It would be difficult to get the politicians to agree on policies.
As a grid gets larger, and more complex, it becomes harder to manage. There's an increasing (though still remote) risk of a cascade failure. https://en.wikipedia.org/wiki/Northeast_blackout_of_2003
According to Wikipedia, the energy density of lithium ion batteries is 250–676 Wh/L. The older acid-quinone battery had about 50 Wh/l. The article says the new chemistry gets about 2/3 of that, around 33 Wh/l. Lead-acid batteries are around 60–110 Wh/l. So this would probably be useless for mobile applications, but good for stationary purposes. Supposedly flow batteries can last indefinately, unlike lead-acid. It sounds like that would make them good for big data center UPS batteries.
Would this be effective for large UPS batteries in data centers? The current lead calcium acid batteries we use are expensive, heavy, and need to be replaced every few years. It would be interesting if we could replace the membrane and pump in some new fluid.
1. Several X window managers had an "icon that stayed there", long before Windows. For example, twm's icon manager had it in 1988.
2. twm, fvwm, etc all had this too.
3. That was possible with many X window managers long before Windows tried it. However, most people found the visible borders useful.
4. I've seen this on both windows and X for years. I can't remember where I saw it first.
5. This page says the scroll wheel was invented by Mouse Systems, then popularized by Microsoft.
6. I believe Xfm did this in '92. I vaguely recall doing this on the Amiga in '86, but I could be wrong.
The only idea I've heard of first from Microsoft (or more accurately havn't heard of anyone else doing first) was transparent file system space saving by having the file system code combine identical files into a single copy on disk. It kept a hash of every file in the filesystem control structures, and when a new file was written, compared the hashes to see if it could merge files. Copying a file just created a second reference to the saved data. Writing to one copy caused the data to be split into seperate copies.
I'm sure Microsoft has invented other original ideas, but not nearly as many as they like to claim. Mostly they reimplement existing ideas.
Wouldn't the extra weight of the motor be partially offset by lighter breaking mechanisms, assuming that regerative breaking is the primary breaking mechanism, and friction breaks are mostly for emergency and parking?
It's technically feasable for the government to insert distributed spyware in their version of linux. A system like this would be technologically capable of monitoring keystrokes for trigger keywords, and reporting them to the government. It could use idle time to examine local files, sniff the local net, or join a distributed computing cluster cracking dissidents' crypto.
Since the government also controls the hardware, they can make it harder to replace the OS. They could make the processor only execute government signed kernels.
I don't know if the current Chinese government would do these things. But I believe they have the technical ability to do them.
Could the servers hide their real address, even from the hub, if the hub->server communications occured over a multicast group, and the server->hub communications used forged source address UDP packets?
A server would start by joining the multicast group. Next it would send an announcement to the hub, forging it's source ip, but including a randomly chosen, probably unique, identifier number. The hub would respond over multicast, using the identifier . They would procede to negotiating the server's inventory and capabilities. Eventually the hub would receive a request from a client that it could match to a server, and forward that request to the server. The server would begin transmitting the data directly to the client using forged source address UDP packets. If the multicast group becomes congested, the hub could create additional ones, and request existing servers switch to it.
I can see how someone could get other traders to digitally sign a letter of reference, then offer a set of those references when requesting files. It's similar to the web of trust concept in PGP. However, how do you keep a mole from tricking someone into signing for him, and letting him into the web? Once he's trusted, he could collect IPs, sign in other moles, etc. Also, how do you distribute negative references?
Unix used to have a steep learning curve. Learning to use it well is still a large task. However, any MS Windows user can learn to use the default Gnome or KDE desktops in a few minutes. They won't be using a tenth of what Unix can do for them, but they'll be able to click around and do about as much as they could in MS Windows.
I've encountered a couple MS Office files that I couldn't open in StarOffice. However, they're rare. I either requested a copy in some other format, or borrowed a MS Windows box. Besides, with new viruses appearing every day, no sane person accepts Office file formats from outside the company.
