"The market decided" my ass. The market wants refillable ink cartridges. Companies like Lexmark and HP are using the DMCA to prevent this from happening.
Don't tell me the market decided on this crappy printer business model, where printer companies are competing to make the worst printer. If it did, the market isn't working properly. The market works when companies are competing to make the best product. This clearly is not the case in this situation: companies are competing to make the worst printer. Presumably something that sells cheap, prints out fast and high quality, but has enormous maintenance costs, and will no longer even have ink-cartridges for it in a few years.
Quite frankly, the market was tricked into accepting this fraudulant business model. All printer companies are frauds. Advertise their printers can print a certain speed, without qualifying the print-content per page, but allowing the consumer to believe that it can print out that many *full text* pages with images at high quality. Advertise that their printers can print really fast, but don't bother to inform the consumers of ink-consumption.
In short, they provide no information with which to allow consumers to make an informed decision on whether or not it would be better to buy one of the cheap-up-front expensive-ink printers or the expensive-up-front cheap-ink printers.
Also, since they never mentioned this back-weighted deal to the consumer, they don't get to bitch because consumers are looking for cheaper ink solutions.
This guy's arguing that his distribution is going to be better because of the reasons he stated. I call bullshit because it's going to require people to upgrade their computers. Sorry, but if anyone has a computer over 100MHz, it's unreasonable to ask them to upgrade their system for your OS. If you require them to do such, you're OS is bloated crap (see Windows & Mac).
The average user doesn't need to think too much to realize the following simple facts:
For the same functionality: (1) Quicker load-time = better (2) Less RAM consumption = better (3) Less CPU-intensive = better (4) Smaller size = better
Or, perhaps even better would be to have user-ratings of performance vs. functionality, allow sorting by performance, functionality, and the average of both.
Btw, new users to GNU/Linux aren't going to be impressed if it's just as slow and bloated as MacOS and Win9x, which is what it will be if you go the KDE road, and make every app have copies of all the libs it uses.
ou may as well compile every app statically and users can just get used to every *&*&%^$ app doing things differently. One point of sharing components is that they can then also share behaviours...And think about trying to fix stuff like the zlib overflow when the &*^*&^*&^ library might (or might not) be duplicated in any app installed on your system (potentially compiled using any number of secure or insecure versions of zlib for that matter).
FINALLY, the first INTELLIGENT response to my original post by someone who is actually THINKING -- unlike most of the morons who replied to my post. And I didn't even *mention* the security issue, which is a big deal.
One other thing I didn't really stress was downloading bandwidth. Most people download stuff, and few -- even on cable modems -- would want to download every library that every program uses every time they download any program.
May not be essential, but they are still *the best* way to do things. No-one wants to waste *their* hard-drive space because programmer write such sloppy code and are too lazy to write good package-mangement systems with good depencency and reverse-dependency determination. The way to deal with "dependency-hell" is to use a good package management system, like Gentoo's portage. Iow, stop using RedHat's crap.
There are other benefits of using shared libraries. One is consistency. If every Gtk program uses its own version of Gtk, you have a ton of consistency problems. Want to get them all the same? You have to do it for each and every program, which means a lot of work. Furthermore, having shared libraries creates common trouble-shooting points for system-wide stability and performance issues.
Your "solution" is simply the lazy man's way out, and presents more problems, some of them just as bad as "dependency hell". The right thing to do use better package management. We should not model ourselves after MacOS' bloated slow resource-pig system.
thanks for the correction. In that case, it really is named poorly. Rename it to/sys-resrc or/sysresrc. We can leave off the unix because we all know what OS we're in.
"Speed is no longer an issue". Bullshit. Some people want to do more with their computer than just open up KDE and admire how pretty (but useless) it is. If you asked me, programmers should clock whatever CPU's their working on back to 100MHz. If their software can't run at 100MHz, it's crap (that's about 10 years ago, by now). What most productivity applications do *should not* require high-powered hardware. It should not take a lot of power to do word-processing, spread-sheet, internet-browsing. If it does, you have a crappy bloated piece of software, which either has too many useless features, or simply isn't coded well.
Sorry, I'm not going to upgrade my computer to run your bloated OS. Until I have a need for something more computationally intense than what I currently do, I'm not going to upgrade my computer. MacOSX is bloatware, plain and simple. If you actually have to do a lot of CPU-intensive things and use several programs at once -- especially one's that require a lot of RAM by the *nature* (not flawed design) of what they're doing [like bayesian phylogenies] -- then MacOSX will suck.
