Yeah, I've been meaning to restore that particular vserver for, like, a year now. This is what happens when there are better things to do than maintain a personal web site.:-)
The site is a mess with all sorts of useability and performance problems. It would be nice if someone just setup a good new framework for it, and then "imported" everyone's crap into it.
If you wrote such a system and offered it to myspace, they'd probably pay big bucks for it. But it would have to handle more traffic than even Slashdot does, do it well, and manage tens of terabytes of data without falling over. And that's just for the user profiles... Then you've got the groups, the music, the mail system,...
I love listening to people who have never built a massive web site before saying how "someone should just do it right". When you build a system and it explodes 1000x faster than you could possibly predict, it's hard to keep up with. Even when you build it correctly and manage to anticipate your traffic loads, serving that number of complex pages is no laughing matter and causes a lot of people to work a LOT of long hours.
On the first paragraph, I highly doubt that the software controlling my fuel injection, engine timing, and numerous other things is just "a few thousand lines of code." Car computers tend to be fairly complex beasts (and, perhaps unfortunately, some of them even run Windows Embedded!). There is an operating system, it's just an RTOS instead of a general purpose OS. The timing on these things is absolutely critical, too, which means much more required accuracy. You have a point on 3rd party drivers and user interaction, though, and that IS a huge thing.
As far as the latter, the more testing I think is the point. It's not nearly impossible to develop stable GUI apps -- nobody bothers though, because they don't want to spend the time testing.
--S (guilty as charged -- I occasionally develop ugly GUI apps too;-)
I see this kind of thinking a lot, but it all falls apart when you suddenly realize that cars, toasters, and ovens actually *contain* software with potentially millions of lines of code...
Somehow, they still seem to work much better than your average piece of "consumer software".
Of course, right after computer technology goes that way, someone will implement this new "file system" concept for this big ugly hunk of unmanageable memory...:-)
Tell that to the enterprise-class SCSI drives *I* work with. Quality Control my ass. (And no, I'm not talking about some third-rate storage vendor here -- thank dog for support contracts, we go through WAY too many drives for my liking).
We'll see if the 15KRPM models are any better than the 10KRPM models.
Windows is only as unstable as the administrator that's managing it.
That said, many Windows administrators are notoriously unstable due to the low barrier to entry. That's one of the main reasons I prefer Linux on my servers for most applications.:-)
And if you don't have expensive monitor calibrator, you'll get consistently bad prints every time:-)
More to the point, printer/monitor calibration tools aren't that expensive these days. You can probably get both for around $300 (which is more than worth it if you're that concerned about print quality, even if you're NOT printing your own photos).
OpenVMS is likely a vastly more complicated operating system than Microsoft Windows ever will be.
It includes, among other things: support for privilege-based security as well as more traditional methods, support for indexed databases as a standard feature of the filesystem, full fault-tolerant clustering, and probably a zillion things I can't think off offhand. And oh yeah, it even has X11, aka DecWindows (though we called it DeathWindows, 'cause it was so damn slow on that old crappy hardware).
VMS is not the same as WindowsNT, but it is almost certainly more complex.
Ironically, lots of the NT architects were old VMS guys, too, from what I've heard. Since they've built a highly reliable system once (despite its problems, it's still far more reliable and secure than Windows, and always has been), I have to question whether the problems in Windows NT are the result of corporate culture rather than programmer ability.
Every time this discussion comes up, all I hear is "think of the poor programmers!". If you want to cry like a baby every time someone suggests that you can do better or that you should be held accountable for your work, then IMNSHO you don't belong in this business.
The fact is, this industry is built on the ability to ship crap-quality software specifically because they can get away with it. Reliable, high-quality software and hardware (from operating systems to major enterprise-class databases to whatever else you want to think about) has existed for a LONG time. Just ask IBM. It's only expensive because so few companies care enough to produce quality software outside the embedded arena.
As for the space program, that's apples and oranges. Participation in the space program is an Elite Activity, which means that there are very few who actually do it. There are a whole lot more minds working on software on a daily basis. The fact is, software companies are just too flipping lazy and/or cheap to fix the problems.
Personally, I think that software houses should be liable in the case of gross negligence. If they cost someone a few million dollars because their software crashes, then they should go to court. If they're found to not have appropriate quality control at the time the software was written, they damn well *should* be liable for producing a flawed product.
The only thing I see that will keep this from ever happening is that not enough non-geek people know anything about software engineering. That will make it damned hard for the judicial system to determine what "appropriate quality control" is, exactly.
Besides, why should I be forced to suffer through non-quality software (and make no mistake, we *are* forced to as low-end consumers, simply by lack of available choice. It's sad when free operating systems work more reliably than commercial ones)? If Kenmore was putting out refridgerators that randomly quit every few days or under the wrong phase of the moon, I doubt they'd sell many refridgerators.