Support is a big issue for business. Microsoft's support is a joke; an expensive joke. Redhat offers support contracts, but I don't know how their price compares with MS's. Redhat is also funding a lot of kernel and Gnome development.
I've been in corporate meetings where Windows vs Unix/Linux were discussed. The decisions are generally neither rational, informed, nor careful. They generally have more to do with over-promoted idiots saving face by ignoring anything they don't understand. I watched my last company waste several million dollars on upgrading all the desktops to NT, and finish just in time to start upgrading them all to 2000.
A lot of the difference in support personel is the desktops. For the Unix solution, he was using graphical terminals. These things are far less complex than a windows desktop, because they have no local application software, no local writable storage, and no local configuration. They're interchangable appliances. So the support staff only has to take care of software on the servers.
Occationally they will have to replace a failed desktop, but it's a simple matter of swapping the entire unit.
The ms windows architecture has software locally installed on all the desktops. This software is almost never the same across all desktops in the company. Some stuff, like Office, will probably be common to all of them. However, in any large organization, different groups require (and pay for) different additional software packages. Different groups also require slightly different desktops with options like CD burners, printers, bar code scanners, check imprinters, etc. Desktops get replaced with newer models that require slightly different drivers.
All of this desktop variation, combined with the fact that they can write to local storage (possibly overwriting operating software) leads to a much more complicated desktop situation. This requires more personel.
The default installations of Gnome or KDE work almost exactly like MS Windows (with some extra features that users can ignore). I've put Windows users in front of both of them, and they've been able to adapt in minutes. In a business setting, new hires will have a learning period for the new interface, but it will be shorter than the time spent learning where supplies, forms, and coffee are kept in the office.
StarOffice reads and writes MS Office file formats. It even does presentations.
I think it's silly to ask the Qt and GTK people to support
two different abstraction layers in two different versions of their libraries. They have enough to do already.
Also, an X server that used directFB would still
need a config file, in addition to the directFB file. So you would have two ugly config files to deal with.
I've seen X add-on systems like Exceed. They don't work nearly as well as a real X desktop. They're never integrated into the local window system. They're duct taped onto the side.
What you propose would be more complicated, and uglier, than the current archetecture. The directFB clients would loose the network transparency and interclient communications of X. They would gain a relatively small amount of display speed (which they could have gotten with the direct draw X extention from vmware).
I think systems like directFB are a step backwards. XFree86 already has shared memory and direct draw extentions (see vmware), designed for high speed local graphics. When running remotely, the X library falls back to it's normal protocol, and the apps slow way down, but still operate. The network transparency of X is far, far too usful to encourage a crop of apps that can't use it.
It allows an amazing level of interprocess communication, but with a total neglect for security. Almost every VB and macro virus has depended on COM. Most of the serious explorer holes have used it. Microsoft is scrapping it for their new.net architecture.
MS Windows NT and 2k may have been intended to be ``rocksolid'', but they completely failed to meet that mark. A quick look through the bugtraq archives at securityfocus.com will prove that.
You also suggest that MS Windows stability problems are caused by faulty hardware and poor drivers. Since the vast majority of drivers are provided on on the OS install disc, or are at least approved by MS, any blame for them still belongs with the OS.
Linux and BSD have repeatedly shown themselves to be more stable than Win98 or NT on the same hardware. I personally have had several machines that could dual boot between various flavors of MS Windows (3.11, 95, 98, or NT) and Linux. Every one of them was more reliable while running Linux. In particular, I had an old Dell, that reached 452 days of uptime under redhat 4.2 before the power was cut. Under windows it had trouble staying up for more than a few days and often crashed during normal use.
Virtually every other experianced admin has similar stories. I regret I can't point you to any independant studies of reliability. Almost all of the tests have concentrated on performance.