What Linux needs is not a unified desktop. What is good for one person, another will hate. Period. What it needs is a way to separate content from appearance. A way for the programmer to say "menu" for a certain program, and then a menu is created, per the user's preferences (e.g., a NeXT menu in WindowMaker, a motif-menu in a motif environment, a KDE-universal-menu in a KDE environment with the mac-like "one menu, etc).
Linux does not need to -- and should not -- become just like Windows and MacOS(X). Just because CPUs and GPUs are becoming faster, and hard-drives and RAM are becoming faster and cheaper, is no excuse for programmers to write crappy bloated code.
Jesus christ. Some people are fucking dense. Not everyone has a 100gig HD. Even if they did, so what? Why should *MY* space be wasted reproducing the same files in many many directories. Every program having its own Gtk library would be absurd bullshit. Firstly, that means that if the user finds out there's an update of gtk with enhanced stability and performance, (s)he has to update every single one of those things. Oh yea, many people are still downloading stuff over the net. I don't think anyone wants to spend 2x-4x as long downloading because every program includes every library it needs, even if that library already exists on your system. The other boneheaded part of your idea is the fact that every program would be using its own different version of the library. This means inconsistency.
Who are you to say "space-waste doesn't matter"? That's your preference. Quite frankly, few others agree. Most people don't want their hard-drive cluttered up with crap, especially when they can just refer to the same thing. Your idea removes most of the benefit of having libraries in the first place. Might as well make the entire program one collosal self-contained executable, HD-space and RAM be damned. Anyone who wants to waste their precious RAM and HD-space because of your crappy engineering decisions is more than welcomed too, however.
The right way to do it is simply have better dependency maintenance. Gentoo does it pretty well.
Windows has a monopoly. 96% market share -- monopoly. Period. MS has used every dirty trick in the book to do as they please. They will be able to up the ante because many of their sales are OEM, where newer and faster computers exist.
That doesn't mean it's the right way. It's the WRONG way. The OS should be sitting in the background taking up minimal resources so that I can use those resources on important things. I should not have to shell out $1000 just to buy a system that is capable of running WinXP.
Maybe what you don't understand is that to people who actually want to get work done, it is important for the OS to take up minimal system resources.
Why should anyone running older computer give a flying fuck about his sucky bloated distribution?
I am not going to upgrade my computer because some asshole says "you need 64MB of RAM, 4GB, a 1GHz computer, and a 32MB graphics card to run my OS". Sorry. The OS and applications just shouldn't be that intensive. I can understand games requiring people to get better computers -- they are improving and becoming more detailed and complex at the same (or greater) pace than CPUs and GPUs are becoming faster. But not the fucking OS, word-processor, internet program, etc.
My old 100MHz Compaq is still good enough for file-browing, web-browing, document-creation, and other daily activity. And it always should be. People should not need the latest greatest system -- or anywhere near that -- for day-to-day activities.
The problem with that is that you get lots of wasted hard-drive space if many applications use the same libraries and reproduce them all in their own directories.
It is superior to have differnet applications calling on the same library, as this reduces bloat on the HD, reduces RAM-usage, and creates a single point from which stability and performance issues can be addressed accross different applications.
The problem is managing these things well so that you don't get into . hell. Gentoo does a pretty damn good job; RedHat does a pretty damn bad job.
Why should every application have it's own private version of said library (say GTK)? This just means that lots of space is wasted on the HD, and the user has to spend more time downloading stuff. Furthermore, if the user wants any performance improvments to be gained from libraries that multiple applications use, (s)he will have to do this for every single application individually.
Ok, so this guy has an article about how confusing it is that there's a zllion GNU/Linux distros, and he wants to add one more -- his own? His own distro which does everything backwards from other distros, so that users can't use any of the help-documents that apply to all GNU/Linux distros?
His complaints abou the file-system hierarchy are noted. However, I believe he is wrong. There is/boot -- the portion you boot from./dev -- where devices (like your CDROM) are./mnt -- where devices are mounted and accessible from./root and/usr -- where most of the applicaitons are. Then there's/home -- where the user's stuff is. How exactly doesn't this make sense? My suggested improvements would be renaming/dev to/devices,/usr to/user, and/mnt to/mounted-devices.
