(Essentially this is an NRA "Guns don't kill people argument..)
The way I see it is that most virus problems are a direct result of cluless users not practicing common sense. If it weren't for viruses, there would be some other way that they would mess up the system. Virus or no, computer support personel -always- have their time full.
But, where do you draw the line between what's OS and what's an app? It's pretty easy to consider bash as an OS bit, and xroach as not, but what about grep? awk? vi? emacs? gcc?
Let's look at gcc specifically : In the unix world, a compiler is considered one of the essential bits of the OS, while in the windows world, it's considered to be a major app, and I doubt the difference in perception is significantly influednce by the presence of an IDE...
Erp... IE tossed my cookies and just posted this as AC..
Everyone seems to be under the assumption that in a power-saving mode a system has to effectively be off? What's wrong w/ a partial powerdown? The CPUs in laptops can do it, why not on servers?
Carrying on the car metaphor so effectively used earlier; yes, a car that's got 100k mi of stop and go driving is going to be in worse shape than one that's been driving long distances on open road.
OTOH, consider two cars used for open freeway driving, one cruises at a respectable 65mph, while the other is constantly redlined, and driving as fast as it can go, the one that's run harder isn't going to last long (ie Indy/Nascar motors generally get replaced after each race)
So, why run a drive at 10k RPM all the time, when it'll more than do the job at 5k? Why run your CPU at 1GHz, when it can handle all essential functions at half, or a quarter of that?
Continuing w/ the car analogy, what's the point of running the engine as hard as you can, if you have to drop down to 2nd gear to flow with trafic? (Perhaps having the clutch partially engaged would be a more accurate way of describing idle CPU cycles)
Dealing with drives could probably be done on the controller in such a way that the OS wouldn't even have to deal with the process, and modern CPUs, with their cores running at nearly 10x their external speeds, would have little problem dropping their multipliers to only 3-4x, with no visible effect to the rest of the system (not to mention the reduced heat production, and it's effect on procs)
And really, the CPU and drives account for most of the energy use in a high-performance computer. Unlike a desktop, where the monitor is another big issue, many servers, such as those in cluster/farm situations, don't even have their own monitors.
In todays rapidly moving hardware world, where this year's high-end servers are next years doorstops, implimenting, and geting adopted new standards should be a relatively simple task, as long as MS agrees to support it (as for the big iron (or Macs...), with thier proprietary OSes, having the technology supported is even less of an issue).
Having just finished reading the book I have to wonder how much things have changed since the adoption of high-level languages, which he mentions as more of a new toy/accademic curosity than a real-world tool.
One would only think that using HLLs would allow people to more effectively interface with and understand somebody else's code.
Getting back to the original article, it seems to me that rewriting the whole thing from scratch is often one's first instinct, as something that's sprung entirely from your own mind is always going to be easier to comprehend and modify than something you've had to learn. Two important questions come to mind tho' : Will it be more effective for you to rewrite than to modify the existing system? Will your version be easier for the next guy to maintain than the existing one?
I'm sure that the guys at google are just as interested as you are, and are pouring over logs + data-files trying to figure out how somebody 'cheated' the system.
It's just not the same... The SID chip had analogue components in it. If you listen to the purists, no DSP is ever going to sound like a piece of analogue kit... something about the sound being sterile, or something...
Of course, there are the purists... and, for the most part, you can't tell analogue from a good DSP once the instrument's in the mix... But we're talking musical instruments here. If you look at it more along the lines of vintage wines/cars/guitars it'll make more sense than looking at it as buying an obsolete sound chip.
The problem with picking a standard is that file formats change over time. Any WP format more complicated than an ASCII text file has gone through numerous invokations as features are added/fixed through various revisions of software. Short of comming up with a new standard, you'll have to make compromises somewhere, be it in portability, reliance on a propriatary package or capability.
Perhaps it's time to actually have some standards body define a standard format for word processing, that presents an acceptable minimum of functionality. The cries of XML! XML!, while partially missing the point, as XML itself isn't up to the task, might be a start, since, at least in theory, an XML-based format would be both extensible and maintain backwards compatability, and have the added bennefit of being relatively easy to write implementations of.
Of course, opinions ( ie (_)*(_)s )what exactly constitutes the minimum acceptable functionality may vary but, as you know, committees are good at making sure that nobody is any happier with the results than anybody else.
