In a word: Yes, Linux can schedule itself to multiple processors.
Lets go back in a time machine -- to Redhat 9, running on a dual PPRO 200 box.
Using the "top" utility on the box right now results in the following report (and note that the "CPU" column indicates the assigned CPU 0 or 1 for the process).
Notice that most of the "system processes" can use SMP. And this is Linux 2.4. 2.6 is far better for heavy-duty SMP. Is the kernel itself preemptive? Mostly. 2.6 has a far better scheduler for SMP. But it really doesn't matter much for my application (which is why the box is on 2.4).
"seems like another O/S based on processes/scheduling/filesystem. Isn't it time to move beyond those? todays needs are much more dynamic than the current 40-year-old O/S model offers."
Running on von Neuman architecture hardware, which is 70 years old (based on Konrad Zuse's work).
Running a graphic interface with a mouse that is 38 years old (1968 was the mother of all demos year), developed by Douglas Engelbart.
Unix showed up in 1970, making it 36 years old. But it was modelled on Multics, which in turn was modelled on older concepts -- I would put the conceptual roots to 1960 (for OS 360), and 1957, when Fritz-Rudolf Güntsch described automatic memory (virtual memory), which makes it 49 years old.
Ok, let's just scrap it all. We need a more dynamic model. And, while we are at it, let's scap 4 wheel cars.
I just started a contract with a company. I met someone I hadn't seen in 10 years, and we had a coffee (seeing as how I am "mellowing" in my old age). I mentioned that I had married, and had kids.
"...they let you reproduce?..." was the stunned response.
Snap! I guess I have left "geek" and am now a "nerd".
Both of these licenses will force an audit. As to certifying the results... There are several ways - you can let an external auditor do it, or let the BSA do it.
Go see your lawyer about the licenses you have executed in your jurisdiction. It would have been helpful to that before executing them.
If you don't like the license, don't do the product. Simple as that. And I (as a business) execute NDA and confidentiality agreements with my clients. As a result, I cannot allow such an audit. As a result, I don't use Monotype, Adobe, Microsoft, and a host of other vendors. As to being outside of the network effect... Tough shit. It's a cost of doing business. (and, thank you, I am fully booked).
"Compliance with Licences. If you are a business, company or organization, you agree that upon request from Adobe or its authorized representative you will within thirty (30) days fully document and certify that use of any and all Adobe software at the time of the request is in conformity with your valid licences from Adobe."
"If you are a business or organization, you agree that upon request from MI or MI's authorized representative, you will with thirty (30) days fully document and certify that use of any and all MI Font Software at the time of the request is in conformity with your valid licences from MI."
I am not bothering with Microsoft right now. You should have the idea already...
There *is* a problem with not having estate taxes. HIGH estate taxes.
And here it is: interest.
The thing that will let me retire! Now, if a billion dollar estate passes down, at 10%, it earns. Basically, it keeps growing. After 30 years, it has 3 billion (more, because of the compounding) added into it. So, the third generation starts with 5 billion, and ends up with 150 billion. The fourth generation: 5 trillion dollars.
Beginning with 40 billion? Its a trillion after the second generation, 30 trillion after the third. The fourth generation? A cool petabuck.
Of course, the economy HAS to compensate -- and does so by inflating.
The idea behind estate taxes was to compensate for compounding in that last generation. That is, equalize to remove the effect. Which means it has to be set at 90% or more (it depends on the interest rates and patterns over the generation).
And note -- generations 2 and onward don't have to do ANYTHING.
"Microsoft 54G Wireless Base Station Includes a 4-port 10/100 Ethernet switch 802.11g technology, transfer data up to 54Mbps Smart Windows CE-powered expandable platform Interoperable with 2.4GHz wireless frequency (11 Mbps and 54 Mbps) Backwards compatible Built-in firewall protection 256-bit Wi-Fi protected access"
So, as you can see, not only is it a Microsoft branded wireless router; it is based on Windows/CE.
How in HELL did you get the idea that I am a Linux zealot? How in HELL did you get the idea that I am a Microsoft hater?
I just posted facts. Do you wish to deny them? Or do you want to go off frothing at the keyboard with another ad hominem attack?
Indeed, the facts are that I am platform agnostic. I do find the head butting to be amusing, though, and I *never* think politics are a waste of effort. I do support Linux, but my ONLY public statement has been "If you need Linux, you will know why and when".