Many of your other statements are also misguided. NT contains numerous design flaws. Obvious examples include the registry, the domain system, and COM. Please note that Microsoft is revamping the first, and scrapping the later two of those.
To the best of my knowledge, Dave Cuttler has designed only one other operating system, VMS. VMS was a good, solid, robust OS in it's day, and is still in use in some specialized sectors. However, there has to be a reason so many Vaxen were reconfigured to run BSD in the 70s and 80s.
I'm afraid your last claim ``Win2k is rocksolid and stable'' is obviously untrue. Just look at bugtraq, and at service pack 1.
We used to have a bunch of HP Apollo workstations that came with scsi cables that had been split this way. They were sliced down to individual wires. We had problems with conductors breaking over time. Our best guess was the the seperated wires flexed too much when vibrated by the fans and drives.
You have ignored 2 of the three people interviewed for the article. You are only translating the
Navy's PR statements.
Anthony DiGiorgio, a civilian engineer with the Atlantic Fleet Technical Support
Center in Norfolk said:
the NT operating system is the source of the Yorktown's computer problems
The article also says:
Ron Redman, deputy technical director of the Fleet Introduction Division of the Aegis Program Executive
Office, said there have been numerous software failures associated with NT aboard the Yorktown.
Even your interpretation of the Navy's statements is incomplete, because it's based on insufficent data. The article actually says
The Yorktown's Standard Monitoring Control System administrator entered zero into the
data field for the Remote Data Base Manager program. That caused the database to overflow and crash all LAN
consoles and miniature remote terminal units, the memo said.
That is not enough information to tell if the application crashed or the os crashed.
I agree that your interpretation, an app only crash, is far more probable.
I have often found non-technical users cannot distinguish between the OS and the application they use.
However, we lack sufficent data to really know what happened. In the absence of that data, I will accept the opinion of two men who do know what happened, DiGiorgio and Redman. They say NT was at fault.
I don't like that it uses special propretary cartriges. There's a seperate programmer that attaches to your computer to write the cartriges. I also got the impression you can't read from a cartridge, only write or erase. So it can't be used for sharing. I think the cartriges are expensive too.
I'd go with a cdr or cdrw unit. In a year or so there will probably be mp3 cd changers for cars. So you can have 6500 minutes of music on 10 cds.
Slashdot Mirror
It would be difficult to get the politicians to agree on policies.
As a grid gets larger, and more complex, it becomes harder to manage. There's an increasing (though still remote) risk of a cascade failure. https://en.wikipedia.org/wiki/Northeast_blackout_of_2003
According to Wikipedia, the energy density of lithium ion batteries is 250–676 Wh/L. The older acid-quinone battery had about 50 Wh/l. The article says the new chemistry gets about 2/3 of that, around 33 Wh/l. Lead-acid batteries are around 60–110 Wh/l. So this would probably be useless for mobile applications, but good for stationary purposes. Supposedly flow batteries can last indefinately, unlike lead-acid. It sounds like that would make them good for big data center UPS batteries.
https://en.wikipedia.org/wiki/Lithium-ion_battery
https://en.wikipedia.org/wiki/Flow_battery#Organic
https://en.wikipedia.org/wiki/Lead%E2%80%93acid_battery
Would this be effective for large UPS batteries in data centers? The current lead calcium acid batteries we use are expensive, heavy, and need to be replaced every few years. It would be interesting if we could replace the membrane and pump in some new fluid.
Autocad, at least a decent viewer.
1. Say something that excites morons.
2. Put up a web store.
3. Profit
Have you tried an electronics contact cleaner like Blue Shower? It's a little pricey, but works well for cleaning electronics.
http://mouser.com/ProductDetail/Techspray/1667-8S/?qs=sGAEpiMZZMvJqaFk9BIiv7vkq027uiu6ZZJKZ6s0jAI%3D
2. twm, fvwm, etc all had this too.
3. That was possible with many X window managers long before Windows tried it. However, most people found the visible borders useful.