Why doesn't such a standard exist already? Simple, no company wants to write a stable, open spec. for a minimal document, and if somebody were to attempt to do so other companies would not likely give it acceptance. This is why some committee, be it an organization, such as ISO, or a group of 'interested parties' agreeing to work together would be the best situation. This is probably outside of JEDEC's charter, but y'all may be able to pass the suggestion onto the appropriate parties.
So, in closing, such a spec would need to be:
simple
powerful
open
extensible
ensure backwards compatability
... and most importantly set in stone
As you mentioned in your write-up, RTF fit most of these, but under the sole control of MicroSoft it kept changing. Perhaps, with their blessing, RTF would be a good basis for the new spec, as long as you can solidify the format, and keep it fixed for a length of time, ensuring backwards compatability between revisions.
This, and the recent 'what can I do for OSS' threads are really great... Next term I'm taking an 'Advanced Operating Systems Class' at school, and, partially 'cuz there's only going to be 3-4 ppl in it, and partially 'cuz it's never been taught before, it's going to be an independant project. Talking to the prof. , the sylabus will basically say "Find some open-source OS project, and do something for it".
Not only is the coursework going to be open-ended, but he's going by a -very- liberal interpretation of 'open-source OS'... I mentioned that the Berlin project was looking for volunteers, and he said it'd be kosher. So, as long as it can reasonably be justified as expanding my knowledge of OS internals, it's a go.
Anybody wanna try recruiting me for their project, and save me the trouble of making a decision myself? (not to mention the ego boost that it'd cause...)
What I meant was that if you didn't go with vigroup you're other option would be to rewrite the system code to handle a new format of the/etc/groups file. While the script may be simpler, and require far less effort, it's a bit of hack...
I don't have the mod points, so I'm forced to post to add my agreement here. If it's more your style, something perly, or whatnot might be better, but either way it's still the simplest solution. Doing it the Right Way would require tracking down and rewritting the neccessary bits of system code, definately a non-trivial task.
Grr... Got dirty power here, and if I run the monitor at anything ovr 60Hz, the picture gets wobbly... I've tried multiple monitors, vid-cards and systems, so either I've got bad power or some massive RF source burried in the walls here..
Re-PC is the bigest one... they're located down by Safeco Field. If you look in the Yellow Pages under Computers - used you should find them, they've also got ads in the monthly free computer magazine (Puget Sound Computer User) that you find at grocery stores and whatnot...
For those who don't like subtlety, these (or the local equivalent) are more good places to look for dealers close to you.
Of course, the obvious answer would be to pick up a >500MHz machine to do the upgrade with (my personal rule of thumb is that un upgrade isn't worthwhile unless it doubles performance), and then claim the old 200MHz system for your firewalling needs.
I know that Seattle's got at least 2 large used computer stores, look around, and see if you do. (I wouldn't reccomend buying a whole used system through the mail... shipping a $20 comp. is just rediculous).
You might wanna try getting friendly with the staff of a local computer store that does a bit of business in system upgrades; all those old P-166s have to go -somewhere-.
Finally, don't forget about usenet. Your local forsale newsgroup is a good place to look for 'obsolete' systems...
Something about it reminds me of the computer lab we had back in highschool. It was a pair of decrepit Xenix machines that were constantly getting rooted by the AP students. In fact we developed a set of unwritten rules for a game we eventually called Root Wars. It centered around trying to get on (and stay on) as root while preven others from doing so, without inconveniencing the normal users (too much).
Of course, there was the time that a fork-bomb killed the IDE controller (actually, it was the idiot TA power-cycling the machine too fast when it stopped responding, but, everone still blames me for the hardware dammage...).
I've always thought that something like this could be a successfull multi-player game, if handled with the right degree of abstraction (on par with the connection between Quake and real combat). A good Gibsonesqe hacking game...
It's only inevitable if we survive the Y2K crisis, err... umm... the second coming... Hrmm... how about a bio-engineered super-plague... err.. new version of Windows..
Slashdot Mirror
(Essentially this is an NRA "Guns don't kill people argument..)
The way I see it is that most virus problems are a direct result of cluless users not practicing common sense. If it weren't for viruses, there would be some other way that they would mess up the system. Virus or no, computer support personel -always- have their time full.
... of course, don't some theories suggest that multi-cellular life is a result of a parasitic relationship 'gone bad'?
Great news... Now I can say that I'm promoting the evolution of the species by not having that nasty rash down there treated.
No... like I said... no skipping. I'm typing this right now, in Netscape, and untarring a compressed Linux kernel... and tarring it back up...
no skipping.