Too bad that the software industry managed to convince consumers that it's somehow immune from common-sense quality standards.
Ah, but how much more often would you turn off your desktop if it were guaranteed to be as you left it when you turned it back on -- and there was no delay in getting to that state?
Think of the power savings if computers only had to be on when we were actually *used* them...:-)
This is definitely a better hook for Serentiy than the ads they had on TV. If it weren't for my supremely twisted sense of humor, I would never have gone to see it. When I *did* go see it, I thought it would be pretty lame, but I needed to get out of the house for a while. Boy, was I surprised.
Their marketing really sucked, to be honest. I'm shocked that it's done as well as it has, and I'm disappointed that the marketers left it in that position, 'cause it deserved much better.
Do an ROI on multi-homed setups.... we did.... I helped many companies do the same.... the ONLY time the returns are truly favorable is in exactly this and ONLY this situation. Every other possibility shows that waiting out outages and rerouting is best for the bottom line for a "normal" non-MS company. Period. To tell people they need to pay for multi-homed access from the very beginning to avoid this type of problem is insane - they just don't NEED it.
You seem to be calling me full of shit for something I never said. How interesting. And how utterly expected on Slashdot.
The part you missed was where I mentioned 'acceptable risk.' Most small web sites can give two hoots if they go down for a few hours (or in many cases even a few days).
On the other hand, if someone's bottom line is tied to site accessibility-related customer satisfaction (aka, any truly major web site) and you're showing them an ROI study that says they don't need it, you're screwing your customers and are completely unqualified to be providing said study.
You missed the point. I'm amazed you got an "informative" (though that's no fault of yours;-).
The point is that non-L3 customers should be able to hit Cogent via other routes, and in many cases can't. Which means that Cogent's other connectivity sucks the business end of a donkey.
Slashdot Mirror
Yeah, I've been meaning to restore that particular vserver for, like, a year now. This is what happens when there are better things to do than maintain a personal web site. :-)
--S
This is a panacea for MySpace. Idiots comprise the largest and most profitable segment that capitalist companies can market to. :-)
--S
If you wrote such a system and offered it to myspace, they'd probably pay big bucks for it. But it would have to handle more traffic than even Slashdot does, do it well, and manage tens of terabytes of data without falling over. And that's just for the user profiles... Then you've got the groups, the music, the mail system,
I love listening to people who have never built a massive web site before saying how "someone should just do it right". When you build a system and it explodes 1000x faster than you could possibly predict, it's hard to keep up with. Even when you build it correctly and manage to anticipate your traffic loads, serving that number of complex pages is no laughing matter and causes a lot of people to work a LOT of long hours.
--S (yes, I've done this before.)
So true. I mean, after all, it's so much cooler to use ^H ^Q^H^H^W -- oh, screw it...
--S
On the first paragraph, I highly doubt that the software controlling my fuel injection, engine timing, and numerous other things is just "a few thousand lines of code." Car computers tend to be fairly complex beasts (and, perhaps unfortunately, some of them even run Windows Embedded!). There is an operating system, it's just an RTOS instead of a general purpose OS. The timing on these things is absolutely critical, too, which means much more required accuracy. You have a point on 3rd party drivers and user interaction, though, and that IS a huge thing.
;-)
As far as the latter, the more testing I think is the point. It's not nearly impossible to develop stable GUI apps -- nobody bothers though, because they don't want to spend the time testing.
--S (guilty as charged -- I occasionally develop ugly GUI apps too
I see this kind of thinking a lot, but it all falls apart when you suddenly realize that cars, toasters, and ovens actually *contain* software with potentially millions of lines of code...
Somehow, they still seem to work much better than your average piece of "consumer software".
--S
Or it could be that many of us have real-world experience with performance-sensitive business applications. They happen.
Mind you, I happen to like C# and Java. Just not for everything.
--S
Of course, right after computer technology goes that way, someone will implement this new "file system" concept for this big ugly hunk of unmanageable memory... :-)
--S
Tell that to the enterprise-class SCSI drives *I* work with. Quality Control my ass. (And no, I'm not talking about some third-rate storage vendor here -- thank dog for support contracts, we go through WAY too many drives for my liking).
We'll see if the 15KRPM models are any better than the 10KRPM models.
I'm not holding my breath.
--S
I can show you such a device.
:-)
There's no rule that says you can't 1:1 NAT everything on your network so that it's accessible from IP's on a different netblock...
Sadly, there are probably networks out there that are configured that way, too
--S
OSS: The only community in which you can get roasted for trying to play nice when people ask you to.
Windows is only as unstable as the administrator that's managing it.
:-)
That said, many Windows administrators are notoriously unstable due to the low barrier to entry. That's one of the main reasons I prefer Linux on my servers for most applications.
--S
On the other hand, the commonly accepted defacto standard is that RAM is volatile. That's why someone came up with the extended acronym NVRAM...
--S
People would probably pay a lot of money for the guaranteed ability to fart super silently, however.
--S
Interestingly, the domain contact's email address is at transmx.com. The web server at www.transmx.com redirects to... ...www.transmx.com.
I am uninspired by the technical prowess of TransMedia's IT department.
--S
And if you don't have expensive monitor calibrator, you'll get consistently bad prints every time :-)
More to the point, printer/monitor calibration tools aren't that expensive these days. You can probably get both for around $300 (which is more than worth it if you're that concerned about print quality, even if you're NOT printing your own photos).
--S
Bottles of Dom Perignon don't come with brand new electronic printheads attached.
:-)
Not that I think that justifies current cartridge prices (I remember the past!), but it's still worth pointing out.
--S
errrr....
OpenVMS is likely a vastly more complicated operating system than Microsoft Windows ever will be.
It includes, among other things: support for privilege-based security as well as more traditional methods, support for indexed databases as a standard feature of the filesystem, full fault-tolerant clustering, and probably a zillion things I can't think off offhand. And oh yeah, it even has X11, aka DecWindows (though we called it DeathWindows, 'cause it was so damn slow on that old crappy hardware).
VMS is not the same as WindowsNT, but it is almost certainly more complex.
Ironically, lots of the NT architects were old VMS guys, too, from what I've heard. Since they've built a highly reliable system once (despite its problems, it's still far more reliable and secure than Windows, and always has been), I have to question whether the problems in Windows NT are the result of corporate culture rather than programmer ability.
--S
Every time this discussion comes up, all I hear is "think of the poor programmers!". If you want to cry like a baby every time someone suggests that you can do better or that you should be held accountable for your work, then IMNSHO you don't belong in this business.
The fact is, this industry is built on the ability to ship crap-quality software specifically because they can get away with it. Reliable, high-quality software and hardware (from operating systems to major enterprise-class databases to whatever else you want to think about) has existed for a LONG time. Just ask IBM. It's only expensive because so few companies care enough to produce quality software outside the embedded arena.
As for the space program, that's apples and oranges. Participation in the space program is an Elite Activity, which means that there are very few who actually do it. There are a whole lot more minds working on software on a daily basis. The fact is, software companies are just too flipping lazy and/or cheap to fix the problems.
Personally, I think that software houses should be liable in the case of gross negligence. If they cost someone a few million dollars because their software crashes, then they should go to court. If they're found to not have appropriate quality control at the time the software was written, they damn well *should* be liable for producing a flawed product.
The only thing I see that will keep this from ever happening is that not enough non-geek people know anything about software engineering. That will make it damned hard for the judicial system to determine what "appropriate quality control" is, exactly.
Besides, why should I be forced to suffer through non-quality software (and make no mistake, we *are* forced to as low-end consumers, simply by lack of available choice. It's sad when free operating systems work more reliably than commercial ones)? If Kenmore was putting out refridgerators that randomly quit every few days or under the wrong phase of the moon, I doubt they'd sell many refridgerators.
Too bad that the software industry managed to convince consumers that it's somehow immune from common-sense quality standards.
--S
Ah, but how much more often would you turn off your desktop if it were guaranteed to be as you left it when you turned it back on -- and there was no delay in getting to that state?
:-)
Think of the power savings if computers only had to be on when we were actually *used* them...
--S
Right, because if lots of people like it, they must all be sheeple that are liking it because they're told to.
Yah. Sure. You betcha.
--S
This is definitely a better hook for Serentiy than the ads they had on TV. If it weren't for my supremely twisted sense of humor, I would never have gone to see it. When I *did* go see it, I thought it would be pretty lame, but I needed to get out of the house for a while. Boy, was I surprised.
Their marketing really sucked, to be honest. I'm shocked that it's done as well as it has, and I'm disappointed that the marketers left it in that position, 'cause it deserved much better.
--S
For another, I could watch it without being bored out of my skull.
You seem to be calling me full of shit for something I never said. How interesting. And how utterly expected on Slashdot.
The part you missed was where I mentioned 'acceptable risk.' Most small web sites can give two hoots if they go down for a few hours (or in many cases even a few days).
On the other hand, if someone's bottom line is tied to site accessibility-related customer satisfaction (aka, any truly major web site) and you're showing them an ROI study that says they don't need it, you're screwing your customers and are completely unqualified to be providing said study.
--S
You missed the point. I'm amazed you got an "informative" (though that's no fault of yours ;-).
The point is that non-L3 customers should be able to hit Cogent via other routes, and in many cases can't. Which means that Cogent's other connectivity sucks the business end of a donkey.
--S