Hell, you can have completely duplicate directory tree for these "Linux base" libraries if you want to hang on to what the distro already had. That would make sense as this stable base would probably never be as up to date as some of the distros.
Some things have their own libraries anyway, as it's the only sane way to support all the distros out there, but it would be a big win if applications could share these. And if they have nothing to do with the distro they're installed on, you could install these on Debian-stable or some old version of Slackware if you wanted. They'd just need kernel support for what they want to do.
"You really *do* know how to properly update your config files"
Yes. I never got breakage because of this.
"and you *aren't* running ~arch, are you?"
No. The warnings in the docs about this were quite clear.
The worst stuff was things like "stable" packages that had masked dependencies (eg famd and KDE, I forget which versions), and USE flags that change their meaning without any warning (Xinerama went from default-on from default-off).
I guess it just depends on your timing, but that's the point. I'm not comfortable with a system that has no assurances in place to make sure the "stable" branch is buildable at all times. People said it had improved when I tried it again late last year, but I had exactly the same issues, so I've pretty much given up on Gentoo for the forseeable future.
I don't hate Gentoo guys for bringing Portage to BSD. I dislike Gentoo zealots because they're really annoying, and I dislike Gentoo because it breaks itself without my help.
It only has half of the BSD philosophy. Good docs by Linux standards, and a centralized mechanism to install most 3rd party software, but terrible reliability.
The BSD philosophy includes a distaste for unreliability and breakage. We'd rather be a bit slower and a bit out of date than bleeding edge if it means we don't have to go around tracking down breakage.
I've tried Gentoo, and I loved how easy it was to install the most up to date software. However, I took a look at myself after a month or two and realized I was spending a lot of time cruising the formus, trying to find out why my system broke itself in the last update. I couldn't assume my system would be up and running at any given time, and I simply wasn't willing to put in the time it takes to keep that from happening.
I still need a Linux for various reasons, and I'd still like it to be up to date-ish. But Gentoo can't give that to me with the time I'm willing to put in. However, Gentoo did convince me to move my home directory to a BSD machine, and then export it via NFS. That way I can jump ship and move to another Linux (or BSD) at a moment's notice, without worrying about whether my filesystem is supported or any of the other issues that make changing OSes annoying.
Jesus Gentoo fanbois can be annoying. For some reason, unlike the users of every other distro, some Gentoo users think everyone would be happier with the decision they've made for themselves.
Some people like Gentoo, but some people have serious issues with it. emerge is a decent package manager, but it's attached to a distro that conservative users aren't going to touch. The more conservative distros have package managers that their users are already perfectly happy with, so it's unlikely to be used anywhere else.
Developers want to be able to release packages that work on all the Linuxes, not just Gentoo. Not everyone wants to make the fast updates/reliability tradeoff necessary to use Gentoo.
If you could convert the whole planet into devices that understand IPv6, and you could make each device out of a few thousand atoms, then we'd be in trouble. Otherwise, no. 2^128 is a big number.
The routing strategies they use cut that down quite a bit, but it's okay because most of the mass of the Earth is tied up as molten rock and stuff like that. We'll be okay unless the Earth gets eaten by nanomachines.
"I had to create a symlink to the executable, which was/usr/local/Adobe/Acrobat7.0/bin/acroread"
I'd rather it be tucked away somewhere like that than put itself somewhere where it might get in the way of something important. Executable installers on a *nix make me really nervous.
download download download...
You know it lets you choose the install path right?
"And finally, there is no indication yet on the direction the gov't wants to go with our oft-cherised "private copying" right, which currently may or may not apply to downloading music onto your computer. (No, it is not clearly legal in Canada, despite what newspapers and other slashdot posters say. It's just very hard to identify and sue infringers.)t"
Truly private copying, like putting songs from a CD I own on an MP3 player that I own is what I'd like to stick around, but from the what I'm reading it looks like they're not out to make the copyright holder omnipotent.