4. I've seen this on both windows and X for years. I can't remember where I saw it first.
5. This page says the scroll wheel was invented by Mouse Systems, then popularized by Microsoft.
6. I believe Xfm did this in '92. I vaguely recall doing this on the Amiga in '86, but I could be wrong.
The only idea I've heard of first from Microsoft (or more accurately havn't heard of anyone else doing first) was transparent file system space saving by having the file system code combine identical files into a single copy on disk. It kept a hash of every file in the filesystem control structures, and when a new file was written, compared the hashes to see if it could merge files. Copying a file just created a second reference to the saved data. Writing to one copy caused the data to be split into seperate copies.
I'm sure Microsoft has invented other original ideas, but not nearly as many as they like to claim. Mostly they reimplement existing ideas.
Wouldn't the extra weight of the motor be partially offset by lighter breaking mechanisms, assuming that regerative breaking is the primary breaking mechanism, and friction breaks are mostly for emergency and parking?
It's technically feasable for the government to insert distributed spyware in their version of linux. A system like this would be technologically capable of monitoring keystrokes for trigger keywords, and reporting them to the government. It could use idle time to examine local files, sniff the local net, or join a distributed computing cluster cracking dissidents' crypto.
Since the government also controls the hardware, they can make it harder to replace the OS. They could make the processor only execute government signed kernels.
I don't know if the current Chinese government would do these things. But I believe they have the technical ability to do them.
Could the servers hide their real address, even from the hub, if the hub->server communications occured over a multicast group, and the server->hub communications used forged source address UDP packets?
A server would start by joining the multicast group. Next it would send an announcement to the hub, forging it's source ip, but including a randomly chosen, probably unique, identifier number. The hub would respond over multicast, using the identifier . They would procede to negotiating the server's inventory and capabilities. Eventually the hub would receive a request from a client that it could match to a server, and forward that request to the server. The server would begin transmitting the data directly to the client using forged source address UDP packets. If the multicast group becomes congested, the hub could create additional ones, and request existing servers switch to it.
I can see how someone could get other traders to digitally sign a letter of reference, then offer a set of those references when requesting files. It's similar to the web of trust concept in PGP. However, how do you keep a mole from tricking someone into signing for him, and letting him into the web? Once he's trusted, he could collect IPs, sign in other moles, etc. Also, how do you distribute negative references?
They used ext3 with the default 4k block size, against ntfs with a 64k block size.
That seems very unlikely. After all, this entire discussion is about MS selling software with a known major security hole.
I don't know about "grouping" them into a single icon. However, twm could alphabetize your icon manager in the late '80s.
I've encountered a couple MS Office files that I couldn't open in StarOffice. However, they're rare. I either requested a copy in some other format, or borrowed a MS Windows box. Besides, with new viruses appearing every day, no sane person accepts Office file formats from outside the company.
Support is a big issue for business. Microsoft's support is a joke; an expensive joke. Redhat offers support contracts, but I don't know how their price compares with MS's. Redhat is also funding a lot of kernel and Gnome development.
I've been in corporate meetings where Windows vs Unix/Linux were discussed. The decisions are generally neither rational, informed, nor careful. They generally have more to do with over-promoted idiots saving face by ignoring anything they don't understand. I watched my last company waste several million dollars on upgrading all the desktops to NT, and finish just in time to start upgrading them all to 2000.
The ms windows architecture has software locally installed on all the desktops. This software is almost never the same across all desktops in the company. Some stuff, like Office, will probably be common to all of them. However, in any large organization, different groups require (and pay for) different additional software packages. Different groups also require slightly different desktops with options like CD burners, printers, bar code scanners, check imprinters, etc. Desktops get replaced with newer models that require slightly different drivers.
All of this desktop variation, combined with the fact that they can write to local storage (possibly overwriting operating software) leads to a much more complicated desktop situation. This requires more personel.
StarOffice reads and writes MS Office file formats. It even does presentations.
