I'm not sure that I have better knowledge, but I do know that you've been able to do this for a while. As you said, I think it's been since loopback mounts were supported.
My theory is that the swap partition is faster, since you are cutting out the middleman of the normal ext2fs/ext3fs/whatever layer, and going straight to disc from the kernel VM layer. But that's just my theory; it could be the case that using a swap file is just as fast. Windows uses a swap file IIRC.
The problem with doing the automatic partitioning is that the user who's going to be the most interested in desktop Linux is the same person who is most likely to need to dual-boot with Windows for the near future. Dual-booting with Windows means that you have to mess with partitions, which means that not needing a swap partition is a little bit of a bonus in terms of installability. Or that's my theory, at least.
Do you have to work hard to spell that poorly, or does it come naturally? I mean, come on! I know you're a troll, but try to at least make it a respectable troll, all right?
Maybe the fact that you leave the side open had more to do with the innards getting hot? It doesn't help if you have ventilation and fans, if the case is open, since that decreases the forced air action that you get with a closed case.
Except that BSA audits are motivated by fear, not by logic. Get enough senior executives afraid of the expense of being audited and the potential for fines if they are not in line on licensing, and they'll prefer to throw money at the problem instead. This is just another item for the BSA to use to create a fear effect in the business community. If you make the licensing prohibitively complex like this, then pretty much everything you do will break some part of your license(s), and you can be audited at will.
The funny thing is that, a few years back, you would be defending them by saying that Microsoft doesn't make any media programs, they just provide APIs for other media software companies to use.
Wake up - the only reason that a particular Windows software niche hasn't been locked up by Microsoft is because they haven't wanted or needed to...yet. The fact that they haven't yet taken over the whole ball of wax doesn't mean that that is not their ambition, as evidenced by their previous actions.
Well, it's not like they were impartial before, right? Now they just have an opportunity to be partial in the other direction. Kind of mirroring Anakin's fall to the Dark Side, if you will:)
I agree with some of the other comments about how the fact that you've ended up in such a situation probably means that some things weren't planned right to begin with.
However, since you've still got to dig out of the hole that you're currently in, I think you're sort-of on the right track. The problem with bringing in a bunch of changes at once is twofold: you have to do the merges properly, and you also have to ensure that bringing in developer A's changes doesn't break what developer B just did.
To get the merges done properly, your best bet is to stick with your current air traffic control strategy, but make sure that you have a controller who runs the show and does the merges. You need someone experienced enough that they can merge most people's code without breaking it, even if they don't quite understand all of how it works. Devoting a person full-time or almost full-time to merging at crunch time would mean that developers can focus more fully on making the right fixes, testing them, and making them easier for you to merge.
As far as verifying that your merges haven't broken something else, probably your best bet is to test rigorously with each new release via automated tests. This is going to be a pain to set up right now if you haven't planned for this, but there's really no other good solution. You can also try to enforce peer review for changes, on the hopes that developer B will notice that developer A's changes will break something that B just fixed, but peer review doesn't always catch everything, especially at crunch time.
I think you've pointed out an interesting failure of current SCM tools - they make it fairly easy to keep things organized, but they trade off speed for that organizing power. I don't think we have very good tools yet that will allow you both speed and organizational power.
If it's the AP feed that they're all picking up, then sure, it's the same story.
What's interesting is when the story is different - if CNN says one thing, the BBC says something a little different, and the Times of India has yet a third viewpoint, then it's interesting to speculate on the editorial biases that are leading to such divergent viewpoints.
Yeah, I'd love to be able to play Wing Commander again - there's a golden opportunity for someone to write a simple DOS app that allows you to adjust the timing of a program that it runs so that things happen at the right speed. Because I don't see the original publisher re-releasing the game using better timing code, unfortunately.
Although we are using decimal currency, which was also Jefferson's idea - apparently decimal money is more of an advantage than other decimal measurements.
I can confirm that all science in the U.S. is done in the metric system - kilograms, liters, etc. Having experienced both, I prefer the metric system, but I don't see that the American populace as a whole will be ready to switch any time soon.
