clusters. IBM, Compaq and Sun are using similar techniques with their upcoming systems (we are getting another system to go with our T3E and J90, so we all got to sit though several days of NDA vendor talks. Was pretty interesting). They all want to get the benefits of large numbers of processors without the hassles of all those CPUs talking to the same memory pool.
Nice interview. It looks like clustering technology is, indeed, advancing by leaps and bounds (In fact, if you substituted IA64 processors for the Alphas, you'd have just about described SGI's SNIA architecture ;-) ...
Anyway, Greg mentioned that when a node goes down, the job running on that node is restarted. What about for scheduled maintenance? Does the system software have any checkpoint/restart features? That is, the ability to write out a jobs state to disk and then restart it later (on the same or other nodes).
On our 272-node T3E checkpoint/restart is really vital as we have many users who run 128node (or larger) jobs that take from 8 to 16 hours. They'd scream bloody murder if we had to restart their jobs from the beginning all the time.
Regards,
Derek
One other detail that should be mentioned is that Sauron was familiar with the Elves, Dwarves, Men, etc. But did not have any real knowledge of Hobbits. Hobbits came on to the scene _after_ the Rings had been created and given and were, thus, largly unaffected by their power. The best example is, of course, Gollum. He, unlike the men, possessed the Ring for many years, yet never fully turned to a wraith. Yes, he was corrupted, but he was still less affected than the humans were.
I often think that the Hobbits were created specifically by one of the Valar (or, possibly, by Illuvetor Himself) in order to counter the Ring's power (similarly to the way in which the dwarves were created). Their origins were always shrouded in a mystery that even Gandalf could not penetrate.