I also support Microsoft -- I am a shareholder in the company.
Of course Microsoft has dogfood in this area. They sell Microsoft branded networking gear. They sell an embedded OS that is suitable for appliances, including network devices.
Microsoft sells routers and access points. Microsoft produces an OS that is claimed to be usable as a commercial embedded system for routers. Doesn't that give them experience in the "networking hardware arena"?
Microsoft continually harps on "TCO" issues (Windows vs. Linux). Are you then claiming that the "TCO" for Windows/CE is higher than Linux?
And, just a final question: What Microsoft competitor are you talking about?
Well... I think that as backup goes, these systems basically suck (too expensive and slow). As a business proposition - I am thinking about it.
My price per gigabyte has just gone down. I am paying $109 CDN ($125.35 CDN after tax) for a 250GB drive (that's "decimal" based GB). Which is 50 cents (CDN) per GB. Translating to US, its 45 cents US per GB.
The drives have a manufacturing guarantee of 3 years, so my GB price per year is 15 cents. (Of course I don't rely on the drive to last - but I will get it replaced).
I use 3 drive RAID 5 - after overhead I get 473GB usable for every three drives. To cut the calculation short, its 24 cents US per GB per year, RAID5 configuration.
I have not allowed for environmentals (power, AC, &etc.). Lets go ahead and cheat. I'll just double it. 48 cents per GB per year.
I back the RAID5 onto DVD, and I pay 10 cents per GB.
I run my own Apache server, and have access to my data remotely.
But the online service is ORDERS OF MAGNITUDE more expensive. Which means that I should get into the business! (and explains why so many companies are in the game now).
Lets look at performance. My current "family dataset" is 130GB is size. It takes 5 minutes to transport a GB of data over my internal network (limited to 50mbps). External DOWNLOAD speed would cause that to be 50 minutes a GB, UPLOAD speed (important for backup) would cause that to be 250 minutes a GB. In other words, to back up my family dataset to a "web backup" would take 23 days, saturating my upload link 100%. Which is why I decided to keep backups locally in the first place.
In conclusion, its "thumbs down" on Web storage and backup providers. Until the price actually outways the performance issues (which the provider can do nothing about). My price point would be 5 cents per GB per month (5 dollars per month, for 100GB). I would sell the service at 50 cents per GB per month (and invest in a tape storage backup solution), because I am a capitalist pig (this undercuts my competition by an order of magnitude). Its "thumbs up" on offering the service locally!
You said "try wxWidgets". I did.I summarize my results:
Windows: Ok results. Application built with a wxWidgets builder.
Move application to Linux x86. Result: horrible. Faults all over. Back to cleaning up source code. Turns out that dialog window objects are somehow ok to use, even after they are out of scope with MS VC. Debugging is nasty. When I finally got it running, it looked reasonable (GNOME, Redhat 9). 5 hours.
Move application to Solaris (Sparc) 8. A few problems left. Go back, figure out wxWidgets issues. But... the application looks TERRIBLE. Fonts don't fit, dialogs don't fit, complete disaster visually. Manually modify to make it acceptable. But, notice that some UI paradigms that the original designer used didn't work (eg. Grayed out text being VITAL -- on the SUN grayed out meant drawn with a pattern, not legible). Modify application to fix these things. 25 hours.
Notice that HELP features don't work across platforms -- fix it up. 25 hours.
Test on AIX and HPUX. 20 hours.
Total time preparing a "cross platform" wxWidgets application: 75 hours. But then, I *may* not know what I am doing.
"Oh yes, I'm sure that _you_ could single handedly emulate every single game in existance, on a different CPU and a different graphics chip and all. Emulation isn't a trivial affair buddy, and it becomes increasingly problematic because of ever increasing complexity of the system you're trying to emulate. Yes, I'm sure everyone can wave Zsnes as proof that a console can be emulated, but look further up the food chain. It took several teams about 6 years to emulate a PSX acceptably. (But any existing emu still doesn't emulate at least a quarter of the PSX titles well enough!) It took more than 6 years of trying to emulate the PS2, and _still_ noone has more than a few demos and games that make it barely past the start menu, to show for that effort."
I am certain that I can't. But... it's Microsofts system (xbox) to Microsofts system (xbox 360). The only way to get a game on that box that Microsoft would even think about "emulating" is if Microsoft approved that game in the first place.