I think this guy's comments are certainly not taylored towards making a good GNU/Linux distributiion overall -- but only one that is good for people with 1+GHz systems. Only allowing people to choose what are clearly the most bloated applications? I don't think so. Obviously, this guy doesn't give a flying fuck about anyone who wants to use Linux for older computers.
Rather than eliminating choices, the distributions should give users the information to make better choices. Mark one e-mail client as the preferred "light" client, and several others as preferred "well-featured" clients for various environments. Also, for categories (in Gentoo) like net-mail, provide a spreadsheet of features and which e-mail clients have those features, as well as binary-sizes, RAM-sizes, and benchmarks of run-time performance, load-time; also, user ratings.
I really don't buy that this article was written by Venter. I may not like the man for various proprietary-related reasons, but he is certainly a better writer than this rubbish in the article.
These may be his ideas expressed in this article, but I don't buy for one second that he actually wrote this. Some journalist probably screwed up all of his quotes, or took notes and simply could not present the ideas clearly.
Putting aside the literary flaws of this article, Venter has some interesting ideas. The question is how to do this in a way that is as open to the public as possible, yet still provide the funding. I do not think that Venter particularly wants to accomplish these goals by proprietary means; however, it does not seem like there are many alternatives. What he wants to do is sequence 10,000 genomes. Remember how long it took just to sequence 5 genomes? Well, he wants to sequence 10,000.
He also wants to develop technologies to allow us to sequence each individual's genome. All of this stuff costs lots of money. It makes sense that the software and algorithms to do this should -- and probably will -- be open, simply so that various methods can be improved upon and scrutinized. However, the costs of the hardware needed to accomplish this is mind-boggling. I'd suggest using beowulf linux clusters with very powerful individual computers; however, it's still going to be an enormous cost.
In regards to sequencing each individual's genome, some of the technology may be on the way. There are ideas about sequencing DNA base-pairs by recording electric readings as DNA is pulled through a pore (the electric readings produced by A, T, C, and G all differ).
Yes, actually it is. Companies always do what they think will make them the most money. They might not always be right. That's because executives aren't perfect -- they make mistakes. If every company always took the ideal course of action to make the most money, no companies would go out of business. However, the companies that stick around are the one's that are good at identifying the best moeny-making courses, and that have no qualms about switching positions hypocritically: in other words, the one's that have no morals.
The cable-TV industry has said there's no reason to adopt such rules, because it has no intention of discriminating against Web sites or limiting new technologies.
So, we're supposed to take their word for that, right?
This is just one more example of why companies are completely hypocritical and can never be taken at their word. If MS, Amazon.com, and Walt Disney were in the position of AOL/TimeWarner, they would take *exactly* the opposite position. Worse yet, if they switched positions with AOL/TimeWarner, then they would switch to *exactly* the opposite position.
Lessig has talked about his in "The Future of Ideas".
None of these companies have the public interest in mind. Only *their* interest. They can make useful allies in the same sense that mercenaries make useful allies: temporary, unloyal, and certainly not trustworthy.
No. The list of prompted directories should be configurable, and it should be up to the distribution to configure
I agree. Protected directories should *not* be hard-coded. Should be up to the distro, admin, and (eventually) user for his own files. Someone else pointed out a superior way to do this, which wouldn't require changing the file-system: create a list of protected directories in a certain file, and have rm refer to it.
When I said/,/boot,/home,/usr, etc should contian such a prompt, I was referring to my own opinion. Depending on what people are doing, some of those directories might not be too important.
o you want the rm command to be asking a shell script if it is sure it wants to delete the/usr directory
Well, I guess the shell script would have to either be programmed to answer yes to that question, override the prompt, or ask the user. That's why you have to test these things;-).
To which I reply: useless. A firealarm that goes off every time you light a match is useless.
there are (at least prototype) undelete utilities for ext2
Alot people use FS' other than ext2 now. I use ReiserFS for it's performance, for example (and for it's journaling capabilities). ext3 is not desireable because it's performance sucks.
