Boo-hoo. Then stop demanding that the rest of us pick up the slack where you left off. You acknowledge it is impossible, why saddle the rest of us with an impossible task, that causes a lot of inconvenience? How about THAT, HMMM??
How is K-Mart requiring carding "demanding that the rest of the [society] pick up the slack"? I see it as a tool for enforcing my values and standards. How is a private company choosing who they sell product X to censorship? How is it "parenting"? Do you have any children of your own?
Wonderful! Gay Day at the local high school is ON. After all, gay people are part of society, let them imprint some values on the next generation, too!
That's not what I said, and you know it. But since you bring up the Gay community, I will respond to that. Like it or not, they are a part of the society. It is up to us parents to teach tolerance and acceptance to those not like ourselves. This narrow-minded "you are wrong because you're different" thought pattern just teaches children to close off their minds. This is exactly what led to the hate-movement of the 1960's, and the hate riots in Germany during the 1990's.
You know what sickens me? The hypocrisy of so-called parents when it comes to who is responsible for their children's behavior.
My spouse and I are ultimately responsible for our children. We know this, and strive for excellence in everything we do with our children, teaching them our values. If I wanted my children to play "Soldier of Fortune", then I would be happy to buy it at K-Mart (or wherever), with my identification. K-Mart's carding is a tool, not censorship.
Go ahead and boycott K-Mart and Wal-Mart. I think they will do fine without your business.
--
Re:It's still a democracy.....use it!
on
Lawsuits Suck
·
· Score: 2
What he's proposing is to create a SIG, not a political party. A SIG lobbies congressmen, and in turn, pays for a part of their campaign.
The wording of the article tells me that the guywho wrote it doesn't have a slightest clue what he is talking about, but wants to scre everyone in sight to death
Heard on NPR, quoting Jesse Ventura, the governor of Minnesota. He stated, and quite correctly, that many media outlets are no longer in the business of reporting the news, but instead, making the news. They're slaves to the almighty buck.
He went on to state that there's nothing wrong with being a slave to the almighty buck, but if they are, they should not be misleading consumers by stating that they're "journalists" and not "tabloid artists".
On the web, as with anywhere else, "consider the source". Many people no longer do this, but this is absolutely critical when judging the accuracy of any information, not limited to that shown in the news. Sensationalism in "professional" journalism is very much alive today as it were in the 1890's, when "yellow" journalism coursed through the pages of "credible" newspapers.
No the world revolves around whiny idiots who can't set limits for their own children. Who don't take the time to look around their kids' rooms, and who don't take the time to run their own household. The intelligent people like us must suffer the existence of these marching morons.
It kind of shows that you don't have children of your own. I believe in privacy, both of myself and of "my" children. It works both ways--I don't go rummaging around in their stuff, and they don't go rummaging around in mine. I lead a very busy schedule (work 60+ hours/week), and so does my fiancee. We welcome all tools that will allow us to enforce our rules that we set, such as this K-Mart bit.
Seriously, there is no way that anyone can reasonably maintain a 24/7 surveillance on their children. Like it or not, this is what will be required to filter most of the social rubbish that will be imprinted on our children. Instead, let the society and the corporations help you in building the tools to monitor the products' influences on our children. This is precisely what K-Mart is doing; as the other posters have suggested, if I wanted my children to grow up playing violent video games, I will personally buy them myself. This new-age "society has no place in rearing my children" rubbish really sickens me; for thousands of years, small communities imprinted their values on the children. This indoctrination still happens daily at schools. I guess that your children will be homeschooled as well?
If you don't like it, fine. Boycott K-Mart and Wal-Mart. That is your right as a consumer. However, when you have your own children, and they're at the age where they want to play video games, even violent ones, and are playing them in your living room, at least have the decency to listen to us when we say, "We told you so.".
The SUV's and the trucks that're quite popular (check out the JD Power's stats) aren't exactly frugal in their petrol expenditures, you know....
Do you want to twist statistics? GM's earnings were *down* this year compared to the last. Toyota is still growing. GM had net revenues of approx. $5 billion, Toyota $4.5 billion. Looking at the rate of growth in both companies, Toyota will surpass GM in net earnings by 2005. Easily.