Distributing copyrighted material without consequences was on borrowed time anyway. I'd prefer these changes be passed during a minority government where they can't afford to piss anyone off than during a liberal supermajority where they can do anything they like.
"In conformity with the WCT and WPPT, the circumvention, for infringing purposes, of technological measures (TPMs) applied to copyright material would itself constitute an infringement of copyright. Copyright would also be infringed by persons who, for infringing purposes, enable or facilitate circumvention or who, without authorization, distribute copyright material from which TPMs have been removed. It would not be legal to circumvent, without authorization, a TPM applied to a sound recording, notwithstanding the exception for private copying."
As I read this, private copying (eg, backups of iTunes songs and protected CDs) is okay.
"This update resolves a newly-discovered, publicly reported vulnerability. A vulnerability exists in the HTML Help ActiveX control in Windows that could allow information disclosure or remote code execution on an affected system."
"Now the chip-in-your computer analogy is ridiculous. Nobody is going to force application developers to sign their executables. For one, they change far too frequently. For another, downloaded software would cease to function."
While I agree with the vast majority of your post, I don't think signing executables would be particularly difficult. Once it becomes common the compiler will do it whenever it rebuilds the binary (does it do that now? I'm on the *nix side of things so I don't know or care), so each new version of the binary is automatically signed already. Downloaded binaries that are signed are still signed after you download them, so that's a non-issue as well.
Still, I'm not sure how much good it would do. There's plenty of ways to get applications to do stuff they're not supposed to. The only way to stop them from doing stuff they're not supposed to is to make them bug free.
"Second, how would a secure computing platform work? You would still need virtualization capabilities. And once you can virtualize a platform, you can do pretty much whatever you want with it."
Yes. The thought of a company with a timetable and a budget making the virtual machine so secure that it can't successfully attacked is hilarious. If the OpenBSD people fuck up occasionally when they don't care about time and don't care much about money, how are Microsoft or Apple going to do it?
And if the virtual machine is programmable by third parties, which it has to be to support new codecs and stuff, they have to write secure code too. Try to imagine all those startups with people working 20 hour days struggling to get something in a month after the deadline... doing security audits.
Moreover, and this is nearly as bad as the practical difficulties of dealing with "secure" hardware the user has complete access to, it's designed by a company with a timetable and a budget.
The best minds in the world fuck up cryptography and security when they have decades of time to work and peer all the review they can handle.
Along comes a company that wants to do DRM. They could do use a very strong cipher but the chip that does that costs $0.05 instead of $0.03. They could open it up to peer review but they want it secret and they want it by the end of next quarter. They could have the code audited for security but that would take an expensive consultant.
Whoops. Now the cipher can be brute-forced a few years down the road. Whoops, their implementation drops bits of the key when the user does a chosen-plaintext attack. Whoops, there's a buffer overflow in in the firmware of the DRM chip. Now it can be reprogrammed to dump the unencrypted audio stream onto the hard drive.
Big business is never going to change the way it thinks. Their decisions will be based on what will give them good margins this quarter and next, not what will keep them secure for years to come. DRM is in a terrible position because it has to go in consumer electronics, where these pressures are at their worst.
The G3 iBook I purchased new in early 2003 came with 128 mb of memory and MacOS 10.2. I went out and purchased a memory upgrade the day after it arrived because it was thrashing when I tried to do stuff like read my mail with a browser window open. It was nearly unusable for even trivial tasks in the default configuration.
"Apple designs computers that work fine out of the box with a stock configuration."
Are you saying the G3 iBooks worked fine with the stock 128 mb? That's funny... I couldn't boot 10.2 without swapping (no applications running) before I upgraded to 384.
512 mb is the minimum if you want to do more than one thing at once without swapping. I have trouble with 384 mb on 10.3.
"The number is higher but a higher number isn't always faster."
That's true. RISC binaries take up more memory than CISC because all the opcodes must be the same length.