I think it's silly to ask the Qt and GTK people to support two different abstraction layers in two different versions of their libraries. They have enough to do already.
Also, an X server that used directFB would still need a config file, in addition to the directFB file. So you would have two ugly config files to deal with.
I've seen X add-on systems like Exceed. They don't work nearly as well as a real X desktop. They're never integrated into the local window system. They're duct taped onto the side.
What you propose would be more complicated, and uglier, than the current archetecture. The directFB clients would loose the network transparency and interclient communications of X. They would gain a relatively small amount of display speed (which they could have gotten with the direct draw X extention from vmware).
I think systems like directFB are a step backwards. XFree86 already has shared memory and direct draw extentions (see vmware), designed for high speed local graphics. When running remotely, the X library falls back to it's normal protocol, and the apps slow way down, but still operate. The network transparency of X is far, far too usful to encourage a crop of apps that can't use it.
It allows an amazing level of interprocess communication, but with a total neglect for security. Almost every VB and macro virus has depended on COM. Most of the serious explorer holes have used it. Microsoft is scrapping it for their new .net architecture.
I have to disagree with most of your statements.
MS Windows NT and 2k may have been intended to be ``rocksolid'', but they completely failed to meet that mark. A quick look through the bugtraq archives at securityfocus.com will prove that.
You also suggest that MS Windows stability problems are caused by faulty hardware and poor drivers. Since the vast majority of drivers are provided on on the OS install disc, or are at least approved by MS, any blame for them still belongs with the OS.
Linux and BSD have repeatedly shown themselves to be more stable than Win98 or NT on the same hardware. I personally have had several machines that could dual boot between various flavors of MS Windows (3.11, 95, 98, or NT) and Linux. Every one of them was more reliable while running Linux. In particular, I had an old Dell, that reached 452 days of uptime under redhat 4.2 before the power was cut. Under windows it had trouble staying up for more than a few days and often crashed during normal use.
Virtually every other experianced admin has similar stories. I regret I can't point you to any independant studies of reliability. Almost all of the tests have concentrated on performance.
Many of your other statements are also misguided. NT contains numerous design flaws. Obvious examples include the registry, the domain system, and COM. Please note that Microsoft is revamping the first, and scrapping the later two of those.
To the best of my knowledge, Dave Cuttler has designed only one other operating system, VMS. VMS was a good, solid, robust OS in it's day, and is still in use in some specialized sectors. However, there has to be a reason so many Vaxen were reconfigured to run BSD in the 70s and 80s.
I'm afraid your last claim ``Win2k is rocksolid and stable'' is obviously untrue. Just look at bugtraq, and at service pack 1.
We used to have a bunch of HP Apollo workstations that came with scsi cables that had been split this way. They were sliced down to individual wires. We had problems with conductors breaking over time. Our best guess was the the seperated wires flexed too much when vibrated by the fans and drives.
Anti-virus filters only work on known viruses. That means they're almost always a day late for things like the love bug.
Anthony DiGiorgio, a civilian engineer with the Atlantic Fleet Technical Support Center in Norfolk said:
The article also says:
Even your interpretation of the Navy's statements is incomplete, because it's based on insufficent data. The article actually says
That is not enough information to tell if the application crashed or the os crashed. I agree that your interpretation, an app only crash, is far more probable. I have often found non-technical users cannot distinguish between the OS and the application they use. However, we lack sufficent data to really know what happened. In the absence of that data, I will accept the opinion of two men who do know what happened, DiGiorgio and Redman. They say NT was at fault.I don't like that it uses special propretary cartriges. There's a seperate programmer that attaches to your computer to write the cartriges. I also got the impression you can't read from a cartridge, only write or erase. So it can't be
used for sharing. I think the cartriges are expensive too.
I'd go with a cdr or cdrw unit. In a year or so there will probably be mp3 cd changers for cars. So you can have 6500 minutes of music on 10 cds.