What does this information actually mean? Absolutely nothing! Why? because both indicators are completely dependent upon the trade agreements and such between the US and other countries, as well as buyer trends and such.
All we managed to prove is that the large companies in the automotive business are rather resilient, and that with deep enough pockets, it's possible to ride out the storm.
Granted, it wasn't a bunch of Linux boxen, but at one of the places I've worked at, we had 500+ Sun boxen running various versions of SunOS, 200+ combined HP and IBM workstations, 150 or so Macs, mainly in the marketing, plus about 100 (or thereabouts) Windows boxen.
This was a few years ago, and since then, they've been bought out by Cadence, so I don't know what's going on there now.
Sorry, I parsed your previous "error in the config" as "error in the configuration file", and not an "error in the hardware configuration as the program currently sees it".
In the case you just described, yes, it was rather idiotic, either the driver or the userland program should've acknowledged the problem and did *something* other than just blanking out.
You're forgetting that both IBM and HP are rather large companies with rather large support services divisions. Take, for example, IBM. IBM Global Services, their consultancy group, will support anything from CP/M to IBM SP2 (and probably Sun E10k) to OS/390 and AS/400 mainframes. Same with HP. They tout themselves as a one-stop support center for everything. Large companies pay huge amounts of money to these organisations for support.
So why are they bulling ahead with this certification lab? It makes supporting these applications (yes, they support applications too) easier if they know the ins and outs of the app. They'll of course find this out while they "certify" the piece of software.
Currently, because of the lack of certification, they will support any app, even the ones that I built at home while drinking wine, and is this unmaintainable clusterf*ck, if it made them some dough. Suppose they certify someone else's OSS project, which does essentially the same thing, except missing features x and y. Will they support my app for the company? Hell no. They will tell them to move to the someone else's app, because it's actually possible to maintain the bloody thing.
See Caldera? See IBM? See HP? See SCO? Hell, see MS? They're all trying to move to a constant revenue model that comes from providing a service, not maintaining the one-time revenue of a product.
How hard can it be to detect an error in the config and drop back to the console - IANACP (I am not a C programmer)
ITYM, "IANAP" (I am not a programmer). As anyone with even a small grounding in language design will tell you, Syntactic errors are bloody easy to detect. Just look at all the typos that your compiler will spit out. Semantic errors, however, are another story. Bloody things are almost impossible to detect at the machine level.
Just about any monkey (or a Turing Machine, whichever) can find that something is syntactically amiss. Tools like lex and yacc make it rather simple to build even the most complicated of configuration scripts. However, validating for semantic correctness in even a moderately complex configuration script is another matter; it is multiple orders of magnitude more difficult and computationally intensive to verify this (and for some languages, like English, whose synctactic definitions certainly do exist, the technology for verifying semantic correctness doesn't yet exist).
Assuming that you're misinformed rather than being a troll, SSL needs the signature to prevent a man-in-the-middle attacks, or any hosts masquerading as another host. Think about it. You're going to amazon.com to buy books. I poisoned your DNS so that www.amazon.com points to my domain, and I act as a proxy for the real amazon.com until you're ready to make a purchase. When you do, through my SSL server, without the certificate, and I log your credit card information for my future use.
With certificates, you can be reasonably sure that the www.amazon.com is the real amazon.com and not my little false fpos thing. Without it, well, it opens things up for all sorts of interesting attacks.
imho there are a lot of things in the article that aren't the result of Debian being "bad" they are the result of incompetence or of overlooking obvious details.
I agree completely. You have to take a proactive measure to security, regardless of what system you run, be it Linux, *BSD, Winblows, whatever. Then, someone will counter, "but I have to install N machines, same (or very similar) configuration for the foo application!".
Well, son, you have several choices.
You can manually configure each and every box. Of course, this is the stupidist way of doing so, but guaranteed to get you more OT. Then again, I have not known a competent SysAdmin who likes work like this.
You can build your own distro. But then, do you really want to do this?
Build your own installation scripts. Not only will this save you time by being able to hand this installation process over to a kid getting minimum wage, but now you get to do something a little more challenging than building a box and making it secure. At the very least, you can build one box properly, install a skeleton OS on the others, and just do a batch copy between the boxen.