I don't think there's an undelete util for ReiserFS; not sure why, since files are deleted there by default pretty much the same way (e.g., mark them as "free space", but don't write zeros over them). However, there is the "unix way" to do it:
grep -a -B[size before] -A[size after] 'text'/dev/[your_partition]
but if it means adding a new flag to the filesystem (and therefore creating a new filesystem format), then before long, someone will decide to add another new feature - an rm option that says `override protection flag -- use with extreme caution'. And then you're back to where we are now, with the -f option overriding the readonly flag
Why would anyone do that? As opposed to just removing said flag. Sure, they could do it. If someone's insistent on blowing their brains out, no safety on a gun is going to stop them. Also, even with such a flag, at least the user would have to consciously make a decision to open hiim/herself up to that.
Presumably if you were to create an rm that reads a list of protected files (eg/etc/dontrm) then someone would add a flag to circumvent that, too.
Well, I know when to admit I'm wrong. Your idea is better than mine (a list of protected files, as opposed to a flag, is better because it does not involve modifying hte file-system, and seems like it would take less work to implement, thus be a cleaner better solution. Yes, someone could circumvent that. Anyone who has root access (presumably all GNU/Linux home users) can circumvent the superuser/user protection system by logging in as root all the time, as well. Why don't they do that? Because every time ne1 goes into an IRC as "root", they immediately get told to make a user account and come back; iow, there's a social pressure. It also might be that most people are intelligent enough to realize that not logging in as root is a good idea.
It still saves directories - the first y is to say `Descend into directory...?', and doesn't actually delete anything. Then when you get a second question after pressing `y' you might realise that it's not about to do what you expected.
But what if you really wanted to delete a directory titled "boot"? Except you wanted to delete './boot' not '/boot'. So then you go yes yes yes yes yes anyways, and then you can't boot.
Personally, I delete stuff because it's taking valuable disk space, so I don't want it to go and take up the same amount of space under the name `recycle bin'.
Probably why most of us delete stuff. That and to eliminate confusing extra information. As for the space-issue. That's why it'd be a good idea to have a logout script that:
Lists the contents of the trashcan.
Beeps if a protected file is in it.
Prompts the user, "Empty trashcan? cancel logout? (y/n/esq,C^C)"
[optional]Waits 5 seconds before accepting input, to force the user to read the list. [disabled at user's preference]
Automatically logs out without emptying the trashcan (n) after 30s. This is in case the admin log out, forgetting that this script happens, to prevent someone else from coming along and being able to be a sysadmin by pressing esq or CTRL-C (C^C).
[?]Requests admins password before deleting files or canceling the logout. This way, if the admin does leave forgetting this, someone else can't just delete the files or get back in in that 30s window.
So does that mean we should put a "spherical clear cover that has to be unhinged" over the return button?
Nope, we press return all the time -- usually for non-destructive reasons, like making a new paragraph. Return.
However, maybe having such a spherical clear cover over the POWER ON/OFF button on the box, AND the POWER ON/OFF button on the surge-protector power-strip, would be helpful.
What exactly is it again about a car that's not easily fucked up?
It is not feasable to make collision-detectable cars affordable for the average US citizen. This would add great costs.
However, cars do have basic good design choices. The breaks and accelerator pads are gripped, and far enough from one another that you won't press one by accident when you meant to press the other. The stearing wheel is gripped. There are seatbelts. The hard-brake lever is out of the way, and cannot accidentally be used.
(1) I suggested a solution. Add code in rm that looks for a PROMPT flag in anything it deletes. If such a PROMPT flag is present, it will prompt the user. Which directories will have this flag should be up to the maker of the distribution (as different directories have differing importance, depending on the distro). No more than 1% of all files/folders should have this flag -- otherwise, it's no different than "-i", which just engrains in the user the habit to press 'y' after typing 'rm'.
, a computer simply cannot consider social realities and human imperfections on the fly
Yep, that's why the people making the program should consider that for the program.
It will differ from machine to machine [which folders should be protected]
Which is why you'd have a flag attributable to folders, that would tell rm to prompt the user "are you sure?". Which folders got this flag would be up to the people who put together the particular distro (of, if they don't do it, the end-user or administrator who wants to harden it).
There's already one level of safeguards in place by aliasing rm to `rm -i'
Wrong. Always being prompted "are you sure (y/n)" is *not* a safeguard at all. It simply builds into you the habit to press 'y' after typing 'rm'. A firealarm is useless if it goes off every 10 minutes. This is why only a very few of the most important folders and files should have the aforementioned flag.