In my experience with Windows, MacOS X, Linux and *BSD (with the "fancy" GUIs) they have very similar memory requirements. 256 is enough for an "office" machine that doesn't have to do more than web, e-mail, and word processing/small spread sheets. 512 mb is enough to comfortably multitask, and nothin' says lovin' like a gig of memory.
I agree with your skepticism of the 20% number, but people do not only buy computers for games. If they did, Intel would not be the larges graphics chipset maker (as their graphics chips suck for games).
I think Apple's share will grow, but the 20% number is pretty optimistic unless they bring out a PowerBook with a bus faster than my Athlon had in 1999.
At the jan CUUG meeting, there was this "religious war" thing and the Apple advocate was talking about how reliable Apple hardware was. I knew better, so I asked:
"Who here has an iBook?"
Maybe 20 people put their hands up.
"Who here is on your original logic board?"
Everyone put their hand down.
Apple sold flawed iBooks from 2001 to 2003, and given that they have access to all the warantee numbers they would have to be completely incompetent to not notice the higher failure rates. I therefore conclude that they knowingly sold flawed laptops to customers. The only question is why.
It's seriously hurt my trust with Apple. Everyone has bad days, but Apple knowingly sold bad machines over many revisions of the line. I feel that their admission of the flaw and offer to repair affected laptops for free was pretty much the minimum they could have done, as they faced a class action lawsuit otherwise.
Slashdot Mirror
Indeed.
Hell, you can have completely duplicate directory tree for these "Linux base" libraries if you want to hang on to what the distro already had. That would make sense as this stable base would probably never be as up to date as some of the distros.
Some things have their own libraries anyway, as it's the only sane way to support all the distros out there, but it would be a big win if applications could share these. And if they have nothing to do with the distro they're installed on, you could install these on Debian-stable or some old version of Slackware if you wanted. They'd just need kernel support for what they want to do.
"You really *do* know how to properly update your config files"
Yes. I never got breakage because of this.
"and you *aren't* running ~arch, are you?"
No. The warnings in the docs about this were quite clear.
The worst stuff was things like "stable" packages that had masked dependencies (eg famd and KDE, I forget which versions), and USE flags that change their meaning without any warning (Xinerama went from default-on from default-off).
I guess it just depends on your timing, but that's the point. I'm not comfortable with a system that has no assurances in place to make sure the "stable" branch is buildable at all times. People said it had improved when I tried it again late last year, but I had exactly the same issues, so I've pretty much given up on Gentoo for the forseeable future.
I don't hate Gentoo guys for bringing Portage to BSD. I dislike Gentoo zealots because they're really annoying, and I dislike Gentoo because it breaks itself without my help.
It only has half of the BSD philosophy. Good docs by Linux standards, and a centralized mechanism to install most 3rd party software, but terrible reliability.
The BSD philosophy includes a distaste for unreliability and breakage. We'd rather be a bit slower and a bit out of date than bleeding edge if it means we don't have to go around tracking down breakage.
I've tried Gentoo, and I loved how easy it was to install the most up to date software. However, I took a look at myself after a month or two and realized I was spending a lot of time cruising the formus, trying to find out why my system broke itself in the last update. I couldn't assume my system would be up and running at any given time, and I simply wasn't willing to put in the time it takes to keep that from happening.
I still need a Linux for various reasons, and I'd still like it to be up to date-ish. But Gentoo can't give that to me with the time I'm willing to put in. However, Gentoo did convince me to move my home directory to a BSD machine, and then export it via NFS. That way I can jump ship and move to another Linux (or BSD) at a moment's notice, without worrying about whether my filesystem is supported or any of the other issues that make changing OSes annoying.
Plenty of people do, plenty of people don't. Unless everyone does it, you can't assume it will be available for what you need.
Jesus Gentoo fanbois can be annoying. For some reason, unlike the users of every other distro, some Gentoo users think everyone would be happier with the decision they've made for themselves.
Some people like Gentoo, but some people have serious issues with it. emerge is a decent package manager, but it's attached to a distro that conservative users aren't going to touch. The more conservative distros have package managers that their users are already perfectly happy with, so it's unlikely to be used anywhere else.