I did this at one of my clients' sites, where we had over 600 SPARC boxen. Of course, took it one step further, by providing an install server, but then, this was also in a very protected subnet.
Do it only once, but for Seldon's sake, do it right!
Looking at the current design and usage patterns of 8-10lane highways popular in the US, there's an upper limit of what can be utilised. Sure, if a person is travelling more than a few km, the person will use the entire highway. However, at major exits and such, the bottleneck isn't just the highway itself, but also the lower-capacity roads below. Frequently, I've found myself moving along at ~120km/h, while the people taking the exit ramp (or the entry ramp to another, clogged, highway) were bogged down for literally hundreds of car-lengths.
Unfortunately, I don't have a solution (Actually, I do--while living in central Tokyo, I commuted on off-peak hours on the train--a solution unworkable in most American metroplitan areas), other than abusing the flex hours.
Sorry, not to be pedantic, but what you just described is 5 529mbps, far in excess of what you just described with your DSL saturation issue.
I think the main issue with your "inability" to saturate your DSL is an issue of bottlenecks; with an analogue modem connection, your modem is the bottleneck, with your DSL, some upstream provider somewhere is.
The question is, will they ever "roll it out" to beyond what it is now? I mean, sure, they use IPv6, sure, their backbones are probably an order of magnitude fatter on a per-host basis, but would "they" ever roll it out, or would the current IPv4-based Internet just migrate to IPv6 when the specs are "done", tunnelling some legacy IPv4-based traffic in a "4-bone", or doing some sort of weird IPv6-IPv4 NAT? Or will the current IPv4-based Internet plod on, NAT-ting everywhere (dear lord, I hope not)?
Then again, when I run traceroute(1) everywhere, I almost always see a 10.x.y.z somewhere:-)
No, but there were looser (read: no) restrictions on who can be on the network. I2 has a rather stringent policy on who can be on and who can't. Please read the FAQ:-).
The biggest problem Is that the internet is already too much of a standard.. it's easy to change standands if they're not too popular or used that much... Want proof of this? In some cases, we're still using the kermit protocol to download.. if anyone doesn't know, this is one of the first protocols written for the internet. So that leaves Internet2 to the scientific/idle rich category. By the way, has anyone done any real world benchmarking? As in, take a copy of the slashdot code, and see how it runs on internet2?
Uh.... I think I'm missing something here. First, Internet2 is (was?) a research-only network. It's not (supposed to be) a place for pr0n and other commercial use.
You speak of standards. The Internet, as we see it, is just a collection of heterogenious networks in a homogeneous naming space (well, mostly--a discussion of NAT is beyond the scope of this comment). You want to move to I2? Fine. Go enroll as student or faculty at one of 180 research institutions. Just don't go there for much Quake use:-)
You also take a look at kermit. Kermit was not an Internet protocol per se, but was a terminal-terminal download protocol that assumes unreliable network streams. Yes, that means certain instances of kermit-over-IP existed, but it doesn't mean that kermit was necessarily an Internet application-level protocol, where most transfer tools assume a TCP-level (or equivalent) functionality, leaving error detection and correction (among other things) to the lower-level protocols, and it dealing more with the application-specific information.
As for benchmarking/. code, I really don't see the difference of its code running on I2, Internet, or my home LAN; provided that the host is mated to a decent backbone, its performance would be more dependent upon load/system configuration than the underlying network itself. I2, with its much smaller user base and a similar backbone, consequently has much lower load, so it should perform better.
Besides, if you wanted to compare/contrast I2 to Internet, you'd probably do a network traffic analysis (peak bandwith, peak latency, multiplexing capabilities, etc), which would be a function of the routers than anything else.
Athlons have, IMO, proved themselves in the field by 1) being out in the "wild" for over a year, and 2) not having any significant issues associated with them, or at least, any significant issues associated with them that mainstream press has decided that it's important enough to tout. Almost all the Intel bugs (the F00F bug notwithstanding) have been discovered within months of their debut.
I run plenty of modules on my system that are not in the kernel. I thought compiling your own kernel was one of the best things ABOUT linux.