Heh, so this was an actual problem in the aircraft industry? I just finished reading Micheal Crichtons 'Airframe'
Ok, so ya caught me. I didn't bother to check and see if this really was a problem in the aircraft industry. I too got this from Airframe. Yep, great book. It seems like something that very well could have happened.
However, google 'uncommanded slats extension' (without quotes). You'll find some interesting results:
Slashdot Mirror
"The market decided" my ass. The market wants refillable ink cartridges. Companies like Lexmark and HP are using the DMCA to prevent this from happening.
Don't tell me the market decided on this crappy printer business model, where printer companies are competing to make the worst printer. If it did, the market isn't working properly. The market works when companies are competing to make the best product. This clearly is not the case in this situation: companies are competing to make the worst printer. Presumably something that sells cheap, prints out fast and high quality, but has enormous maintenance costs, and will no longer even have ink-cartridges for it in a few years.
Quite frankly, the market was tricked into accepting this fraudulant business model. All printer companies are frauds. Advertise their printers can print a certain speed, without qualifying the print-content per page, but allowing the consumer to believe that it can print out that many *full text* pages with images at high quality. Advertise that their printers can print really fast, but don't bother to inform the consumers of ink-consumption.
In short, they provide no information with which to allow consumers to make an informed decision on whether or not it would be better to buy one of the cheap-up-front expensive-ink printers or the expensive-up-front cheap-ink printers.
Also, since they never mentioned this back-weighted deal to the consumer, they don't get to bitch because consumers are looking for cheaper ink solutions.
Instead of complaining about HP ink prices, promote those suppliers who make refillable printers
(1) Name any major suppliers that make refillable printers?
(2) If there are any, please tell me how many of them support FS/OSS?
Who cares? IP laws are supposed to benefit the consumers. Thus, because Static Control's product benefits the consumers, it should be left alone.
That'd be a valid line of logic if we were talking about some entrepreneur no-one's ever heard of before. We're not. We're talking about Craig Venter.
This guy's arguing that his distribution is going to be better because of the reasons he stated. I call bullshit because it's going to require people to upgrade their computers. Sorry, but if anyone has a computer over 100MHz, it's unreasonable to ask them to upgrade their system for your OS. If you require them to do such, you're OS is bloated crap (see Windows & Mac).
The average user doesn't need to think too much to realize the following simple facts:
For the same functionality:
(1) Quicker load-time = better
(2) Less RAM consumption = better
(3) Less CPU-intensive = better
(4) Smaller size = better
Or, perhaps even better would be to have user-ratings of performance vs. functionality, allow sorting by performance, functionality, and the average of both.
Btw, new users to GNU/Linux aren't going to be impressed if it's just as slow and bloated as MacOS and Win9x, which is what it will be if you go the KDE road, and make every app have copies of all the libs it uses.
ou may as well compile every app statically and users can just get used to every *&*&%^$ app doing things differently. One point of sharing components is that they can then also share behaviours...And think about trying to fix stuff like the zlib overflow when the &*^*&^*&^ library might (or might not) be duplicated in any app installed on your system (potentially compiled using any number of secure or insecure versions of zlib for that matter).
FINALLY, the first INTELLIGENT response to my original post by someone who is actually THINKING -- unlike most of the morons who replied to my post. And I didn't even *mention* the security issue, which is a big deal.
One other thing I didn't really stress was downloading bandwidth. Most people download stuff, and few -- even on cable modems -- would want to download every library that every program uses every time they download any program.
May not be essential, but they are still *the best* way to do things. No-one wants to waste *their* hard-drive space because programmer write such sloppy code and are too lazy to write good package-mangement systems with good depencency and reverse-dependency determination. The way to deal with "dependency-hell" is to use a good package management system, like Gentoo's portage. Iow, stop using RedHat's crap.
There are other benefits of using shared libraries. One is consistency. If every Gtk program uses its own version of Gtk, you have a ton of consistency problems. Want to get them all the same? You have to do it for each and every program, which means a lot of work. Furthermore, having shared libraries creates common trouble-shooting points for system-wide stability and performance issues.
Your "solution" is simply the lazy man's way out, and presents more problems, some of them just as bad as "dependency hell". The right thing to do use better package management. We should not model ourselves after MacOS' bloated slow resource-pig system.
thanks for the correction. In that case, it really is named poorly. Rename it to /sys-resrc or /sysresrc. We can leave off the unix because we all know what OS we're in.