Um, no there aren't. They just have to target 80386+ glibc2 systems and maybe the latest PPC machines. There aren't any proprietary software companies that support every single Linux configuration; it's silly to say that somehow AOL would have to. They could just check the system libraries, etc. before installing, and refuse to install if the system is deficient. Many proprietary products on Linux already do this - ClearCASE 4.1, for example. Heck, it's not like the latest AOL will work on Windows 3.1 any more, so why should we expect it to work on a libc4 or libc5 Linux system?
A better reason is that the market for the easy-to-use Internet on Linux systems is currently vanishingly small. Maybe if there was a big push into Linux-based "web appliances", it would be worthwhile for AOL, but the last time they tried that it wasn't an astounding success.
No company is going to leave its web site unavailable to 34 million people, not when they can make the site standards-compliant and work in both IE and AOL. Netscape 4.x is a different matter, granted - you almost had to have two copies of your site for a while. But if AOL uses Gecko, then the web's back to only one site for everyone, and all of us using non-IE browsers will benefit from that. AOL's 34 million newbies could be the best thing that's happened to the web in the last couple years (betcha never saw that coming:).
Good point. I was assuming that if you were Microsoft and you didn't want to let on that you were using open-source code, you wouldn't just leave a zlib.dll laying around. So I was assuming dynamic linking for open-source projects (not necessarily true) and static linking for closed-source projects (also not necessarily true). My mistake.
Trust me - if he was really trying to blow up the airplane, would he try to go through security with all of those wires, electronics, etc? The whole point of being a terrorist is getting by security, not getting hassled by them.
Slashdot Mirror
What, so you can get it first? I don't think so! Hahahaha...
Actually, the apocalypse is when he quits saying that, and actually releases the damn thing :)
I'm not sure that I have better knowledge, but I do know that you've been able to do this for a while. As you said, I think it's been since loopback mounts were supported.
My theory is that the swap partition is faster, since you are cutting out the middleman of the normal ext2fs/ext3fs/whatever layer, and going straight to disc from the kernel VM layer. But that's just my theory; it could be the case that using a swap file is just as fast. Windows uses a swap file IIRC.
The problem with doing the automatic partitioning is that the user who's going to be the most interested in desktop Linux is the same person who is most likely to need to dual-boot with Windows for the near future. Dual-booting with Windows means that you have to mess with partitions, which means that not needing a swap partition is a little bit of a bonus in terms of installability. Or that's my theory, at least.
Wow, you know they've plummeted when they've reached the same moral level as GraceNote :)
[as seen on a board around here]
The partner at Anderson said "Ship the Enron documents to the Feds." But you heard "Rip the Enron documents to shreds."
Turns out it was just a case of bad cellular.
Sprint PCS, etc., etc. :)
Do you have to work hard to spell that poorly, or does it come naturally? I mean, come on! I know you're a troll, but try to at least make it a respectable troll, all right?
Maybe the fact that you leave the side open had more to do with the innards getting hot? It doesn't help if you have ventilation and fans, if the case is open, since that decreases the forced air action that you get with a closed case.
Or at least that's what I'd always heard.
Spoken like someone who's never seen sendmail machines swapped out for Exchange servers :)
Except that BSA audits are motivated by fear, not by logic. Get enough senior executives afraid of the expense of being audited and the potential for fines if they are not in line on licensing, and they'll prefer to throw money at the problem instead. This is just another item for the BSA to use to create a fear effect in the business community. If you make the licensing prohibitively complex like this, then pretty much everything you do will break some part of your license(s), and you can be audited at will.
The funny thing is that, a few years back, you would be defending them by saying that Microsoft doesn't make any media programs, they just provide APIs for other media software companies to use.
Wake up - the only reason that a particular Windows software niche hasn't been locked up by Microsoft is because they haven't wanted or needed to...yet. The fact that they haven't yet taken over the whole ball of wax doesn't mean that that is not their ambition, as evidenced by their previous actions.
I haven't seen any of his actions that I would qualify as "dancing". Maybe bounced, bounded, thundered, or berzerked, but not "danced".
Well, it's not like they were impartial before, right? Now they just have an opportunity to be partial in the other direction. Kind of mirroring Anakin's fall to the Dark Side, if you will :)
I agree with some of the other comments about how the fact that you've ended up in such a situation probably means that some things weren't planned right to begin with.
However, since you've still got to dig out of the hole that you're currently in, I think you're sort-of on the right track. The problem with bringing in a bunch of changes at once is twofold: you have to do the merges properly, and you also have to ensure that bringing in developer A's changes doesn't break what developer B just did.
To get the merges done properly, your best bet is to stick with your current air traffic control strategy, but make sure that you have a controller who runs the show and does the merges. You need someone experienced enough that they can merge most people's code without breaking it, even if they don't quite understand all of how it works. Devoting a person full-time or almost full-time to merging at crunch time would mean that developers can focus more fully on making the right fixes, testing them, and making them easier for you to merge.
As far as verifying that your merges haven't broken something else, probably your best bet is to test rigorously with each new release via automated tests. This is going to be a pain to set up right now if you haven't planned for this, but there's really no other good solution. You can also try to enforce peer review for changes, on the hopes that developer B will notice that developer A's changes will break something that B just fixed, but peer review doesn't always catch everything, especially at crunch time.
I think you've pointed out an interesting failure of current SCM tools - they make it fairly easy to keep things organized, but they trade off speed for that organizing power. I don't think we have very good tools yet that will allow you both speed and organizational power.
The problem is that nature can always make a crappier programmer, but our supply of smart debugger-writers is finite.
Like my pointers? They're made out of MONEY!!
Wow, I think this is the first time I've ever seen an AC say something nice about Bruce Perens. I'm in shock!
More people (in this forum, at least) would recognize that if QuickTime (Sorenson et al.) worked properly on Linux.
If it's the AP feed that they're all picking up, then sure, it's the same story.
What's interesting is when the story is different - if CNN says one thing, the BBC says something a little different, and the Times of India has yet a third viewpoint, then it's interesting to speculate on the editorial biases that are leading to such divergent viewpoints.
Uh oh, the wear indicators on the tracks are showing, time to rotate the disks :)
Yeah, I'd love to be able to play Wing Commander again - there's a golden opportunity for someone to write a simple DOS app that allows you to adjust the timing of a program that it runs so that things happen at the right speed. Because I don't see the original publisher re-releasing the game using better timing code, unfortunately.
Although we are using decimal currency, which was also Jefferson's idea - apparently decimal money is more of an advantage than other decimal measurements.
I can confirm that all science in the U.S. is done in the metric system - kilograms, liters, etc. Having experienced both, I prefer the metric system, but I don't see that the American populace as a whole will be ready to switch any time soon.
Um, no there aren't. They just have to target 80386+ glibc2 systems and maybe the latest PPC machines. There aren't any proprietary software companies that support every single Linux configuration; it's silly to say that somehow AOL would have to. They could just check the system libraries, etc. before installing, and refuse to install if the system is deficient. Many proprietary products on Linux already do this - ClearCASE 4.1, for example. Heck, it's not like the latest AOL will work on Windows 3.1 any more, so why should we expect it to work on a libc4 or libc5 Linux system?
A better reason is that the market for the easy-to-use Internet on Linux systems is currently vanishingly small. Maybe if there was a big push into Linux-based "web appliances", it would be worthwhile for AOL, but the last time they tried that it wasn't an astounding success.
The answer is: Market Share
No company is going to leave its web site unavailable to 34 million people, not when they can make the site standards-compliant and work in both IE and AOL. Netscape 4.x is a different matter, granted - you almost had to have two copies of your site for a while. But if AOL uses Gecko, then the web's back to only one site for everyone, and all of us using non-IE browsers will benefit from that. AOL's 34 million newbies could be the best thing that's happened to the web in the last couple years (betcha never saw that coming :).
Good point. I was assuming that if you were Microsoft and you didn't want to let on that you were using open-source code, you wouldn't just leave a zlib.dll laying around. So I was assuming dynamic linking for open-source projects (not necessarily true) and static linking for closed-source projects (also not necessarily true). My mistake.
Trust me - if he was really trying to blow up the airplane, would he try to go through security with all of those wires, electronics, etc? The whole point of being a terrorist is getting by security, not getting hassled by them.