And I'm still using a PPro200. Maybe it's 'cuz I'm only using an ISA SB16...
Now, on windows NT... I get audio skipping when moving windows arround...
Damn...
PPro-200, 64M, adaptec 2940UW, and some fast-wide 5k/7k RPM drives...
no skipping.
Catchs somebody having extramarital sex with an animal and, using pictures of the act, create a photomosiac of their spouse.
Patents don't cover criminal activities, do they?
Perhaps you meant goatse.cx...
But, where do you draw the line between what's OS and what's an app? It's pretty easy to consider bash as an OS bit, and xroach as not, but what about grep? awk? vi? emacs? gcc?
Let's look at gcc specifically : In the unix world, a compiler is considered one of the essential bits of the OS, while in the windows world, it's considered to be a major app, and I doubt the difference in perception is significantly influednce by the presence of an IDE...
Erp... IE tossed my cookies and just posted this as AC..
Everyone seems to be under the assumption that in a power-saving mode a system has to effectively be off? What's wrong w/ a partial powerdown? The CPUs in laptops can do it, why not on servers?
Carrying on the car metaphor so effectively used earlier; yes, a car that's got 100k mi of stop and go driving is going to be in worse shape than one that's been driving long distances on open road.
OTOH, consider two cars used for open freeway driving, one cruises at a respectable 65mph, while the other is constantly redlined, and driving as fast as it can go, the one that's run harder isn't going to last long (ie Indy/Nascar motors generally get replaced after each race)
So, why run a drive at 10k RPM all the time, when it'll more than do the job at 5k? Why run your CPU at 1GHz, when it can handle all essential functions at half, or a quarter of that?
Continuing w/ the car analogy, what's the point of running the engine as hard as you can, if you have to drop down to 2nd gear to flow with trafic? (Perhaps having the clutch partially engaged would be a more accurate way of describing idle CPU cycles)
Dealing with drives could probably be done on the controller in such a way that the OS wouldn't even have to deal with the process, and modern CPUs, with their cores running at nearly 10x their external speeds, would have little problem dropping their multipliers to only 3-4x, with no visible effect to the rest of the system (not to mention the reduced heat production, and it's effect on procs)
And really, the CPU and drives account for most of the energy use in a high-performance computer. Unlike a desktop, where the monitor is another big issue, many servers, such as those in cluster/farm situations, don't even have their own monitors.
In todays rapidly moving hardware world, where this year's high-end servers are next years doorstops, implimenting, and geting adopted new standards should be a relatively simple task, as long as MS agrees to support it (as for the big iron (or Macs...), with thier proprietary OSes, having the technology supported is even less of an issue).
Having just finished reading the book I have to wonder how much things have changed since the adoption of high-level languages, which he mentions as more of a new toy/accademic curosity than a real-world tool.
One would only think that using HLLs would allow people to more effectively interface with and understand somebody else's code.
Getting back to the original article, it seems to me that rewriting the whole thing from scratch is often one's first instinct, as something that's sprung entirely from your own mind is always going to be easier to comprehend and modify than something you've had to learn. Two important questions come to mind tho' : Will it be more effective for you to rewrite than to modify the existing system? Will your version be easier for the next guy to maintain than the existing one?
Actually, you'd need "news +for nerds" since the 'for' is normally ignored, being such a common word.
I'm sure that the guys at google are just as interested as you are, and are pouring over logs + data-files trying to figure out how somebody 'cheated' the system.
160-some-odd posts and nobody's suggested a beowulf of these things?
And a reference to Sandra Bullock instead of Natalie Portman...
my-oh-my, times are a'changin'
Dunno, but after that thing's been /. for a a day or two, their budge might be tapped, and they'll just have to make do with a few XTs and some arcnet.
It's just not the same... The SID chip had analogue components in it. If you listen to the purists, no DSP is ever going to sound like a piece of analogue kit... something about the sound being sterile, or something...
Of course, there are the purists... and, for the most part, you can't tell analogue from a good DSP once the instrument's in the mix... But we're talking musical instruments here. If you look at it more along the lines of vintage wines/cars/guitars it'll make more sense than looking at it as buying an obsolete sound chip.
Perhaps it's time to actually have some standards body define a standard format for word processing, that presents an acceptable minimum of functionality. The cries of XML! XML!, while partially missing the point, as XML itself isn't up to the task, might be a start, since, at least in theory, an XML-based format would be both extensible and maintain backwards compatability, and have the added bennefit of being relatively easy to write implementations of.
Of course, opinions ( ie (_)*(_)s )what exactly constitutes the minimum acceptable functionality may vary but, as you know, committees are good at making sure that nobody is any happier with the results than anybody else.
Why doesn't such a standard exist already? Simple, no company wants to write a stable, open spec. for a minimal document, and if somebody were to attempt to do so other companies would not likely give it acceptance. This is why some committee, be it an organization, such as ISO, or a group of 'interested parties' agreeing to work together would be the best situation. This is probably outside of JEDEC's charter, but y'all may be able to pass the suggestion onto the appropriate parties.
So, in closing, such a spec would need to be:
- simple
- powerful
- open
- extensible
- ensure backwards compatability
... and most importantly set in stone
As you mentioned in your write-up, RTF fit most of these, but under the sole control of MicroSoft it kept changing. Perhaps, with their blessing, RTF would be a good basis for the new spec, as long as you can solidify the format, and keep it fixed for a length of time, ensuring backwards compatability between revisions.Just remember, Linux is not an OS... so it'd be:
..."
"GNU's not Unix's not Unix's not Unix... Unix's/Not an OS running in a browser on
This, and the recent 'what can I do for OSS' threads are really great... Next term I'm taking an 'Advanced Operating Systems Class' at school, and, partially 'cuz there's only going to be 3-4 ppl in it, and partially 'cuz it's never been taught before, it's going to be an independant project. Talking to the prof. , the sylabus will basically say "Find some open-source OS project, and do something for it".
Not only is the coursework going to be open-ended, but he's going by a -very- liberal interpretation of 'open-source OS'... I mentioned that the Berlin project was looking for volunteers, and he said it'd be kosher. So, as long as it can reasonably be justified as expanding my knowledge of OS internals, it's a go.
Anybody wanna try recruiting me for their project, and save me the trouble of making a decision myself? (not to mention the ego boost that it'd cause...)
What I meant was that if you didn't go with vigroup you're other option would be to rewrite the system code to handle a new format of the /etc/groups file. While the script may be simpler, and require far less effort, it's a bit of hack...
I don't have the mod points, so I'm forced to post to add my agreement here. If it's more your style, something perly, or whatnot might be better, but either way it's still the simplest solution. Doing it the Right Way would require tracking down and rewritting the neccessary bits of system code, definately a non-trivial task.
Grr... Got dirty power here, and if I run the monitor at anything ovr 60Hz, the picture gets wobbly... I've tried multiple monitors, vid-cards and systems, so either I've got bad power or some massive RF source burried in the walls here..
Re-PC is the bigest one... they're located down by Safeco Field. If you look in the Yellow Pages under Computers - used you should find them, they've also got ads in the monthly free computer magazine (Puget Sound Computer User) that you find at grocery stores and whatnot...
For those who don't like subtlety, these (or the local equivalent) are more good places to look for dealers close to you.
Of course, the obvious answer would be to pick up a >500MHz machine to do the upgrade with (my personal rule of thumb is that un upgrade isn't worthwhile unless it doubles performance), and then claim the old 200MHz system for your firewalling needs.
I know that Seattle's got at least 2 large used computer stores, look around, and see if you do. (I wouldn't reccomend buying a whole used system through the mail... shipping a $20 comp. is just rediculous).
You might wanna try getting friendly with the staff of a local computer store that does a bit of business in system upgrades; all those old P-166s have to go -somewhere-.
Finally, don't forget about usenet. Your local forsale newsgroup is a good place to look for 'obsolete' systems...
Something about it reminds me of the computer lab we had back in highschool. It was a pair of decrepit Xenix machines that were constantly getting rooted by the AP students. In fact we developed a set of unwritten rules for a game we eventually called Root Wars. It centered around trying to get on (and stay on) as root while preven others from doing so, without inconveniencing the normal users (too much).
Of course, there was the time that a fork-bomb killed the IDE controller (actually, it was the idiot TA power-cycling the machine too fast when it stopped responding, but, everone still blames me for the hardware dammage...).
I've always thought that something like this could be a successfull multi-player game, if handled with the right degree of abstraction (on par with the connection between Quake and real combat). A good Gibsonesqe hacking game...
It's only inevitable if we survive the Y2K crisis, err... umm... the second coming... Hrmm... how about a bio-engineered super-plague... err.. new version of Windows..