"Speed is no longer an issue". Bullshit. Some people want to do more with their computer than just open up KDE and admire how pretty (but useless) it is. If you asked me, programmers should clock whatever CPU's their working on back to 100MHz. If their software can't run at 100MHz, it's crap (that's about 10 years ago, by now). What most productivity applications do *should not* require high-powered hardware. It should not take a lot of power to do word-processing, spread-sheet, internet-browsing. If it does, you have a crappy bloated piece of software, which either has too many useless features, or simply isn't coded well.
Sorry, I'm not going to upgrade my computer to run your bloated OS. Until I have a need for something more computationally intense than what I currently do, I'm not going to upgrade my computer. MacOSX is bloatware, plain and simple. If you actually have to do a lot of CPU-intensive things and use several programs at once -- especially one's that require a lot of RAM by the *nature* (not flawed design) of what they're doing [like bayesian phylogenies] -- then MacOSX will suck.
What Linux needs is not a unified desktop. What is good for one person, another will hate. Period. What it needs is a way to separate content from appearance. A way for the programmer to say "menu" for a certain program, and then a menu is created, per the user's preferences (e.g., a NeXT menu in WindowMaker, a motif-menu in a motif environment, a KDE-universal-menu in a KDE environment with the mac-like "one menu, etc).
Linux does not need to -- and should not -- become just like Windows and MacOS(X). Just because CPUs and GPUs are becoming faster, and hard-drives and RAM are becoming faster and cheaper, is no excuse for programmers to write crappy bloated code.
Jesus christ. Some people are fucking dense. Not everyone has a 100gig HD. Even if they did, so what? Why should *MY* space be wasted reproducing the same files in many many directories. Every program having its own Gtk library would be absurd bullshit. Firstly, that means that if the user finds out there's an update of gtk with enhanced stability and performance, (s)he has to update every single one of those things. Oh yea, many people are still downloading stuff over the net. I don't think anyone wants to spend 2x-4x as long downloading because every program includes every library it needs, even if that library already exists on your system. The other boneheaded part of your idea is the fact that every program would be using its own different version of the library. This means inconsistency.
Who are you to say "space-waste doesn't matter"? That's your preference. Quite frankly, few others agree. Most people don't want their hard-drive cluttered up with crap, especially when they can just refer to the same thing. Your idea removes most of the benefit of having libraries in the first place. Might as well make the entire program one collosal self-contained executable, HD-space and RAM be damned. Anyone who wants to waste their precious RAM and HD-space because of your crappy engineering decisions is more than welcomed too, however.
The right way to do it is simply have better dependency maintenance. Gentoo does it pretty well.
Windows has a monopoly. 96% market share -- monopoly. Period. MS has used every dirty trick in the book to do as they please. They will be able to up the ante because many of their sales are OEM, where newer and faster computers exist.
That doesn't mean it's the right way. It's the WRONG way. The OS should be sitting in the background taking up minimal resources so that I can use those resources on important things. I should not have to shell out $1000 just to buy a system that is capable of running WinXP.
Maybe what you don't understand is that to people who actually want to get work done, it is important for the OS to take up minimal system resources.
Why should anyone running older computer give a flying fuck about his sucky bloated distribution?
I am not going to upgrade my computer because some asshole says "you need 64MB of RAM, 4GB, a 1GHz computer, and a 32MB graphics card to run my OS". Sorry. The OS and applications just shouldn't be that intensive. I can understand games requiring people to get better computers -- they are improving and becoming more detailed and complex at the same (or greater) pace than CPUs and GPUs are becoming faster. But not the fucking OS, word-processor, internet program, etc.
My old 100MHz Compaq is still good enough for file-browing, web-browing, document-creation, and other daily activity. And it always should be. People should not need the latest greatest system -- or anywhere near that -- for day-to-day activities.
The problem with that is that you get lots of wasted hard-drive space if many applications use the same libraries and reproduce them all in their own directories.
It is superior to have differnet applications calling on the same library, as this reduces bloat on the HD, reduces RAM-usage, and creates a single point from which stability and performance issues can be addressed accross different applications.
The problem is managing these things well so that you don't get into . hell. Gentoo does a pretty damn good job; RedHat does a pretty damn bad job.
Why should every application have it's own private version of said library (say GTK)? This just means that lots of space is wasted on the HD, and the user has to spend more time downloading stuff. Furthermore, if the user wants any performance improvments to be gained from libraries that multiple applications use, (s)he will have to do this for every single application individually.