Microsoft has no need to reverse engineer the box. Nor the games. Any "emulation problems" must be either (1) implementation issues, or (2) design issues. Given that Microsoft is the largest supplier (or one of the largest) of x86 architecture commercial software, we have to assume a certain level on competance with implementation. So, the problem must be design... Crap, that is worse -- it must be an implementation issue... Crap, that doesn't bode well for a company making revenue selling x86 OSs...
In short, I don't know.
Ratboy.
What does this mean? If Microsoft is having "emulation problems", it must be either (1) an implementation
The original poster made some statements which were pretty much content free. I just reversed the thesis; it still held together (given that there was NO CONTENT).
In other words, a pure troll.
Now, on to your questions: most people do email, browsing, wp, spreadsheet, calendaring. In most cases, most information is imparted via words.
The "text interface" is a very lightweight, resource conserving approach to these problems. Typically, more efficient in use of compute resources. The "graphical interface" is heavier.
As to which is better? Servers typically do better with a "TUI". Simply because more resources can be devoted to the problem being solved. Clients? These days, computers are cheap, so the GUI tends to win.
But, my original point holds (and was smartly modded "redundent") -- the post I was replying to was content-free and a troll.
You are nuts. There are reasons why folks use Unix, TSO and DOS: most applications are textual in nature and benefit from that environment. Windows is not a good platform for building a text interface. Hell, it wasn't even a good platform for building a graphical interface. It did the job it was supposed to: provide an interface to the IBM PC hardware, but that is about it.
2 - Hack USB stick to flip removable bit. (More difficult, but easy enough). Combine with (1) for extra attack points.
3 - DIY USB stick. This one is custom. Easy enough -- the engineering would only cost 5 to 10K. This one is a USB stick, with extra logic to trigger DMA through USB, and potentially rooting x86 systems WITHOUT being visable at all. Hardware for this: cheap, software: can be as creative as you want to go.
Given that (1) worked 15 out of 20, it isn't even worth exploring 2 and 3, But that's just my opinion. I defend against all three possible attacks:
USB is limited to keyboard/mouse and storage ON AN X SERVER. Since the X server is booted from R/O media and doesn't have local disk, a rootkit generally doesn't matter. Nor does "auto exec" and double clicking a trojan doesn't work.
An attack is still possible, but the user would need to copy the file, make it executable, and run it.
Really, getting people to run an EXE from the 'net under the guise of trying to determine if they are "smarter" than the anti-malware crowd. Good one. If you run it, you are obviously dumber, no matter WHAT the result is.
Now, on to malware on Linux/Unix, and root-kits. Sure, it CAN happen, and it is quickly dealt with. I simply use hashes on files, and off-site them (tripwire).
Periodically, the hardware is refreshed with the files corresponding to the correct hash. Which ensures that the MAXIMUM time a root-kit can live is the time to refresh. Of course, the original vector could be exploited again -- I rely on regular security updates to plug those.
"There comes a time when legacy support just isn't viable anymore"
Let me comment on this.
My house. Built in 1970. Uses aluminum wiring, which is "not supported" anymore. Hasn't been for 20 years. I can still find inspectors to look at it for insurance purposes. Previous house used "knob and tube" wiring. Same deal, even though it hasn't been "supported" for, what, 50 years?
I drive a 1996 Saturn. Still supported (and passes its "drive-clean" with flying colours). I expect another 2 years from the vehicle.
Most of my other appliances are under 10 years old, but I fully expect them to be usuable beyond a 10 year window (TV, washer, drier, stove, microwave, fridge, &etc.).
So, why can't I expect at LEAST 10 years from my home computer?
I do. My main server is a PPRO 200x2, (vintage 1996). I did move to Redhat 9 a couple of years ago on this box. I use an IBM GL 300 (vintage 7 years) as an X terminal. Second X terminal is a Pentium 166 (vintage 13? years). Again, I upgraded the software a couple of years ago (to PXES 0.8 on the X terminals). I am retiring the P166, and getting a new computer this year.
The Compaq cannot run Linux, because it runs some "Windows only" stuff. That is probamtic under Linux/Wine. It works, and works well (on its third hard disk now!). As I said, the ONLY software that needs "updating" is the web browser. If Firefox doesn't want to support it, jolly for them. I'll go to Opera, or enable a proxy, or support a branched Firefox (after mourning for a bit, I like Firefox).