First, you're getting the kernel modules with the pluggable service model of the microkernels mixed up. In the Linux module world, all the modules run in kernel space, but in a microkernel, a fair bit of the services run in user space. Take, for example, mkLinux. The "hardware" that mkLinux runs on is the Mach Microkernel. All the hardware drivers and other system-specific bits and bobs have been rewritten to use the Mach services instead. Rip out the memory manager (because that's Mach's job, not the Linux service's), and you have an oversimplified view of mkLinux, where the Linux "service" runs in user space.
You can compile your own kernel in other unices too. Granted, you won't compile from the source, but both BSDI and SunOS 4 will let you compile the kernel mainly from object files, and all that. Most commercial unices now don't come with this feature out of the box, because it's an expensive-add-on-feature.
Axil computer, a now-defunct SPARC systems manufacturer, used to bake their systems in a 45C oven, and require them to run for a month without crashing. I would assume that Sun does something similar with their boxen. Afterall, they're still alive:-)
Sun is a corporation that sells hardware and software. If it pisses off enough of its customers (all within its legal rights), they will go away and stop buying highly overpriced Sun hardware.
And you get to spend the hideous amounts of money once again, in relicensing all your software. Think about it. Oracle and co will license you their software for only one platform at a time; to move over to say, an RS/6000, you would need to repurchase all the licenses. Plus, you get to port all your Sun-specific software to the said new platform, and after all that, you get to worry if the vendor's commitment to the platform will be ongoing.
At least with Sun, you know that their SPARC platform isn't going away anytime soon, unlike IBM, with their myopic and rather erratic shifts of policy regarding their multitude of platforms, HP, with the PA-RISC/Itanic, etc, and SGI, with, well, it's SGI. They may not be around in a few years....
Don't get me wrong, I like AIX 4.x and IRIX 6.4 is among the sweetest I've programmed for (and like it better than Solaris), but if I'm the CIO, you'd better believe it that I'm probably going to buy Sun kits. Then again, I'm not currently the CIO here. Hrm....:-)
It's all anecdotal evidence, but at my site, we have a number of E10k's, and other than the occassional kernel updates and custom drivers, our E10k's have been up almost 100%. Granted, software stability is another issue, but that's my problem:-).
I wonder if the p4 will really swoop in and take the lead at this rate... between rambus and 1.0 GHz problems, the pIII is looking pretty bad in middle age. The p4 will be like a sleek and solid dream machine by comparison.
As I've stated elsewhere, I have a different opinion. With all the problems Intel (and AMD too) have had with their latest cores, I believe the P4 will have at least 1, if not more, issues that will require massive engineering attention. Every new core, from the 486 to the P6, have had some issues associated with it, and I don't think the P4 will be any different.
Of course, I would like to see Intel absolutely shine in their P4 efforts, but then I'm biased; I have several friends who work for various Intel subcontractors that I don't want to see out in the streets =P
Intel is jamming FUD down the uninformed throats in an attempt to stem the blood loss until they can rush the P4 out the door.
With all the "blood loss", one would have to wonder if the P4 will be truly ready, or merely "ready enough" when it comes out.
I've grown a bit leery of anyone's "latest and greatest" processor cores; I will be buying Athlons and Lower-end PIII's for my servers for a bit to come, until other people have beta-tested the sledgehammers and P4's.
Slashdot Mirror
How is K-Mart requiring carding "demanding that the rest of the [society] pick up the slack"? I see it as a tool for enforcing my values and standards. How is a private company choosing who they sell product X to censorship? How is it "parenting"? Do you have any children of your own?
That's not what I said, and you know it. But since you bring up the Gay community, I will respond to that. Like it or not, they are a part of the society. It is up to us parents to teach tolerance and acceptance to those not like ourselves. This narrow-minded "you are wrong because you're different" thought pattern just teaches children to close off their minds. This is exactly what led to the hate-movement of the 1960's, and the hate riots in Germany during the 1990's.
My spouse and I are ultimately responsible for our children. We know this, and strive for excellence in everything we do with our children, teaching them our values. If I wanted my children to play "Soldier of Fortune", then I would be happy to buy it at K-Mart (or wherever), with my identification. K-Mart's carding is a tool, not censorship.
Go ahead and boycott K-Mart and Wal-Mart. I think they will do fine without your business.
--
What he's proposing is to create a SIG, not a political party. A SIG lobbies congressmen, and in turn, pays for a part of their campaign.
--
Heard on NPR, quoting Jesse Ventura, the governor of Minnesota. He stated, and quite correctly, that many media outlets are no longer in the business of reporting the news, but instead, making the news. They're slaves to the almighty buck.
He went on to state that there's nothing wrong with being a slave to the almighty buck, but if they are, they should not be misleading consumers by stating that they're "journalists" and not "tabloid artists".
On the web, as with anywhere else, "consider the source". Many people no longer do this, but this is absolutely critical when judging the accuracy of any information, not limited to that shown in the news. Sensationalism in "professional" journalism is very much alive today as it were in the 1890's, when "yellow" journalism coursed through the pages of "credible" newspapers.
--
It kind of shows that you don't have children of your own. I believe in privacy, both of myself and of "my" children. It works both ways--I don't go rummaging around in their stuff, and they don't go rummaging around in mine. I lead a very busy schedule (work 60+ hours/week), and so does my fiancee. We welcome all tools that will allow us to enforce our rules that we set, such as this K-Mart bit.
Seriously, there is no way that anyone can reasonably maintain a 24/7 surveillance on their children. Like it or not, this is what will be required to filter most of the social rubbish that will be imprinted on our children. Instead, let the society and the corporations help you in building the tools to monitor the products' influences on our children. This is precisely what K-Mart is doing; as the other posters have suggested, if I wanted my children to grow up playing violent video games, I will personally buy them myself. This new-age "society has no place in rearing my children" rubbish really sickens me; for thousands of years, small communities imprinted their values on the children. This indoctrination still happens daily at schools. I guess that your children will be homeschooled as well?
If you don't like it, fine. Boycott K-Mart and Wal-Mart. That is your right as a consumer. However, when you have your own children, and they're at the age where they want to play video games, even violent ones, and are playing them in your living room, at least have the decency to listen to us when we say, "We told you so.".
--
Remember the fuel price drop of the 1980's?
The SUV's and the trucks that're quite popular (check out the JD Power's stats) aren't exactly frugal in their petrol expenditures, you know....
Do you want to twist statistics? GM's earnings were *down* this year compared to the last. Toyota is still growing. GM had net revenues of approx. $5 billion, Toyota $4.5 billion. Looking at the rate of growth in both companies, Toyota will surpass GM in net earnings by 2005. Easily.
What does this information actually mean? Absolutely nothing! Why? because both indicators are completely dependent upon the trade agreements and such between the US and other countries, as well as buyer trends and such.
All we managed to prove is that the large companies in the automotive business are rather resilient, and that with deep enough pockets, it's possible to ride out the storm.
--
Just to reply to a troll....
Granted, it wasn't a bunch of Linux boxen, but at one of the places I've worked at, we had 500+ Sun boxen running various versions of SunOS, 200+ combined HP and IBM workstations, 150 or so Macs, mainly in the marketing, plus about 100 (or thereabouts) Windows boxen.
This was a few years ago, and since then, they've been bought out by Cadence, so I don't know what's going on there now.
--
Sorry, I parsed your previous "error in the config" as "error in the configuration file", and not an "error in the hardware configuration as the program currently sees it".
In the case you just described, yes, it was rather idiotic, either the driver or the userland program should've acknowledged the problem and did *something* other than just blanking out.
--
You're forgetting that both IBM and HP are rather large companies with rather large support services divisions. Take, for example, IBM. IBM Global Services, their consultancy group, will support anything from CP/M to IBM SP2 (and probably Sun E10k) to OS/390 and AS/400 mainframes. Same with HP. They tout themselves as a one-stop support center for everything. Large companies pay huge amounts of money to these organisations for support.
So why are they bulling ahead with this certification lab? It makes supporting these applications (yes, they support applications too) easier if they know the ins and outs of the app. They'll of course find this out while they "certify" the piece of software.
Currently, because of the lack of certification, they will support any app, even the ones that I built at home while drinking wine, and is this unmaintainable clusterf*ck, if it made them some dough. Suppose they certify someone else's OSS project, which does essentially the same thing, except missing features x and y. Will they support my app for the company? Hell no. They will tell them to move to the someone else's app, because it's actually possible to maintain the bloody thing.
See Caldera? See IBM? See HP? See SCO? Hell, see MS? They're all trying to move to a constant revenue model that comes from providing a service, not maintaining the one-time revenue of a product.
--
ITYM, "IANAP" (I am not a programmer). As anyone with even a small grounding in language design will tell you, Syntactic errors are bloody easy to detect. Just look at all the typos that your compiler will spit out. Semantic errors, however, are another story. Bloody things are almost impossible to detect at the machine level.
Just about any monkey (or a Turing Machine, whichever) can find that something is syntactically amiss. Tools like lex and yacc make it rather simple to build even the most complicated of configuration scripts. However, validating for semantic correctness in even a moderately complex configuration script is another matter; it is multiple orders of magnitude more difficult and computationally intensive to verify this (and for some languages, like English, whose synctactic definitions certainly do exist, the technology for verifying semantic correctness doesn't yet exist).
--
Assuming that you're misinformed rather than being a troll, SSL needs the signature to prevent a man-in-the-middle attacks, or any hosts masquerading as another host. Think about it. You're going to amazon.com to buy books. I poisoned your DNS so that www.amazon.com points to my domain, and I act as a proxy for the real amazon.com until you're ready to make a purchase. When you do, through my SSL server, without the certificate, and I log your credit card information for my future use.
With certificates, you can be reasonably sure that the www.amazon.com is the real amazon.com and not my little false fpos thing. Without it, well, it opens things up for all sorts of interesting attacks.
--
I agree completely. You have to take a proactive measure to security, regardless of what system you run, be it Linux, *BSD, Winblows, whatever. Then, someone will counter, "but I have to install N machines, same (or very similar) configuration for the foo application!".
Well, son, you have several choices.
I did this at one of my clients' sites, where we had over 600 SPARC boxen. Of course, took it one step further, by providing an install server, but then, this was also in a very protected subnet.
Do it only once, but for Seldon's sake, do it right!
--
Thank you. I didn't know that IPv6 is actually in use right now
--
Looking at the current design and usage patterns of 8-10lane highways popular in the US, there's an upper limit of what can be utilised. Sure, if a person is travelling more than a few km, the person will use the entire highway. However, at major exits and such, the bottleneck isn't just the highway itself, but also the lower-capacity roads below. Frequently, I've found myself moving along at ~120km/h, while the people taking the exit ramp (or the entry ramp to another, clogged, highway) were bogged down for literally hundreds of car-lengths.
Unfortunately, I don't have a solution (Actually, I do--while living in central Tokyo, I commuted on off-peak hours on the train--a solution unworkable in most American metroplitan areas), other than abusing the flex hours.
Any ideas?
--
Sorry, not to be pedantic, but what you just described is 5 529mbps, far in excess of what you just described with your DSL saturation issue.
I think the main issue with your "inability" to saturate your DSL is an issue of bottlenecks; with an analogue modem connection, your modem is the bottleneck, with your DSL, some upstream provider somewhere is.
--
The question is, will they ever "roll it out" to beyond what it is now? I mean, sure, they use IPv6, sure, their backbones are probably an order of magnitude fatter on a per-host basis, but would "they" ever roll it out, or would the current IPv4-based Internet just migrate to IPv6 when the specs are "done", tunnelling some legacy IPv4-based traffic in a "4-bone", or doing some sort of weird IPv6-IPv4 NAT? Or will the current IPv4-based Internet plod on, NAT-ting everywhere (dear lord, I hope not)?
Then again, when I run traceroute(1) everywhere, I almost always see a 10.x.y.z somewhere :-)
--
No, but there were looser (read: no) restrictions on who can be on the network. I2 has a rather stringent policy on who can be on and who can't. Please read the FAQ :-).
--
Uh.... I think I'm missing something here. First, Internet2 is (was?) a research-only network. It's not (supposed to be) a place for pr0n and other commercial use.
You speak of standards. The Internet, as we see it, is just a collection of heterogenious networks in a homogeneous naming space (well, mostly--a discussion of NAT is beyond the scope of this comment). You want to move to I2? Fine. Go enroll as student or faculty at one of 180 research institutions. Just don't go there for much Quake use :-)
You also take a look at kermit. Kermit was not an Internet protocol per se, but was a terminal-terminal download protocol that assumes unreliable network streams. Yes, that means certain instances of kermit-over-IP existed, but it doesn't mean that kermit was necessarily an Internet application-level protocol, where most transfer tools assume a TCP-level (or equivalent) functionality, leaving error detection and correction (among other things) to the lower-level protocols, and it dealing more with the application-specific information.
As for benchmarking /. code, I really don't see the difference of its code running on I2, Internet, or my home LAN; provided that the host is mated to a decent backbone, its performance would be more dependent upon load/system configuration than the underlying network itself. I2, with its much smaller user base and a similar backbone, consequently has much lower load, so it should perform better.
Besides, if you wanted to compare/contrast I2 to Internet, you'd probably do a network traffic analysis (peak bandwith, peak latency, multiplexing capabilities, etc), which would be a function of the routers than anything else.
--
Athlons have, IMO, proved themselves in the field by 1) being out in the "wild" for over a year, and 2) not having any significant issues associated with them, or at least, any significant issues associated with them that mainstream press has decided that it's important enough to tout. Almost all the Intel bugs (the F00F bug notwithstanding) have been discovered within months of their debut.
--
First, you're getting the kernel modules with the pluggable service model of the microkernels mixed up. In the Linux module world, all the modules run in kernel space, but in a microkernel, a fair bit of the services run in user space. Take, for example, mkLinux. The "hardware" that mkLinux runs on is the Mach Microkernel. All the hardware drivers and other system-specific bits and bobs have been rewritten to use the Mach services instead. Rip out the memory manager (because that's Mach's job, not the Linux service's), and you have an oversimplified view of mkLinux, where the Linux "service" runs in user space.
You can compile your own kernel in other unices too. Granted, you won't compile from the source, but both BSDI and SunOS 4 will let you compile the kernel mainly from object files, and all that. Most commercial unices now don't come with this feature out of the box, because it's an expensive-add-on-feature.
--
No wonder the bloody site is down all the time! It's running on Win98! :-)
--
Axil computer, a now-defunct SPARC systems manufacturer, used to bake their systems in a 45C oven, and require them to run for a month without crashing. I would assume that Sun does something similar with their boxen. Afterall, they're still alive :-)
--
Flammable. Treat with care.
And you get to spend the hideous amounts of money once again, in relicensing all your software. Think about it. Oracle and co will license you their software for only one platform at a time; to move over to say, an RS/6000, you would need to repurchase all the licenses. Plus, you get to port all your Sun-specific software to the said new platform, and after all that, you get to worry if the vendor's commitment to the platform will be ongoing.
At least with Sun, you know that their SPARC platform isn't going away anytime soon, unlike IBM, with their myopic and rather erratic shifts of policy regarding their multitude of platforms, HP, with the PA-RISC/Itanic, etc, and SGI, with, well, it's SGI. They may not be around in a few years....
Don't get me wrong, I like AIX 4.x and IRIX 6.4 is among the sweetest I've programmed for (and like it better than Solaris), but if I'm the CIO, you'd better believe it that I'm probably going to buy Sun kits. Then again, I'm not currently the CIO here. Hrm.... :-)
--
It's all anecdotal evidence, but at my site, we have a number of E10k's, and other than the occassional kernel updates and custom drivers, our E10k's have been up almost 100%. Granted, software stability is another issue, but that's my problem :-).
--
As I've stated elsewhere, I have a different opinion. With all the problems Intel (and AMD too) have had with their latest cores, I believe the P4 will have at least 1, if not more, issues that will require massive engineering attention. Every new core, from the 486 to the P6, have had some issues associated with it, and I don't think the P4 will be any different.
Of course, I would like to see Intel absolutely shine in their P4 efforts, but then I'm biased; I have several friends who work for various Intel subcontractors that I don't want to see out in the streets =P
--
With all the "blood loss", one would have to wonder if the P4 will be truly ready, or merely "ready enough" when it comes out.
I've grown a bit leery of anyone's "latest and greatest" processor cores; I will be buying Athlons and Lower-end PIII's for my servers for a bit to come, until other people have beta-tested the sledgehammers and P4's.
--