Ok, so this guy has an article about how confusing it is that there's a zllion GNU/Linux distros, and he wants to add one more -- his own? His own distro which does everything backwards from other distros, so that users can't use any of the help-documents that apply to all GNU/Linux distros?
/boot -- the portion you boot from. /dev -- where devices (like your CDROM) are. /mnt -- where devices are mounted and accessible from. /root and /usr -- where most of the applicaitons are. Then there's /home -- where the user's stuff is. How exactly doesn't this make sense? My suggested improvements would be renaming /dev to /devices, /usr to /user, and /mnt to /mounted-devices.
His complaints abou the file-system hierarchy are noted. However, I believe he is wrong. There is
I think this guy's comments are certainly not taylored towards making a good GNU/Linux distributiion overall -- but only one that is good for people with 1+GHz systems. Only allowing people to choose what are clearly the most bloated applications? I don't think so. Obviously, this guy doesn't give a flying fuck about anyone who wants to use Linux for older computers.
Rather than eliminating choices, the distributions should give users the information to make better choices. Mark one e-mail client as the preferred "light" client, and several others as preferred "well-featured" clients for various environments. Also, for categories (in Gentoo) like net-mail, provide a spreadsheet of features and which e-mail clients have those features, as well as binary-sizes, RAM-sizes, and benchmarks of run-time performance, load-time; also, user ratings.
I really don't buy that this article was written by Venter. I may not like the man for various proprietary-related reasons, but he is certainly a better writer than this rubbish in the article.
These may be his ideas expressed in this article, but I don't buy for one second that he actually wrote this. Some journalist probably screwed up all of his quotes, or took notes and simply could not present the ideas clearly.
Putting aside the literary flaws of this article, Venter has some interesting ideas. The question is how to do this in a way that is as open to the public as possible, yet still provide the funding. I do not think that Venter particularly wants to accomplish these goals by proprietary means; however, it does not seem like there are many alternatives. What he wants to do is sequence 10,000 genomes. Remember how long it took just to sequence 5 genomes? Well, he wants to sequence 10,000.
He also wants to develop technologies to allow us to sequence each individual's genome. All of this stuff costs lots of money. It makes sense that the software and algorithms to do this should -- and probably will -- be open, simply so that various methods can be improved upon and scrutinized. However, the costs of the hardware needed to accomplish this is mind-boggling. I'd suggest using beowulf linux clusters with very powerful individual computers; however, it's still going to be an enormous cost.
In regards to sequencing each individual's genome, some of the technology may be on the way. There are ideas about sequencing DNA base-pairs by recording electric readings as DNA is pulled through a pore (the electric readings produced by A, T, C, and G all differ).
Yes, actually it is. Companies always do what they think will make them the most money. They might not always be right. That's because executives aren't perfect -- they make mistakes. If every company always took the ideal course of action to make the most money, no companies would go out of business. However, the companies that stick around are the one's that are good at identifying the best moeny-making courses, and that have no qualms about switching positions hypocritically: in other words, the one's that have no morals.
The cable-TV industry has said there's no reason to adopt such rules, because it has no intention of discriminating against Web sites or limiting new technologies.
So, we're supposed to take their word for that, right?
This is just one more example of why companies are completely hypocritical and can never be taken at their word. If MS, Amazon.com, and Walt Disney were in the position of AOL/TimeWarner, they would take *exactly* the opposite position. Worse yet, if they switched positions with AOL/TimeWarner, then they would switch to *exactly* the opposite position.
Lessig has talked about his in "The Future of Ideas".
None of these companies have the public interest in mind. Only *their* interest. They can make useful allies in the same sense that mercenaries make useful allies: temporary, unloyal, and certainly not trustworthy.
No. The list of prompted directories should be configurable, and it should be up to the distribution to configure
/, /boot, /home, /usr, etc should contian such a prompt, I was referring to my own opinion. Depending on what people are doing, some of those directories might not be too important.
/usr directory
;-).
I agree. Protected directories should *not* be hard-coded. Should be up to the distro, admin, and (eventually) user for his own files. Someone else pointed out a superior way to do this, which wouldn't require changing the file-system: create a list of protected directories in a certain file, and have rm refer to it.