Slashdot Mirror
What the hell!
If you can't afford it, DON'T USE IT. End of story. There are alternatives. I cannot condone software piracy; that simply undermines the GPL.
I can't run GAMES. WHAAA!
Ratboy
In a word: Yes, Linux can schedule itself to multiple processors.
Lets go back in a time machine -- to Redhat 9, running on a dual PPRO 200 box.
Using the "top" utility on the box right now results in the following report (and note that the "CPU" column indicates the assigned CPU 0 or 1 for the process).
88 processes: 87 sleeping, 1 running, 0 zombie, 0 stopped
CPU0 states: 2.0% user 3.0% system 0.0% nice 0.0% iowait 93.0% idle
CPU1 states: 0.0% user 0.0% system 0.0% nice 0.0% iowait 100.0% idle
Mem: 125408k av, 105312k used, 20096k free, 0k shrd, 13944k buff
60224k actv, 1672k in_d, 964k in_c
Swap: 279712k av, 35784k used, 243928k free 52080k cached
PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME CPU COMMAND
9004 fred 19 0 1224 1224 888 R 7.3 0.9 0:00 0 top
1 root 15 0 532 488 456 S 0.0 0.3 0:29 0 init
2 root RT 0 0 0 0 SW 0.0 0.0 0:00 0 migration/0
3 root RT 0 0 0 0 SW 0.0 0.0 0:00 1 migration/1
4 root 15 0 0 0 0 SW 0.0 0.0 0:06 0 keventd
5 root 34 19 0 0 0 SWN 0.0 0.0 0:08 0 ksoftirqd_CPU
6 root 34 19 0 0 0 SWN 0.0 0.0 0:00 1 ksoftirqd_CPU
11 root 16 0 0 0 0 SW 0.0 0.0 0:00 1 bdflush
7 root 15 0 0 0 0 SW 0.0 0.0 3:56 1 kswapd
8 root 15 0 0 0 0 SW 0.0 0.0 0:19 1 kscand/DMA
9 root 15 0 0 0 0 SW 0.0 0.0 9:17 0 kscand/Normal
10 root 15 0 0 0 0 SW 0.0 0.0 0:00 0 kscand/HighMe
12 root 15 0 0 0 0 SW 0.0 0.0 0:25 1 kupdated
13 root 25 0 0 0 0 SW 0.0 0.0 0:00 0 mdrecoveryd
17 root 15 0 0 0 0 SW 0.0 0.0 2:13 1 kjournald
[fred@jupiter fred]$
Notice that most of the "system processes" can use SMP. And this is Linux 2.4. 2.6 is far better for heavy-duty SMP. Is the kernel itself preemptive? Mostly. 2.6 has a far better scheduler for SMP. But it really doesn't matter much for my application (which is why the box is on 2.4).
YMMV
Ratboy
"seems like another O/S based on processes/scheduling/filesystem. Isn't it time to move beyond those? todays needs are much more dynamic than the current 40-year-old O/S model offers."
Running on von Neuman architecture hardware, which is 70 years old (based on Konrad Zuse's work).
Running a graphic interface with a mouse that is 38 years old (1968 was the mother of all demos year), developed by Douglas Engelbart.
Unix showed up in 1970, making it 36 years old. But it was modelled on Multics, which in turn was modelled on older concepts -- I would put the conceptual roots to 1960 (for OS 360), and 1957, when Fritz-Rudolf Güntsch described automatic memory (virtual memory), which makes it 49 years old.
Ok, let's just scrap it all. We need a more dynamic model. And, while we are at it, let's scap 4 wheel cars.
Ratboy
Love it!
I just started a contract with a company. I met someone I hadn't seen in 10 years, and we had a coffee (seeing as how I am "mellowing" in my old age). I mentioned that I had married, and had kids.
"...they let you reproduce?..." was the stunned response.
Snap! I guess I have left "geek" and am now a "nerd".
YMMV
Ratboy
There is a provision that allows for information to be gathered. But, credit to Microsoft, the audit clause isn't in the "retail" products.
Which is good, because it means I can use Windows and Office in the course of my business.
Ratboy.
The BSA may be the authorized representative.
Both of these licenses will force an audit. As to certifying the results... There are several ways - you can let an external auditor do it, or let the BSA do it.
Go see your lawyer about the licenses you have executed in your jurisdiction. It would have been helpful to that before executing them.
If you don't like the license, don't do the product. Simple as that. And I (as a business) execute NDA and confidentiality agreements with my clients. As a result, I cannot allow such an audit. As a result, I don't use Monotype, Adobe, Microsoft, and a host of other vendors. As to being outside of the network effect... Tough shit. It's a cost of doing business. (and, thank you, I am fully booked).
YMMV
Ratboy
For spot audits?
That would be the license. Read the fine license sometime -- a lot of them allow the vendor to do an audit.
Lets take the Adobe License for Photoshop for example:
http://www.adobe.com/products/eulas/index.html
"Compliance with Licences. If you are a business, company or organization, you agree that upon request from Adobe or its authorized representative you will within thirty (30) days fully document and certify that use of any and all Adobe software at the time of the request is in conformity with your valid licences from Adobe."
Lets take Monotype next:
http://www.fonts.com/Legal/MI-EULA.htm
"If you are a business or organization, you agree that upon request from MI or MI's authorized representative, you will with thirty (30) days fully document and certify that use of any and all MI Font Software at the time of the request is in conformity with your valid licences from MI."
I am not bothering with Microsoft right now. You should have the idea already...
Ratboy
And my favorite commands on the ol' HP-2000 mini:
SANCTIFY and DESECRATE
"Sanctify file" moved the file to drum (basically, one-drive RAID 0 for all you young-uns). Desecrate moved it to the regular hard disk.
YMMV
Ratboy
There *is* a problem with not having estate taxes. HIGH estate taxes.
And here it is: interest.
The thing that will let me retire! Now, if a billion dollar estate passes down, at 10%, it earns. Basically, it keeps growing. After 30 years, it has 3 billion (more, because of the compounding) added into it. So, the third generation starts with 5 billion, and ends up with 150 billion. The fourth generation: 5 trillion dollars.
Beginning with 40 billion? Its a trillion after the second generation, 30 trillion after the third. The fourth generation? A cool petabuck.
Of course, the economy HAS to compensate -- and does so by inflating.
The idea behind estate taxes was to compensate for compounding in that last generation. That is, equalize to remove the effect. Which means it has to be set at 90% or more (it depends on the interest rates and patterns over the generation).
And note -- generations 2 and onward don't have to do ANYTHING.
Ratboy
Just the first link I found...
p ?catalog_name=CaliforniaComputer&category_name=&pr oduct_id=R84-00001&cookie_test=1
http://www.californiacomputer.com/Shop/product.as
google for "Microsoft router"
"Microsoft 54G Wireless Base Station
Includes a 4-port 10/100 Ethernet switch 802.11g technology, transfer data up to 54Mbps Smart Windows CE-powered expandable platform Interoperable with 2.4GHz wireless frequency (11 Mbps and 54 Mbps) Backwards compatible Built-in firewall protection 256-bit Wi-Fi protected access"
So, as you can see, not only is it a Microsoft branded wireless router; it is based on Windows/CE.
That covers both of my points.
Ratboy
How in HELL did you get the idea that I am a Linux zealot? How in HELL did you get the idea that I am a Microsoft hater?
I just posted facts. Do you wish to deny them? Or do you want to go off frothing at the keyboard with another ad hominem attack?
Indeed, the facts are that I am platform agnostic. I do find the head butting to be amusing, though, and I *never* think politics are a waste of effort. I do support Linux, but my ONLY public statement has been "If you need Linux, you will know why and when".
I also support Microsoft -- I am a shareholder in the company.
YMMV
Ratboy
Of course Microsoft has dogfood in this area. They sell Microsoft branded networking gear. They sell an embedded OS that is suitable for appliances, including network devices.
And Microsoft has been doing this for years.
Ratboy.
Microsoft sells routers and access points. Microsoft produces an OS that is claimed to be usable as a commercial embedded system for routers. Doesn't that give them experience in the "networking hardware arena"?
Microsoft continually harps on "TCO" issues (Windows vs. Linux). Are you then claiming that the "TCO" for Windows/CE is higher than Linux?
And, just a final question: What Microsoft competitor are you talking about?
Ratboy
Well... I think that as backup goes, these systems basically suck (too expensive and slow). As a business proposition - I am thinking about it.
My price per gigabyte has just gone down. I am paying $109 CDN ($125.35 CDN after tax) for a 250GB drive (that's "decimal" based GB). Which is 50 cents (CDN) per GB. Translating to US, its 45 cents US per GB.
The drives have a manufacturing guarantee of 3 years, so my GB price per year is 15 cents. (Of course I don't rely on the drive to last - but I will get it replaced).
I use 3 drive RAID 5 - after overhead I get 473GB usable for every three drives. To cut the calculation short, its 24 cents US per GB per year, RAID5 configuration.
I have not allowed for environmentals (power, AC, &etc.). Lets go ahead and cheat. I'll just double it. 48 cents per GB per year.
I back the RAID5 onto DVD, and I pay 10 cents per GB.
I run my own Apache server, and have access to my data remotely.
But the online service is ORDERS OF MAGNITUDE more expensive. Which means that I should get into the business! (and explains why so many companies are in the game now).
Lets look at performance. My current "family dataset" is 130GB is size. It takes 5 minutes to transport a GB of data over my internal network (limited to 50mbps). External DOWNLOAD speed would cause that to be 50 minutes a GB, UPLOAD speed (important for backup) would cause that to be 250 minutes a GB. In other words, to back up my family dataset to a "web backup" would take 23 days, saturating my upload link 100%. Which is why I decided to keep backups locally in the first place.
In conclusion, its "thumbs down" on Web storage and backup providers. Until the price actually outways the performance issues (which the provider can do nothing about). My price point would be 5 cents per GB per month (5 dollars per month, for 100GB). I would sell the service at 50 cents per GB per month (and invest in a tape storage backup solution), because I am a capitalist pig (this undercuts my competition by an order of magnitude). Its "thumbs up" on offering the service locally!
YMMV
Ratboy
"Which planet only has one moon?
Regards,
Jorge Boosh"
Um... ours?
fullphaser - pluto now has THREE moons (that we know of): Charon, Nix and Hydra.
More trivia: How many planets in our solar system have only a single moon? (answer: one)
How many have NO moons? (two or three, mecury, venus, and maybe xena)
Ratboy
You said "try wxWidgets". I did.I summarize my results:
Windows: Ok results. Application built with a wxWidgets builder.
Move application to Linux x86. Result: horrible. Faults all over. Back to cleaning up source code. Turns out that dialog window objects are somehow ok to use, even after they are out of scope with MS VC. Debugging is nasty. When I finally got it running, it looked reasonable (GNOME, Redhat 9). 5 hours.
Move application to Solaris (Sparc) 8. A few problems left. Go back, figure out wxWidgets issues. But... the application looks TERRIBLE. Fonts don't fit, dialogs don't fit, complete disaster visually. Manually modify to make it acceptable. But, notice that some UI paradigms that the original designer used didn't work (eg. Grayed out text being VITAL -- on the SUN grayed out meant drawn with a pattern, not legible). Modify application to fix these things. 25 hours.
Notice that HELP features don't work across platforms -- fix it up. 25 hours.
Test on AIX and HPUX. 20 hours.
Total time preparing a "cross platform" wxWidgets application: 75 hours. But then, I *may* not know what I am doing.
YMMV
Ratboy.
"Oh yes, I'm sure that _you_ could single handedly emulate every single game in existance, on a different CPU and a different graphics chip and all. Emulation isn't a trivial affair buddy, and it becomes increasingly problematic because of ever increasing complexity of the system you're trying to emulate. Yes, I'm sure everyone can wave Zsnes as proof that a console can be emulated, but look further up the food chain. It took several teams about 6 years to emulate a PSX acceptably. (But any existing emu still doesn't emulate at least a quarter of the PSX titles well enough!) It took more than 6 years of trying to emulate the PS2, and _still_ noone has more than a few demos and games that make it barely past the start menu, to show for that effort."
I am certain that I can't. But... it's Microsofts system (xbox) to Microsofts system (xbox 360). The only way to get a game on that box that Microsoft would even think about "emulating" is if Microsoft approved that game in the first place.
Microsoft has no need to reverse engineer the box. Nor the games. Any "emulation problems" must be either (1) implementation issues, or (2) design issues. Given that Microsoft is the largest supplier (or one of the largest) of x86 architecture commercial software, we have to assume a certain level on competance with implementation. So, the problem must be design... Crap, that is worse -- it must be an implementation issue... Crap, that doesn't bode well for a company making revenue selling x86 OSs...
In short, I don't know.
Ratboy.
What does this mean? If Microsoft is having "emulation problems", it must be either (1) an implementation
"Manual Writting" -- that's good.
I use LyX for input, and LaTex for the hard stuff. Seems to work out ok...
YMMV
Ratboy
The original poster made some statements which were pretty much content free. I just reversed the thesis; it still held together (given that there was NO CONTENT).
In other words, a pure troll.
Now, on to your questions: most people do email, browsing, wp, spreadsheet, calendaring. In most cases, most information is imparted via words.
The "text interface" is a very lightweight, resource conserving approach to these problems. Typically, more efficient in use of compute resources. The "graphical interface" is heavier.
As to which is better? Servers typically do better with a "TUI". Simply because more resources can be devoted to the problem being solved. Clients? These days, computers are cheap, so the GUI tends to win.
But, my original point holds (and was smartly modded "redundent") -- the post I was replying to was content-free and a troll.
Ratboy
You are nuts. There are reasons why folks use Unix, TSO and DOS: most applications are textual in nature and benefit from that environment. Windows is not a good platform for building a text interface. Hell, it wasn't even a good platform for building a graphical interface. It did the job it was supposed to: provide an interface to the IBM PC hardware, but that is about it.
The name "Windows" was chosen because Microsoft was selling vapour to forestall purchase of a product called "VisiOn".
Simple enough
1 - Rely on social engineering (easy, works well)
2 - Hack USB stick to flip removable bit. (More difficult, but easy enough). Combine with (1) for
extra attack points.
3 - DIY USB stick. This one is custom. Easy enough -- the engineering would only cost 5 to 10K. This one is a USB stick, with extra logic to trigger DMA through USB, and potentially rooting x86 systems WITHOUT being visable at all. Hardware for this: cheap, software: can be as creative as you want to go.
Given that (1) worked 15 out of 20, it isn't even worth exploring 2 and 3, But that's just my opinion. I defend against all three possible attacks:
USB is limited to keyboard/mouse and storage ON AN X SERVER. Since the X server is booted from R/O media and doesn't have local disk, a rootkit generally doesn't matter. Nor does "auto exec" and double clicking a trojan doesn't work.
An attack is still possible, but the user would need to copy the file, make it executable, and run it.
YMMV
Ratboy
Really, getting people to run an EXE from the 'net under the guise of trying to determine if they are "smarter" than the anti-malware crowd. Good one. If you run it, you are obviously dumber, no matter WHAT the result is.
Now, on to malware on Linux/Unix, and root-kits. Sure, it CAN happen, and it is quickly dealt with. I simply use hashes on files, and off-site them (tripwire).
Periodically, the hardware is refreshed with the files corresponding to the correct hash. Which ensures that the MAXIMUM time a root-kit can live is the time to refresh. Of course, the original vector could be exploited again -- I rely on regular security updates to plug those.
YMMV
Ratboy
"There comes a time when legacy support just isn't viable anymore"
Let me comment on this.
My house. Built in 1970. Uses aluminum wiring, which is "not supported" anymore. Hasn't been for 20 years. I can still find inspectors to look at it for insurance purposes. Previous house used "knob and tube" wiring. Same deal, even though it hasn't been "supported" for, what, 50 years?
I drive a 1996 Saturn. Still supported (and passes its "drive-clean" with flying colours). I expect another 2 years from the vehicle.
Most of my other appliances are under 10 years old, but I fully expect them to be usuable beyond a 10 year window (TV, washer, drier, stove, microwave, fridge, &etc.).
So, why can't I expect at LEAST 10 years from my home computer?
I do. My main server is a PPRO 200x2, (vintage 1996). I did move to Redhat 9 a couple of years ago on this box. I use an IBM GL 300 (vintage 7 years) as an X terminal. Second X terminal is a Pentium 166 (vintage 13? years). Again, I upgraded the software a couple of years ago (to PXES 0.8 on the X terminals). I am retiring the P166, and getting a new computer this year.
The Compaq cannot run Linux, because it runs some "Windows only" stuff. That is probamtic under Linux/Wine. It works, and works well (on its third hard disk now!). As I said, the ONLY software that needs "updating" is the web browser. If Firefox doesn't want to support it, jolly for them. I'll go to Opera, or enable a proxy, or support a branched Firefox (after mourning for a bit, I like Firefox).
Ratboy