Developers want to be able to release packages that work on all the Linuxes, not just Gentoo. Not everyone wants to make the fast updates/reliability tradeoff necessary to use Gentoo.
If you could convert the whole planet into devices that understand IPv6, and you could make each device out of a few thousand atoms, then we'd be in trouble. Otherwise, no. 2^128 is a big number.
The routing strategies they use cut that down quite a bit, but it's okay because most of the mass of the Earth is tied up as molten rock and stuff like that. We'll be okay unless the Earth gets eaten by nanomachines.
"I had to create a symlink to the executable, which was /usr/local/Adobe/Acrobat7.0/bin/acroread"
I'd rather it be tucked away somewhere like that than put itself somewhere where it might get in the way of something important. Executable installers on a *nix make me really nervous.
download download download...
You know it lets you choose the install path right?
"And finally, there is no indication yet on the direction the gov't wants to go with our oft-cherised "private copying" right, which currently may or may not apply to downloading music onto your computer. (No, it is not clearly legal in Canada, despite what newspapers and other slashdot posters say. It's just very hard to identify and sue infringers.)t"
Truly private copying, like putting songs from a CD I own on an MP3 player that I own is what I'd like to stick around, but from the what I'm reading it looks like they're not out to make the copyright holder omnipotent.
Distributing copyrighted material without consequences was on borrowed time anyway. I'd prefer these changes be passed during a minority government where they can't afford to piss anyone off than during a liberal supermajority where they can do anything they like.
From the FAQ:
"In conformity with the WCT and WPPT, the circumvention, for infringing purposes, of technological measures (TPMs) applied to copyright material would itself constitute an infringement of copyright. Copyright would also be infringed by persons who, for infringing purposes, enable or facilitate circumvention or who, without authorization, distribute copyright material from which TPMs have been removed. It would not be legal to circumvent, without authorization, a TPM applied to a sound recording, notwithstanding the exception for private copying."
As I read this, private copying (eg, backups of iTunes songs and protected CDs) is okay.
The US is in no position to use NAFTA demand concessions from Canada right now.
"When did you ever hear of an exploit caused by the Microsoft help system?
Using mshtml in the help system or as the desktop is NOT a security problem and never has been. You are spouting more idiotic FUD."
When? I believe the last time was January 11, 2005.
"This update resolves a newly-discovered, publicly reported vulnerability. A vulnerability exists in the HTML Help ActiveX control in Windows that could allow information disclosure or remote code execution on an affected system."
Apparently they mirror their library on some very high bandwidth sites, and their options for individually encrypted streams are limited there.
"Now the chip-in-your computer analogy is ridiculous. Nobody is going to force application developers to sign their executables. For one, they change far too frequently. For another, downloaded software would cease to function."
While I agree with the vast majority of your post, I don't think signing executables would be particularly difficult. Once it becomes common the compiler will do it whenever it rebuilds the binary (does it do that now? I'm on the *nix side of things so I don't know or care), so each new version of the binary is automatically signed already. Downloaded binaries that are signed are still signed after you download them, so that's a non-issue as well.
Still, I'm not sure how much good it would do. There's plenty of ways to get applications to do stuff they're not supposed to. The only way to stop them from doing stuff they're not supposed to is to make them bug free.
"Second, how would a secure computing platform work? You would still need virtualization capabilities. And once you can virtualize a platform, you can do pretty much whatever you want with it."
Yes. The thought of a company with a timetable and a budget making the virtual machine so secure that it can't successfully attacked is hilarious. If the OpenBSD people fuck up occasionally when they don't care about time and don't care much about money, how are Microsoft or Apple going to do it?
And if the virtual machine is programmable by third parties, which it has to be to support new codecs and stuff, they have to write secure code too. Try to imagine all those startups with people working 20 hour days struggling to get something in a month after the deadline... doing security audits.
Right.
Moreover, and this is nearly as bad as the practical difficulties of dealing with "secure" hardware the user has complete access to, it's designed by a company with a timetable and a budget.
The best minds in the world fuck up cryptography and security when they have decades of time to work and peer all the review they can handle.
Along comes a company that wants to do DRM. They could do use a very strong cipher but the chip that does that costs $0.05 instead of $0.03. They could open it up to peer review but they want it secret and they want it by the end of next quarter. They could have the code audited for security but that would take an expensive consultant.
Whoops. Now the cipher can be brute-forced a few years down the road. Whoops, their implementation drops bits of the key when the user does a chosen-plaintext attack. Whoops, there's a buffer overflow in in the firmware of the DRM chip. Now it can be reprogrammed to dump the unencrypted audio stream onto the hard drive.
Big business is never going to change the way it thinks. Their decisions will be based on what will give them good margins this quarter and next, not what will keep them secure for years to come. DRM is in a terrible position because it has to go in consumer electronics, where these pressures are at their worst.
The G3 iBook I purchased new in early 2003 came with 128 mb of memory and MacOS 10.2. I went out and purchased a memory upgrade the day after it arrived because it was thrashing when I tried to do stuff like read my mail with a browser window open. It was nearly unusable for even trivial tasks in the default configuration.
Apple wouldn't replace my machine when I asked them to after the third bad logic board.
a) Yes. You can boot a 32-bit OS in 32-bit mode on x86-64 chips.
b) Qualified yes. You can run 32-bit programs on a 64-bit OS on x86-64 chips, provided that the OS supports this. Some do, some don't.
Water cooling just increases the surface area of the heatsink by using water to move the heat around. You still can't get below ambient.
You need active cooling such as a peltier unit to get below ambient.
"Apple designs computers that work fine out of the box with a stock configuration."
Are you saying the G3 iBooks worked fine with the stock 128 mb? That's funny... I couldn't boot 10.2 without swapping (no applications running) before I upgraded to 384.
512 mb is the minimum if you want to do more than one thing at once without swapping. I have trouble with 384 mb on 10.3.
"The number is higher but a higher number isn't always faster."
That's true. RISC binaries take up more memory than CISC because all the opcodes must be the same length.
In my experience with Windows, MacOS X, Linux and *BSD (with the "fancy" GUIs) they have very similar memory requirements. 256 is enough for an "office" machine that doesn't have to do more than web, e-mail, and word processing/small spread sheets. 512 mb is enough to comfortably multitask, and nothin' says lovin' like a gig of memory.
I agree with your skepticism of the 20% number, but people do not only buy computers for games. If they did, Intel would not be the larges graphics chipset maker (as their graphics chips suck for games).
I think Apple's share will grow, but the 20% number is pretty optimistic unless they bring out a PowerBook with a bus faster than my Athlon had in 1999.
At the jan CUUG meeting, there was this "religious war" thing and the Apple advocate was talking about how reliable Apple hardware was. I knew better, so I asked:
"Who here has an iBook?"
Maybe 20 people put their hands up.
"Who here is on your original logic board?"
Everyone put their hand down.
Apple sold flawed iBooks from 2001 to 2003, and given that they have access to all the warantee numbers they would have to be completely incompetent to not notice the higher failure rates. I therefore conclude that they knowingly sold flawed laptops to customers. The only question is why.
It's seriously hurt my trust with Apple. Everyone has bad days, but Apple knowingly sold bad machines over many revisions of the line. I feel that their admission of the flaw and offer to repair affected laptops for free was pretty much the minimum they could have done, as they faced a class action lawsuit otherwise.
It might be the case that all the hardware you've tried works, but there's a TON of stuff that doesn't work.
PCI cards that don't work on MacOS are more common that those that do, and stuff like USB ethernet adaptors are very hit-or-miss.
"Let me guess: you are one of them? heh certainly gave you the biggest wake-up call you have had in your entire fucking life."
Actually no. I don't post on misc@.
I watch misc@ on and off.
I've seen him chew out people that deserve it.
I've seen him chew out people that didn't deserve it.