When I said
o you want the rm command to be asking a shell script if it is sure it wants to delete the
Well, I guess the shell script would have to either be programmed to answer yes to that question, override the prompt, or ask the user. That's why you have to test these things
To which I reply: useless. A firealarm that goes off every time you light a match is useless.
there are (at least prototype) undelete utilities for ext2
Alot people use FS' other than ext2 now. I use ReiserFS for it's performance, for example (and for it's journaling capabilities). ext3 is not desireable because it's performance sucks.
I don't think there's an undelete util for ReiserFS; not sure why, since files are deleted there by default pretty much the same way (e.g., mark them as "free space", but don't write zeros over them). However, there is the "unix way" to do it:
Why would anyone do that? As opposed to just removing said flag. Sure, they could do it. If someone's insistent on blowing their brains out, no safety on a gun is going to stop them. Also, even with such a flag, at least the user would have to consciously make a decision to open hiim/herself up to that.
Presumably if you were to create an rm that reads a list of protected files (eg /etc/dontrm) then someone would add a flag to circumvent that, too.
Well, I know when to admit I'm wrong. Your idea is better than mine (a list of protected files, as opposed to a flag, is better because it does not involve modifying hte file-system, and seems like it would take less work to implement, thus be a cleaner better solution. Yes, someone could circumvent that. Anyone who has root access (presumably all GNU/Linux home users) can circumvent the superuser/user protection system by logging in as root all the time, as well. Why don't they do that? Because every time ne1 goes into an IRC as "root", they immediately get told to make a user account and come back; iow, there's a social pressure. It also might be that most people are intelligent enough to realize that not logging in as root is a good idea.
It still saves directories - the first y is to say `Descend into directory...?', and doesn't actually delete anything. Then when you get a second question after pressing `y' you might realise that it's not about to do what you expected.
But what if you really wanted to delete a directory titled "boot"? Except you wanted to delete './boot' not '/boot'. So then you go yes yes yes yes yes anyways, and then you can't boot.
Personally, I delete stuff because it's taking valuable disk space, so I don't want it to go and take up the same amount of space under the name `recycle bin'.
Probably why most of us delete stuff. That and to eliminate confusing extra information. As for the space-issue. That's why it'd be a good idea to have a logout script that:
So does that mean we should put a "spherical clear cover that has to be unhinged" over the return button?
Nope, we press return all the time -- usually for non-destructive reasons, like making a new paragraph. Return.
However, maybe having such a spherical clear cover over the POWER ON/OFF button on the box, AND the POWER ON/OFF button on the surge-protector power-strip, would be helpful.
What exactly is it again about a car that's not easily fucked up?
It is not feasable to make collision-detectable cars affordable for the average US citizen. This would add great costs.
However, cars do have basic good design choices. The breaks and accelerator pads are gripped, and far enough from one another that you won't press one by accident when you meant to press the other. The stearing wheel is gripped. There are seatbelts. The hard-brake lever is out of the way, and cannot accidentally be used.
(1) I suggested a solution. Add code in rm that looks for a PROMPT flag in anything it deletes. If such a PROMPT flag is present, it will prompt the user. Which directories will have this flag should be up to the maker of the distribution (as different directories have differing importance, depending on the distro). No more than 1% of all files/folders should have this flag -- otherwise, it's no different than "-i", which just engrains in the user the habit to press 'y' after typing 'rm'.
(2) Two words: libtrash and perltrash.
, a computer simply cannot consider social realities and human imperfections on the fly
Yep, that's why the people making the program should consider that for the program.
It will differ from machine to machine [which folders should be protected]
Which is why you'd have a flag attributable to folders, that would tell rm to prompt the user "are you sure?". Which folders got this flag would be up to the people who put together the particular distro (of, if they don't do it, the end-user or administrator who wants to harden it).
There's already one level of safeguards in place by aliasing rm to `rm -i'
Wrong. Always being prompted "are you sure (y/n)" is *not* a safeguard at all. It simply builds into you the habit to press 'y' after typing 'rm'. A firealarm is useless if it goes off every 10 minutes. This is why only a very few of the most important folders and files should have the aforementioned flag.
Also, there should be a trashcan.
Ok, so ya caught me. I didn't bother to check and see if this really was a problem in the aircraft industry. I too got this from Airframe. Yep, great book. It seems like something that very well could have happened.
However, google 'uncommanded slats extension' (without quotes). You'll find some interesting results: