The "are you sure y/n" method is ONLY useful if it is an uncommon message.
I agree completely. A firealarm is of no use if it goes off all the time for no reason, because people will learn to ignore it. So, I agree with you, that it should be a rare occasion when you are prompted y/n. Directories like/,/boot,/home,/usr should contain such a prompt. The user deletes them with enough rarity that a y/n would be out of the ordinary. Likewise, the user should be able to specify some of his own files that (s)he wants to be prompted on (though, the user should be warned that placing this on every file is useless). I'd say about 1% is a good maximum.
This also illustrates why there should be a trash can.
No. The first purpose of the computer should be to make as MANY THINGS as possible, possible. Doing so in the easiest way is the second purpose. If optimizing for the simple case causes the complex case to fail, you shouldn't do it that way. Such is the way with "rm". The only way to make mistakes impossible with it is to reduce its utility.
Sorry, but this is just wrong-headed. A computer is a tool to get a job done (or rather, many jobs done). Overall, the best system is one which minimizes the time the user has to spend in front of it to get their work done. The idea is not to make mistakes impossible with rm: that would prevent sysadmins from deleting things they really wanted to delete; the idea is to make it a little harder to do so by accident. And -- preferrably -- to have some kind of trashcan, so that even if that happens, things are ok.
The other idea is that some folder's are so important that they should come with a tag by default requesting that rm prompt the user if (s)he really wants to remove them. Furthermore, user's should be able to place such a "confirm prompt" tag on any file/folder they deem that important.
With a computer interface, it's not the same. Even when the user does something strange, that might really be a valid action. You can't have the OS trying to out-guess the user. There might be a valid reason to rm -rf/.
This would be a valid point, except you don't consider the asymmetry of consequence here. Sure, there could be a valid reason for deleting the entire root partition. It is more likely, however, that the user just fucked up. So, for such important directories, it makes sense to place a tag on those directories making sure the user is prompted whether or not to delete them. If you really wanted to delete it, how much does it slow you down? Maybe a few seconds. If you didn't, it saves you a huge headache.
There is nothing new or innovative in this design. Simply taking a scroll wheel and extending it in a 2nd dimension does not count as something worthy of a patent. Neither does a scroll wheel.
I really tire quickly when people continue to defend shit.
Let's take your butcher example. Yes, butcher knives are sharp. Yep, butchers can cut themselves. That does not, however, mean that those making them ignore safety. You notice how butcher knives have high-friction grips on them, right? Notice how the handles aren't made out of smooth mahogony polished so finely that it's slippery, right? Could that be good social design? Maybe? Just maybe?
How is prompting the user if he really wants to recursively delete the *entire* hard-drive ('rm -rf/') a bad thing? Sure, it'll slow down that expert user who does it once a year by a few seconds. But it'll save others hours work from not having to restore backups.
if you're a good admin you've already planned for yourself fucking up
LOL, why is it that all of it falls on the user? According to you, good programmers don't plan on the user fucking up? What would be the harm in: (1) Placing destructive rm options far away from non-destructive ones; (2) Prompting on recursive deletes of important folders; (3) Instead of removing files, placing them in the trashcan, with an auto-script on logout listing/trash contents and prompting "empty? (YES/n)"? You need to consider that there is a *huge* asymmetry of consequence for many y/n actions -- especially "delete".
But let's just say I'm a good system administrator, and I'm using your rm program which assumes I never fuck up and blindly follows me off a cliff. But I do fuck up: about once a year, and I've planned for that by writing a script that auto-backs up things on a daily basis (backing up progressively in a diff-format, for performance). So, it's day 365 of the year, right? Time for me to fuck up. So, I fuck up and accidentally delete a binary directory. It doesn't contain much, just a rare-used program that gets used about every week or so. So, it takes about a week for me to realize something's wrong.
After a week, someone reports to me, "Hey, this fuckin' program isnt' working. I type "program", and it says "command not found": fix it asshole". This pisses me off. After all, this's the first time I've fucked up all year. How dare this fuck call me an asshole? So, I spend about 5 minutes seething in anger.
After I'm finished pouting, I grumpily try to figure out what's going on. You see, I don't necessarily know that the program's been deleted. So, I look in/usr/bin and check out the program link. Yep, shortcut's there. Then I check in where the program directory should be. Aha! Not there. So, now I go look for "program" in my backup archives. After a minute, I find it. Then I copy it over to where it should have been. Bingo! Problem solved. Test it out and it works. This took about another 2-5 minutes.
So, I go back to said asshole who called me an asshole. I shortly say, "Problem solved". He grudgingly thanks me.
Now, before this prick interrupted me, I was doing something. But, my concentration was disrupted, and a period of high productivity on my part was ended. It takes me a minute to remember what I was doing, and when I get back to it, I'm not really in the zone. Plus, I'm still pissed off. Consequently, my productivity for the rest of the day falls 25%.
Now, let's consider another "rm" program written by some software guy names Joel. This rm program prompts me upon deleting important directories, like/bin. However, this was just a specific program directory, so it doesn't prompt me -- perfectly possible I really wanted to remove it. However, instead of deleting those files, it moves them to a trashcan. Upon logging out, a script lists the contents of the trashcan, and prompts me if I want to delete them (y/n) or cancel the logoug (esc, C^C, etc). The program does not accept input (other than esc) for 5sec, thus making it worth my while to at least look over what's there. As a security feature, it automatically logs out after 30seconds without deleting the
Obviously the 85% posting mindless FPs and flames, or simply not posting at all, since most everyone here seems to hate MS Windows (or, gee...maybe they just come here more at work [where they use windows] than at home [where they use GNU/Linux]).;-)
you give the computer a command it should execute it.
No. The purpose of the computer should be to on average allow the user to get work done as fast as possible. If -- because the computer stupidly executed a destructive typo -- it destroys a days worth of work, it has completely defeated it's purpose: to allow work to be done faster. Part of the way to resolve this is simply to have more intelligent command-line options. Command line options that perform destructive tasks should use letters that are *physically* *far* appart from the letters used to invoke harmless command-line options.
How often is someone going to 'rm -rf/"? Not very often at all. (in fact, in such a scenario, it would probably be more intelligent to simply reformat that partition). But, in 9/10 times where someone types 'rm -rf/boot', they really meant to type 'rm -rf./boot'. Period. Now, wouldn't it make sense to have some way of specifying directories that you *don't* want to delete recursively, and prompting the rm-command to prompt hte user "are you sure" for such important directories? Also, libtrash as a default library to intercept such destructive commands and move things to a trashcan would be good
User error is user error and not the computers fault, you are free to add checks but it doesn't make the statement made any less true.
Correct. A truly careless user will tend to fuck things up, even if you prompt him "really want to recursively delete entire home directory?" (shorter is better...the longer a message, the less likely the user will read it). However, you can at least put a speedbump along the road to oblivion. It might actually stop a semi-conscious user from deleting all their important info, and save them time. This is good.
Why? Because the best program that allows the user to get his/her job done the fastest. A program that increases the likelihood that the user will spend 5-10 minutes finding backups and restoring information is bad. Bad, because those 5-10 minutes are wasted. Then the user is pissed off. And his concentration is broken. Odds are, the rest of that users day are going to be unproductive. See "Joel on Software".
cryptic, and difficult to use for a reason
What reason would that be? So raving technocrats can scream, "RTFM, bitch whore!"? The GNU/Linux command line is like it is to allow for enormous power, allowing users to do things the program-creators never conceived of. That end-goal does not require that it be so easy for users to shoot themselves in the foot because they were a little bit sleepy one day.
No, I do not consider GNU/Linux to be "that hard". It's called hyperbolization. Iow, a joke. But the point is still there.
How often have I accidentally deleted something I dind't want to? Well, very rarely, in fact. Within the last year or so, I can only remember a few cases. In each case, it was a minor file, and that was easily recovered from. Other's might not have been so lucky, or so good.
Even if you're the Stephen Hawking of CLIs, you're going to fuck up on rare occasion. It's a statistical certainty. Why? Because you aren't perfect. You aren't perfect, I'm not perfect, and the normal end-user actually using an account -- that guy who just uses the computer you administer -- sure as hell isn't anywhere near perfect.
The rm examples from this book are particularly good. Your counter-point that anyone doing anything important is going to be more careful than that just doesn't hold. (1) People use GNU/Linux for their home computers. Many of these people are not administrators. They may clobber important directories. (2) Ordinary users on the computers you administer have important files. I realize that you don't give a fuck about their files, as you are in no way responsible for them accidentally deleting that document they've been working on all day, or worse -- an entire directories worth of documents, by typing 'rm *.foo' instead of 'ram *.foo'. They, however, do care about this.
The simple fact is, many commands, utilities, and programs in GNU/Linux simply do not consider social engineering. (maybe there is a positive correlation between the anti-social nature of the program and the anti-social nature of the creator?) Some particularly poor design is when two command options -- one that performs a destructive activity, the other that performs a harmless activity -- either use the same key in lower/upper case, or two keys near eachother.
As an example of *good social design*, consider the command "perltrash". Perltrash has a -e and -l option. -e empties the trash. -l lists the contents of it. These keys are very far away from eachother on the keyboard. Now, they could have just as easily used -e for empty and -d for directory listing of trash. This would, however, be crappy social design. Why? Because it does not account for the fact that it is extremely easy to accidentally type an "e" instead of a "d". Additionally, the creator of perltrash did another correct thing by using -r for restore and -R for restore here. Earth-shattering! Using the lower- and upper-case versions of the same letter for a similar activity!
Regarding accidental deletion of files, the best thing to do would be to simply have libtrash (a prog which intercepts all libc delete/remove commands and changes them to move the info to the trashcan) installed by default on the supposedly "easy-to-use" GNU/Linux' like RedHat, Suse, Lindows, et al; and to have packages available for the more "make everything from scratch" distros, like Slackware, Debian, and Gentoo (I'm annoyed that there is no.ebuild for that for Gentoo...I'll have to work on one).
There may not be any room for it -- but it *still* happens. Period. End of discussion. It's a statistical certainty. Even among sys admin root-users as nearly-perfect as (apparently) yourself will eventually fuck up. It is a statistical *certainty*.
Also, you seem to assume htat user's individual documents aren't important. Well, they are. And individual users certainly aren't as perfect as you.
When making something, it is simply *bad engineering* to not consider social phenomena. Technically, there would be nothing wrong with a monitor 40 inches wide and 5 inches high. But the monitor is still fucked up crappy design, because people do not have such a wide field of vision.
I'd really like to see you design airplaines. Maybe fighter planes. Just to make sure pilots are *extra careful*, the button for "drop bombs" should be right next to the button for "lower landing-gear". So what if they blow up the runway, right. Pilots are supposed to be perfect and should *not* fuck up.
The comment of one macho veteran Slashdotter have so annoyed me that I think it deserves being a main-comment for criticism:
There are no such things as "rm disasters". There are only mistakes, stop making them, or at least think before you execute.
Exactly the kind of bullshit macho attitude I was talking about.
Why don't you try doing that if you're a car company, and sell a car that can so easily be fucked up? Oh, yea, instead of having an out of-the-way hard break lever, we put a hard-break button right next to the defog button...but don't fucking bitch at us if you accidentally press the hard-break button (which is right next to the defog button) when trying to defog your windows, and your car spins around and crashes on an icy road.
Does that kind of bullshit macho attitude apply for companies making airplanes? When people making airplaies discovered that slats switches were being turned on accidentally, did they say:
"Yea, so what the slats extention switch can be accidentally turned on by an unintentional movement, possibly causing passenter-injury. Tell the pilots to be more careful and not fuck up."
No, they didn't. They said,
"Ok, so this is a problem. Why don't we cover the slats switch with a spherical clear cover that has to be unhinged before extending slats -- that way, they won't get extended at 500mph and cause the plane to trolly."
Just because many of these problems are socialogical not technological doesn't mean they're not problems. People are not robots. People fuck up -- quite a bit actually. To you perfect people writing a reply to this boldly telling me that people shouldn't "fuck up", how many times did you have to use backspace in writing that response?
The comment of one macho veteran Slashdotter have so annoyed me that I think it deserves being a main-comment for criticism:
There are no such things as "rm disasters". There are only mistakes, stop making them, or at least think before you execute.
Exactly the kind of bullshit macho attitude I was talking about.
Why don't you try doing that if you're a car company, and sell a car that can so easily be fucked up? Oh, yea, instead of having an out of-the-way hard break lever, we put a hard-break button right next to the defog button...but don't fucking bitch at us if you accidentally press the hard-break button (which is right next to the defog button) when trying to defog your windows, and your car spins around and crashes on an icy road.
Does that kind of bullshit macho attitude apply for companies making airplanes? When people making airplaies discovered that slats switches were being turned on accidentally, did they say:
"Yea, so what the slats extention switch can be accidentally turned on by an unintentional movement, possibly causing passenter-injury. Tell the pilots to be more careful and not fuck up."
No, they didn't. They said,
"Ok, so this is a problem. Why don't we cover the slats switch with a spherical clear cover that has to be unhinged before extending slats -- that way, they won't get extended at 500mph and cause the plane to trolly."
Just because many of these problems are socialogical not technological doesn't mean they're not problems. People are not robots. People fuck up -- quite a bit actually. To you perfect people writing a reply to this boldly telling me that people shouldn't "fuck up", how many times did you have to use backspace in writing that response?
Point taken. You could call the BSD's Unix', though I wouldn't instult them by saying such. What I really meant to do was say, "I don't give a flying fuck about proprietary OS' -- period". GNU/Linux, BSD/Linux (there's a distro with BSD-tools), *BSD, and any other FS/OSS, however, I do care about. Yea, each GNU/Linux distro has it's own base. But all are basically the same. The install programs, package management schemes, and compile-level optimization are among the biggest differences. But WindowMaker 0.6 in Gentoo is the same as WindowMaker 0.6 in RedHat is the same as WindowMaker 0.6 in Debian.
Ok, so, I think everyone on/. knows that I like GNU/Linux. So, you expect that what follows is going to be a ranting rave about how much this book sucks, right? Wrong. This book is great, and here's why.
Many here have pointed out that alot of these very same problems exist elsewhere. Hidden files are a social-engineering security problem on Windows and Mac as well; likewise with undeleteable files.
So what? Saying, "well, their OS sucks too" doesn't make our OS any better. Since when is it ok for me to accept my own flaws just because everyone else around me also has those same flaws, or others?
The stuff written in this book shouldn't be seen as MS/Mac propaganda. I think most people who are going to be reading it are GNU/Linux users, and aren't going to be switching anytime soon, irrelevant of how much the authors hate *nix. (btw, if *nix sucks so much, why is Mac basing OSX around it, and why do we keep hearing rumors about MS doing such as well?).
There are many valid and important criticisms of *nix in that book. We should consider ourselves lucky that this book is narrowly targetted to *nix and doesn't address any of the same problems win Windows and MacOS -- we've received solid constructive criticism which others haven't, and that's a good thing.
Look, there's a lot of things that one can learn from reading this carefully, and considering it carefully.
Regarding the rm disasters, we need to lose this macho attitude of "tough shit" and make a solution, or implement a solution. I believe libtrash is a good one, which should work with any programs that use libc function.
Btw, I think even in Linux (I'm not going to talk about UNIX, because I don't give a flying fuck about proprietary crap like Slowlaris), rm does not actually destroy data. It simply marks that area on the HD as being "empty" and available, in that it does *not* write zeros over that area. This is something that has privacy experts in a panic, but I see it as no problem since there are shredder progs.
I'm not saying that overall Windows and Mac are better than Linux (btw, regarding Linux, almost all distributions are perfectly [or nearly so] cross-compatable). In most things, Windows and Mac simply suck -- namely, stability, security capability, and performance. However, GNU/Linux needs to be more concerned with accidental users mistakes, like typing 'rm * foo' instead of 'rm *.foo'.
I agree with the FSF's position on the LGPL in regards to licenses: you should only use it on libraries where there is already a functionally equivalent proprietary library. Otherwise, you should GPL the library to give FS/OSS developers and advantage over proprietary developers.
So what? That's their fucking problem. To me, $300 dollars is a lot of money to pay -- especially for crap technical support with mindless zombies on the other end reading from a cookbook. This is fraudulent advertising here. If they're going to do that, they should just say, "if you want real tech-support, it'll cost you $1000...otherwise, search our web-page".
Just because you can look at the code, doesn't mean you are going to.
No, but someone is going to. With millions of people using the software, it's almost a statistical certainty.
Wrong. The BSA can still audit you.
Let them. They can also audit my dog-house (if I had one, I hate pets [aside from snakes]). In an entirely FS/OSS environment, their audit's will be meaningless, requiring no defense. On the other hand, if you have proprietary software, you're in big trouble, because you have to provide proof that your licensed for every piece of proprietary software you have. Of course, the BSA would not target an entirely FS/OSS organization -- that would waste their money. That entire organization is criminal and corrupt anyways -- they participate in enterprise-level extortion and blackmail, and are granted priviledges analagous to those only law-enforcement officials should be granted.
Uhm, SCO ring a bell? SCO vs. IBM. Keep dreaming
God, you are a moron. That lawsuite is a pathetic joke. No-one anywhere takes that bullshit seriously. It is difficult to comprehend why SCO has made such stupid movies, but that seems to be their history. The only possible explanation is that they want to be bought out by IBM -- more enterprise-level extortion. That lawsuite won't be posing any problems for anyone anywhere. It's a fine example of where some secretary should have just read the thing and said, "That's fucking bullshit," and thrown it in the trash.
Right. If your experience with Oracle is that it is slow, than it is because you are an idiot.
That is the experience of many with Oracle in comparison to more specialized databases.
Uh, right. Buying a platinum support contract from Oracle is much cheaper than buying a few developers and have them implement the features Oracle has, right after they do a code audit of an open source system
Typical short-sighted thoughts of those in the industry thinking maybe 5 years ahead. In the long-run, it would be cheaper for companies needing a service like Oracle to band together and jointly pay for the development of such, as opposed to everyone paying Oracle continuously. Paying an external company for this software necessarily means they're (everyone who's paying for it cummulatively) paying for more than it costed to develop it, which means they're getting fucked over, compared to if they'd banded together and jointly funded it's production.
Regarding database clustering, I believe there are such enterprise-level features in the kernel-compilation program (menuconfig) for Linux 2.4. I'm on a university computer right now; when I get back to my box, I'll check it out.
Regarding the limiations of FS/OSS: yep, that's exactly why we need to continue using FS/OSS and not be tempted into using proprietary software. So that people feel the need to develop these solutions, because they need to use them.
I'm curious as to what Verizon has to gain here. They can't possibly be doing this out of principle. So I really don't see why they're doing it. Pursuing this is costing them millions, and making them powerful enemies. And it's not like most users have heard of this case -- so they're not gaining themselves a significant number of customers. I really don't see what they have to gain.
No worry of BSA-auditing and multimillion dollar extortion schemes.
No licensing headaches.
Infinite scaleability per each individually bought copy (as in, you can install an infinite number of copies with one purchased [or downloaded] CD).
Due to #4, ever-increasing savings as the number of computers onto which you install the software grows.
Assurance that the product will not die off simply because a company goes out of business, as it is FS/OSS. Any worthy project will be taken up by others if it's original developers move on.
Related to #6, ability to develop/implement your own features for your specific needs.
On another note regarding Oracle, it is basically slow crap. The executable alone is 18MB, so it naturally has poor performance; specialized database-systems will outperform it. Btw, data assurance from Oracle doesn't come for free. It costs quite a bit. And for that extra money you spend on it, it'd be better just spending that money doing an audit of FS/OSS code to insure that it won't lose data, and creating backup systems. Using journaling file systems like ReiserFS and XFS is also useful.
Enterprise != Personal systems.
Completely correct. The benefits of using FS/OSS at the enterprise level are even greater. Refer to the many research papers and discussions of companies saving millions by using GNU/Linux over Windows-2000/XP/2003. The MITRE study comes to mind:
http://www.egovos.org/pdf/dodfoss.pdf
This is a study funded by the government to get an objective evaluation; not some crackpot study funded by MS to make them look better.
Your $300 sale from Gateway doesn't mean shit. A $3M sale, does. They don't give a shit about you. Deal with it.
Firstly, this is irrelevant to the rest of the discussion. This was simply a personal digression of mine. The point was that you can get excellent technical support for free within a community of intelligent members.
If my $300 doesn't mean shit to Gateway, then they and every other OEM should stop their false advertising of "tech-support" -- because all they're doing is reading from a cookbook which we could have found online. Btw, I don't how many customers Gateway has. Let's say they have 1-million home-user customers, and each customer pays $100 for tech-support (these are obviously conservative numbers). That amounts to $100 million in tech support paid to Gateway by home-users. They damn well better care about the quality of tech support they're giving to home-users.
Lets see some open source clusters
Where have you been the last five years? Some of the world's most powerful supercomputers are Beowulf clusters, using GNU/Linux. See an O'Reilly article for an overview. In particular, GNU/Linux Beowulf clusters are being used for:
Yep, this FS/OSS stuff is really useless. It's only made the movie industry more money then from any other movie (see Titanic), assisted in the sequencing of the human genome, and assisted in the prediction of weather patterns, potentially saving lives.
What about SAN support?
Granted, I can not find any FS/OSS implementations at the moment, but there is commercial support available for GNU/Linux:
99% of everything Balmer has said is complete FUD.
No installation standards? Maybe not one universal standard -- but each distribution has its own standards. Gentoo has.ebuilds, Debian has.debs, Redhat has.rpms...all of this is pretty much automatic installation. Since one corporation is going to use one distribution, that doesn't really matter much. Furthermore, distributions are perfectly capable of sharing install utilities. The one thing I think would be good is if distributions worked on making sure that that could be done more seamlessly. This would only be relevant if a company was moving from one distribution to another, however.
Regarding a consistent GUI, however, he is right. The problem isn't, however, that there are different UI's. That's a good thing. The problem is that Motif applications look out of place in WindowMaker, for example. The problem is that on each individual user end, it is difficult for all applications to look consistent. This could be solved by completely divorcing content from appearance. The programmer should not be deciding how the program looks to the user. The program should specify, "Menu, with such and such items", and "toolbar, with such and such items", and then depending on the user's UI setup, it would display differently. This is sort of the idea for PicoGUI. IOW, an application written by a programmer would look NeXTish on Windowmaker, Motif'ish on a motif-WM, and Windowish on KDE and GNOME.
This would be a good thing, but would require programmers to collaborate on creating some kind of universal translator for menu => appearance, depending on the user preferences. Preferrably, this standard would be lightweight as possible, and allow for maximum integration between apps.
They want support, they want assurance, they want compliance with existing standards
All of which FS/OSS offers at competitve prices, which is much better deal than you can get from MS. Btw, you get no assurance with MS software -- all software licenses explicitly deny any assurance. So that's just fuzzy buzzword thinking on the part of stupid executives who don't really know wtf they're talking about.
As for support, that is purchased at competitive rates which are much better than anything you can get from MS. Furthermore, you'll get better support, precisely because there is competition. On a personal note, I get better support for free from Gentoo Forums than I get from Gateway for $300.
The benefits of using FS/OSS also scale very well, in that the more computers you use an FS/OSS product on, the more money you save, compared to using MS NT/2k/XP/2.003k. Oh yea, and there's also the fact that you don't have to worry about hundred-million dollar extortion-attempts from the BSA. These benefits -- though providing the most savings for large companies -- are extremely crucial for smaller companies.
*Note:There is also no point in continuing this threat if we have fundamentally different underlying ideologies as to what the ideal is for the end-goal of software. Any time two individuals have fundamentally different underlying values, it will be impossible for them to reach an agreement on any issue upon which they happen to disagree,
So I will state flat out that regarding software, I agree with the FSF on the most important end-goals of software (see the four essential freedom's mentioned below...skim to them). If you disagree with me that these four freedom's are essential, then all we have to talk about is the basis for that disagreement. This is not a way to close down meaningful debate on my part, but simply a way to place the debate in meaningful area.
The use of proprietary software for the end-user is no different then that of use of GPL software, yet you describe it as clubbing them over the head.
You obviously refused to read the text of the GPL I provided, and the link describing MS' latest EULA I provided. I cannot generalize between all EULA's, because they are extremely variable; however, the common feature is to deny the end-user various rights. Most of them specifically prevent multiple copies of the same program from being run on different computers, or multiple instances of it from running. They also tend to give MS or the respective copyright holder the right to termiante the user's license to use for any reason at any time.
Could you name maybe one project that "couldn't compete" becuase it was BSD?
I believe *BSD comes to mind, in that I believe WinNT/2k/XP has more server market share than it, despite being an inferior product. I'm not sure about this, however, but I seem to remember Windows having the largest server-share of any OS. Even if there isn't an example right now, the point is it *could* happen: that's bad enough.
I will admit that I was full of shit on guild-socialism being something you made up. However, I from it's definition -- industrial self-government through national worker-controlled guilds -- I fail to see how it relates much to socialism in general (which in practice had nothing to do with slf-government). It most certainly has nothing to do with the FSF or any other Free Software organization.
Well at least you admitted it when you said "The FSF specifically chose not to enumerate end-user rights"
Are you trying to purposefully mis the point? Consider all of the user-rights that would have to be enumerated for the list to be considered full. Your talking about hundreds -- possibly thousands -- of rights. And then if one right isn't enumerated, everone will panic and assume that that the user doesn't have that right. Consequently, the GPL specifically states that it places no restrictions on the use of the software by the user -- both for practical and ideological reasons.
Practically, it is impossible to fully enumerate all of the user's rights, partially because we can't even think of them all. Ideologically, if those rights were enumerated in the GPL, that would effectively be implying that any rights not enumerated are specifically denied, something which the FSF does not want to state. The FSF wants to make it clear that the GPL is *not* a EULA, and that they feel *no restrictions* should be placed on the user's use of the program. Thus, because the GPL places no restrictions on the user's use of the program (your example of the output only applies when the output is GPL'ed code, which only applies if redistributing).
You have either chosen not to read the references I gave you, or have patently lied in your response. I specifically referenced an article mentioning many restrictions which EULA's place on you. Since you obviously refused to RTFR, I will quote some relevant points regarding MS' WinXP EULA. For your convenience -- since you appear to have trouble distinguishing between things important to developers and things important to users -- I will enum
Yes, the BSD-license allows proprietary developers to club their users over the head. The GPL does not. One is anarchistic in that regard, the other is not.
If I release something BSD, someone can come along and change it so me or my friends don't have access to it any more?
What I'm saying is that you won't necessarily have access to the source of modifications/additions on your code. Whereas in the GPL, you will. This is why BSD-licensed software simply cannot compete with proprietary software -- because proprietary developers can simply incorporate everything into their own code, and return nothing; thus, always being ahead of FS/OSS developers.
Its guild socialism.
Lets not make up meaningless phrases here. Socialism means that everyone's forced to give what they've worked for and their creations to the community, whether they use community resource X or not. It also means you can't keep anything for yourself, privately. This is not the GPL. Giving aid to a neighbor on the condition that (s)he give aid to another when the other's in need is not socialism; it is simply being a good neighbor, and doing that which will make a good neighborhood. Kantianistic philosophy.
Activities other than copying, distribution and modification are not covered by this License; they are outside its scope. The act of running the Program is not restricted, and the output from the Program is covered only if its contents constitute a work based on the Program (independent of having been made by running the Program). Whether that is true depends on what the Program does.
In fact, users aren't even required to accept the GPL
5. You are not required to accept this License, since you have not signed it. However, nothing else grants you permission to modify or distribute the Program or its derivative works.
In other words, by default under the GPL, the end-user has every user right conceivable; no restrictions are placed on the use of GPL'ed software. The GPL guarantees that no restricitons be placed on the use of the software by including such a phrase. The FSF specifically chose not to enumerate end-user rights because that would require many more clauses, and because it would imply that the user does not have the right to use the software unless that right is granted. Also see clause #5: no-one is even required to accept the GPL.
Does the GPL guarantee that you get the software for free? No, it does not (in theory). So what. The FSF and FS advocates are not concerned with getting something for free. That is the slashdot gimme gimme gimme attitude. In practice, however, GPL'ed software tends to be either free or priced at very competitive and affordable prices. The GPL grants end-users many rights by default (that is, it does not place any restrictions on the user of the software) that EULAs deny (refer to/.'s reference to MS' EULA vs. the GPL.
zealot...RMS worshiper...upset your ideologies
Why is it that any time an individual has firm beliefs, (s)he's accused of being a zealot? Simply because I agree with RMS on a few issues does not mean I worship him. I disagree with him on a few issues too; see the Open Software License, and patent-arguments. I form my own opinions on various issues, and change them accordingly as new information and cases present themselves.
Re:Right tool for the job
on
Linus on DRM
·
· Score: 1
Please see The Tanenbaum-Torvalds Debate:
http://www.oreilly.com/catalog/opensources/book/ ap pa.html
Quite frankly, it has everything to do with monolithic vs. micro kernel.
Aside from that, you have the whole thing turned upside down. What makes Linux valuable is all the software around it -- a whole lot of which is GNU software. Without all that software around it, Linux is just a kernel: not particularly useful in and of itself.
The FSF and RMS were and are working on many things; not just the kernel. Linus and his group mainly just works on the kernel. Thus, it is quite easy to see how he could get this done quickly initially. Once he had a working prototype up, developers flocked to it, because it was there first. This is part of the reason why HURD has never taken off -- because developers went to something that worked, and didn't see the need to start over again.
But, according to you, the FSF is just a bunch of people so concerned with philosophy that their coders aren't actually productive. To say this ignores the fact that the FSF effectively built a house from the top down, rather rapidly. Tons of useful software -- I'll single out gcc and screen, the basis of ratpoison -- was developed by FSF developers.
Oh please. When you're done worshiping worshipping Linux Torvalds, please call us so that we can actually engage in an intelligent discussion.
No-one is perfect. No-one's even close. And everyone is fallible in some way or another. I hardly see how RMS, BP, or ESR have been corrupted. That is the most outrageous thing I've ever heard.
Is Stallman eccentric and bizarre and even annoying at times: Yes. Is Linus often times indifferent and dismissive of important political issues: Yes. Perens and Raymond have their own shortcomings, as well. So do I (I've been known to be a little hot-headed at times), and so do you, and so does everyone else.
That's why there is no "one true" leader of the OSS and FS movements (which are broadly overlapping).
As for the free/freedom/liberty thing, simply being a long-time member of Slashdot does not make you any more capable of forming judgements on issues than anyone else. Some would argue that there is an inverse relationship between time spent on Slashdot and intelligent thought.
I think we need to face up to the reality that there is no such thing as absolute freedom for human beings. The closest we can ever come to absolute freedom is if we're in total isolation, and can do whatever we want, and are limited only by physical and natural factors (even then, we're not totally free).
Once we're in social groups, simply having no restrictions on what we can do does not grant us absolute freedom. It's called anarchy.
It grants us freedom, until someone decides to come along and chain us to a stone while they rape our wives and children.
Freedom, until someone decides to cut us up into several dismembered pieces.
Freedom, until someone decides to steal the food we have, and our other posessions.
In short, any freedom granted by anarchy is temporary. I think that the potential for a similar situation exists with BSD-licensed software. Any proprietary developer can come along and make modifications and release the modified version under a proprietary license, which denies the users many rights. Because there is no requirement by the BSD for others to contribute source back, proprietary corporations could take from BSD projects and give nothing back: meaning that there's no way the FS / OSS project could *ever* compete with the proprietary one. This effectively means that the FS / OSS project will die off if any proprietary project decides to take such a tact.
The GPL prevents that. In preventing that, it denies developers the freedom to proprietize modifications/additions to GPL'ed code; but, by doing so, it ensure that end-users and other developers have the freedoms granted by the GPL. So, what we are really talking about here is rights. The GPL ensures that certain rights will be in place, and prevents anyone from taking away those rights (if they base their code off of GPL'ed code).
The entire purpose of the GPL is to ensure the end-user and developers certain freedoms (which RMS deemed important), regarding the use and distribution of GPL and GPL-based software. This requries restricting one freedom of developers, the freedom of developers to proprietize the source code on modifications; which RMS viewed as more of a harmful power, than a freedom.
Slashdot Mirror
The "are you sure y/n" method is ONLY useful if it is an uncommon message.
/, /boot, /home, /usr should contain such a prompt. The user deletes them with enough rarity that a y/n would be out of the ordinary. Likewise, the user should be able to specify some of his own files that (s)he wants to be prompted on (though, the user should be warned that placing this on every file is useless). I'd say about 1% is a good maximum.
I agree completely. A firealarm is of no use if it goes off all the time for no reason, because people will learn to ignore it. So, I agree with you, that it should be a rare occasion when you are prompted y/n. Directories like
This also illustrates why there should be a trash can.
No. The first purpose of the computer should be to make as MANY THINGS as possible, possible. Doing so in the easiest way is the second purpose. If optimizing for the simple case causes the complex case to fail, you shouldn't do it that way. Such is the way with "rm". The only way to make mistakes impossible with it is to reduce its utility.
Sorry, but this is just wrong-headed. A computer is a tool to get a job done (or rather, many jobs done). Overall, the best system is one which minimizes the time the user has to spend in front of it to get their work done. The idea is not to make mistakes impossible with rm: that would prevent sysadmins from deleting things they really wanted to delete; the idea is to make it a little harder to do so by accident. And -- preferrably -- to have some kind of trashcan, so that even if that happens, things are ok.
The other idea is that some folder's are so important that they should come with a tag by default requesting that rm prompt the user if (s)he really wants to remove them. Furthermore, user's should be able to place such a "confirm prompt" tag on any file/folder they deem that important.
With a computer interface, it's not the same. Even when the user does something strange, that might really be a valid action. You can't have the OS trying to out-guess the user. There might be a valid reason to rm -rf /.
This would be a valid point, except you don't consider the asymmetry of consequence here. Sure, there could be a valid reason for deleting the entire root partition. It is more likely, however, that the user just fucked up. So, for such important directories, it makes sense to place a tag on those directories making sure the user is prompted whether or not to delete them. If you really wanted to delete it, how much does it slow you down? Maybe a few seconds. If you didn't, it saves you a huge headache.
There is nothing new or innovative in this design. Simply taking a scroll wheel and extending it in a 2nd dimension does not count as something worthy of a patent. Neither does a scroll wheel.
I really tire quickly when people continue to defend shit.
/') a bad thing? Sure, it'll slow down that expert user who does it once a year by a few seconds. But it'll save others hours work from not having to restore backups.
/trash contents and prompting "empty? (YES/n)"? You need to consider that there is a *huge* asymmetry of consequence for many y/n actions -- especially "delete".
/usr/bin and check out the program link. Yep, shortcut's there. Then I check in where the program directory should be. Aha! Not there. So, now I go look for "program" in my backup archives. After a minute, I find it. Then I copy it over to where it should have been. Bingo! Problem solved. Test it out and it works. This took about another 2-5 minutes.
/bin. However, this was just a specific program directory, so it doesn't prompt me -- perfectly possible I really wanted to remove it. However, instead of deleting those files, it moves them to a trashcan. Upon logging out, a script lists the contents of the trashcan, and prompts me if I want to delete them (y/n) or cancel the logoug (esc, C^C, etc). The program does not accept input (other than esc) for 5sec, thus making it worth my while to at least look over what's there. As a security feature, it automatically logs out after 30seconds without deleting the
Let's take your butcher example. Yes, butcher knives are sharp. Yep, butchers can cut themselves. That does not, however, mean that those making them ignore safety. You notice how butcher knives have high-friction grips on them, right? Notice how the handles aren't made out of smooth mahogony polished so finely that it's slippery, right? Could that be good social design? Maybe? Just maybe?
How is prompting the user if he really wants to recursively delete the *entire* hard-drive ('rm -rf
if you're a good admin you've already planned for yourself fucking up
LOL, why is it that all of it falls on the user? According to you, good programmers don't plan on the user fucking up? What would be the harm in: (1) Placing destructive rm options far away from non-destructive ones; (2) Prompting on recursive deletes of important folders; (3) Instead of removing files, placing them in the trashcan, with an auto-script on logout listing
But let's just say I'm a good system administrator, and I'm using your rm program which assumes I never fuck up and blindly follows me off a cliff. But I do fuck up: about once a year, and I've planned for that by writing a script that auto-backs up things on a daily basis (backing up progressively in a diff-format, for performance). So, it's day 365 of the year, right? Time for me to fuck up. So, I fuck up and accidentally delete a binary directory. It doesn't contain much, just a rare-used program that gets used about every week or so. So, it takes about a week for me to realize something's wrong.
After a week, someone reports to me, "Hey, this fuckin' program isnt' working. I type "program", and it says "command not found": fix it asshole". This pisses me off. After all, this's the first time I've fucked up all year. How dare this fuck call me an asshole? So, I spend about 5 minutes seething in anger.
After I'm finished pouting, I grumpily try to figure out what's going on. You see, I don't necessarily know that the program's been deleted. So, I look in
So, I go back to said asshole who called me an asshole. I shortly say, "Problem solved". He grudgingly thanks me.
Now, before this prick interrupted me, I was doing something. But, my concentration was disrupted, and a period of high productivity on my part was ended. It takes me a minute to remember what I was doing, and when I get back to it, I'm not really in the zone. Plus, I'm still pissed off. Consequently, my productivity for the rest of the day falls 25%.
Now, let's consider another "rm" program written by some software guy names Joel. This rm program prompts me upon deleting important directories, like
Obviously the 85% posting mindless FPs and flames, or simply not posting at all, since most everyone here seems to hate MS Windows (or, gee...maybe they just come here more at work [where they use windows] than at home [where they use GNU/Linux]). ;-)
you give the computer a command it should execute it.
/"? Not very often at all. (in fact, in such a scenario, it would probably be more intelligent to simply reformat that partition). But, in 9/10 times where someone types 'rm -rf /boot', they really meant to type 'rm -rf ./boot'. Period. Now, wouldn't it make sense to have some way of specifying directories that you *don't* want to delete recursively, and prompting the rm-command to prompt hte user "are you sure" for such important directories? Also, libtrash as a default library to intercept such destructive commands and move things to a trashcan would be good
No. The purpose of the computer should be to on average allow the user to get work done as fast as possible. If -- because the computer stupidly executed a destructive typo -- it destroys a days worth of work, it has completely defeated it's purpose: to allow work to be done faster. Part of the way to resolve this is simply to have more intelligent command-line options. Command line options that perform destructive tasks should use letters that are *physically* *far* appart from the letters used to invoke harmless command-line options.
How often is someone going to 'rm -rf
User error is user error and not the computers fault, you are free to add checks but it doesn't make the statement made any less true.
Correct. A truly careless user will tend to fuck things up, even if you prompt him "really want to recursively delete entire home directory?" (shorter is better...the longer a message, the less likely the user will read it). However, you can at least put a speedbump along the road to oblivion. It might actually stop a semi-conscious user from deleting all their important info, and save them time. This is good.
Why? Because the best program that allows the user to get his/her job done the fastest. A program that increases the likelihood that the user will spend 5-10 minutes finding backups and restoring information is bad. Bad, because those 5-10 minutes are wasted. Then the user is pissed off. And his concentration is broken. Odds are, the rest of that users day are going to be unproductive. See "Joel on Software".
cryptic, and difficult to use for a reason
What reason would that be? So raving technocrats can scream, "RTFM, bitch whore!"? The GNU/Linux command line is like it is to allow for enormous power, allowing users to do things the program-creators never conceived of. That end-goal does not require that it be so easy for users to shoot themselves in the foot because they were a little bit sleepy one day.
No, I do not consider GNU/Linux to be "that hard". It's called hyperbolization. Iow, a joke. But the point is still there.
.foo' instead of 'ram *.foo'. They, however, do care about this.
.ebuild for that for Gentoo...I'll have to work on one).
How often have I accidentally deleted something I dind't want to? Well, very rarely, in fact. Within the last year or so, I can only remember a few cases. In each case, it was a minor file, and that was easily recovered from. Other's might not have been so lucky, or so good.
Even if you're the Stephen Hawking of CLIs, you're going to fuck up on rare occasion. It's a statistical certainty. Why? Because you aren't perfect. You aren't perfect, I'm not perfect, and the normal end-user actually using an account -- that guy who just uses the computer you administer -- sure as hell isn't anywhere near perfect.
The rm examples from this book are particularly good. Your counter-point that anyone doing anything important is going to be more careful than that just doesn't hold. (1) People use GNU/Linux for their home computers. Many of these people are not administrators. They may clobber important directories. (2) Ordinary users on the computers you administer have important files. I realize that you don't give a fuck about their files, as you are in no way responsible for them accidentally deleting that document they've been working on all day, or worse -- an entire directories worth of documents, by typing 'rm *
The simple fact is, many commands, utilities, and programs in GNU/Linux simply do not consider social engineering. (maybe there is a positive correlation between the anti-social nature of the program and the anti-social nature of the creator?) Some particularly poor design is when two command options -- one that performs a destructive activity, the other that performs a harmless activity -- either use the same key in lower/upper case, or two keys near eachother.
As an example of *good social design*, consider the command "perltrash". Perltrash has a -e and -l option. -e empties the trash. -l lists the contents of it. These keys are very far away from eachother on the keyboard. Now, they could have just as easily used -e for empty and -d for directory listing of trash. This would, however, be crappy social design. Why? Because it does not account for the fact that it is extremely easy to accidentally type an "e" instead of a "d". Additionally, the creator of perltrash did another correct thing by using -r for restore and -R for restore here. Earth-shattering! Using the lower- and upper-case versions of the same letter for a similar activity!
Regarding accidental deletion of files, the best thing to do would be to simply have libtrash (a prog which intercepts all libc delete/remove commands and changes them to move the info to the trashcan) installed by default on the supposedly "easy-to-use" GNU/Linux' like RedHat, Suse, Lindows, et al; and to have packages available for the more "make everything from scratch" distros, like Slackware, Debian, and Gentoo (I'm annoyed that there is no
There may not be any room for it -- but it *still* happens. Period. End of discussion. It's a statistical certainty. Even among sys admin root-users as nearly-perfect as (apparently) yourself will eventually fuck up. It is a statistical *certainty*.
Also, you seem to assume htat user's individual documents aren't important. Well, they are. And individual users certainly aren't as perfect as you.
When making something, it is simply *bad engineering* to not consider social phenomena. Technically, there would be nothing wrong with a monitor 40 inches wide and 5 inches high. But the monitor is still fucked up crappy design, because people do not have such a wide field of vision.
I'd really like to see you design airplaines. Maybe fighter planes. Just to make sure pilots are *extra careful*, the button for "drop bombs" should be right next to the button for "lower landing-gear". So what if they blow up the runway, right. Pilots are supposed to be perfect and should *not* fuck up.
The comment of one macho veteran Slashdotter have so annoyed me that I think it deserves being a main-comment for criticism:
There are no such things as "rm disasters". There are only mistakes, stop making them, or at least think before you execute.
Exactly the kind of bullshit macho attitude I was talking about.
Why don't you try doing that if you're a car company, and sell a car that can so easily be fucked up? Oh, yea, instead of having an out of-the-way hard break lever, we put a hard-break button right next to the defog button...but don't fucking bitch at us if you accidentally press the hard-break button (which is right next to the defog button) when trying to defog your windows, and your car spins around and crashes on an icy road.
Does that kind of bullshit macho attitude apply for companies making airplanes? When people making airplaies discovered that slats switches were being turned on accidentally, did they say:
"Yea, so what the slats extention switch can be accidentally turned on by an unintentional movement, possibly causing passenter-injury. Tell the pilots to be more careful and not fuck up."
No, they didn't. They said,
"Ok, so this is a problem. Why don't we cover the slats switch with a spherical clear cover that has to be unhinged before extending slats -- that way, they won't get extended at 500mph and cause the plane to trolly."
Just because many of these problems are socialogical not technological doesn't mean they're not problems. People are not robots. People fuck up -- quite a bit actually. To you perfect people writing a reply to this boldly telling me that people shouldn't "fuck up", how many times did you have to use backspace in writing that response?
The comment of one macho veteran Slashdotter have so annoyed me that I think it deserves being a main-comment for criticism:
There are no such things as "rm disasters". There are only mistakes, stop making them, or at least think before you execute.
Exactly the kind of bullshit macho attitude I was talking about.
Why don't you try doing that if you're a car company, and sell a car that can so easily be fucked up? Oh, yea, instead of having an out of-the-way hard break lever, we put a hard-break button right next to the defog button...but don't fucking bitch at us if you accidentally press the hard-break button (which is right next to the defog button) when trying to defog your windows, and your car spins around and crashes on an icy road.
Does that kind of bullshit macho attitude apply for companies making airplanes? When people making airplaies discovered that slats switches were being turned on accidentally, did they say:
"Yea, so what the slats extention switch can be accidentally turned on by an unintentional movement, possibly causing passenter-injury. Tell the pilots to be more careful and not fuck up."
No, they didn't. They said,
"Ok, so this is a problem. Why don't we cover the slats switch with a spherical clear cover that has to be unhinged before extending slats -- that way, they won't get extended at 500mph and cause the plane to trolly."
Just because many of these problems are socialogical not technological doesn't mean they're not problems. People are not robots. People fuck up -- quite a bit actually. To you perfect people writing a reply to this boldly telling me that people shouldn't "fuck up", how many times did you have to use backspace in writing that response?
Point taken. You could call the BSD's Unix', though I wouldn't instult them by saying such. What I really meant to do was say, "I don't give a flying fuck about proprietary OS' -- period". GNU/Linux, BSD/Linux (there's a distro with BSD-tools), *BSD, and any other FS/OSS, however, I do care about. Yea, each GNU/Linux distro has it's own base. But all are basically the same. The install programs, package management schemes, and compile-level optimization are among the biggest differences. But WindowMaker 0.6 in Gentoo is the same as WindowMaker 0.6 in RedHat is the same as WindowMaker 0.6 in Debian.
Ok, so, I think everyone on /. knows that I like GNU/Linux. So, you expect that what follows is going to be a ranting rave about how much this book sucks, right? Wrong. This book is great, and here's why.
Many here have pointed out that alot of these very same problems exist elsewhere. Hidden files are a social-engineering security problem on Windows and Mac as well; likewise with undeleteable files.
So what? Saying, "well, their OS sucks too" doesn't make our OS any better. Since when is it ok for me to accept my own flaws just because everyone else around me also has those same flaws, or others?
The stuff written in this book shouldn't be seen as MS/Mac propaganda. I think most people who are going to be reading it are GNU/Linux users, and aren't going to be switching anytime soon, irrelevant of how much the authors hate *nix. (btw, if *nix sucks so much, why is Mac basing OSX around it, and why do we keep hearing rumors about MS doing such as well?).
There are many valid and important criticisms of *nix in that book. We should consider ourselves lucky that this book is narrowly targetted to *nix and doesn't address any of the same problems win Windows and MacOS -- we've received solid constructive criticism which others haven't, and that's a good thing.
Look, there's a lot of things that one can learn from reading this carefully, and considering it carefully.
Regarding the rm disasters, we need to lose this macho attitude of "tough shit" and make a solution, or implement a solution. I believe libtrash is a good one, which should work with any programs that use libc function.
Btw, I think even in Linux (I'm not going to talk about UNIX, because I don't give a flying fuck about proprietary crap like Slowlaris), rm does not actually destroy data. It simply marks that area on the HD as being "empty" and available, in that it does *not* write zeros over that area. This is something that has privacy experts in a panic, but I see it as no problem since there are shredder progs.
I'm not saying that overall Windows and Mac are better than Linux (btw, regarding Linux, almost all distributions are perfectly [or nearly so] cross-compatable). In most things, Windows and Mac simply suck -- namely, stability, security capability, and performance. However, GNU/Linux needs to be more concerned with accidental users mistakes, like typing 'rm * foo' instead of 'rm *.foo'.
I agree with the FSF's position on the LGPL in regards to licenses: you should only use it on libraries where there is already a functionally equivalent proprietary library. Otherwise, you should GPL the library to give FS/OSS developers and advantage over proprietary developers.
So what? That's their fucking problem. To me, $300 dollars is a lot of money to pay -- especially for crap technical support with mindless zombies on the other end reading from a cookbook. This is fraudulent advertising here. If they're going to do that, they should just say, "if you want real tech-support, it'll cost you $1000...otherwise, search our web-page".
No, but someone is going to. With millions of people using the software, it's almost a statistical certainty.
Wrong. The BSA can still audit you.
Let them. They can also audit my dog-house (if I had one, I hate pets [aside from snakes]). In an entirely FS/OSS environment, their audit's will be meaningless, requiring no defense. On the other hand, if you have proprietary software, you're in big trouble, because you have to provide proof that your licensed for every piece of proprietary software you have. Of course, the BSA would not target an entirely FS/OSS organization -- that would waste their money. That entire organization is criminal and corrupt anyways -- they participate in enterprise-level extortion and blackmail, and are granted priviledges analagous to those only law-enforcement officials should be granted.
Uhm, SCO ring a bell? SCO vs. IBM. Keep dreaming
God, you are a moron. That lawsuite is a pathetic joke. No-one anywhere takes that bullshit seriously. It is difficult to comprehend why SCO has made such stupid movies, but that seems to be their history. The only possible explanation is that they want to be bought out by IBM -- more enterprise-level extortion. That lawsuite won't be posing any problems for anyone anywhere. It's a fine example of where some secretary should have just read the thing and said, "That's fucking bullshit," and thrown it in the trash.
Right. If your experience with Oracle is that it is slow, than it is because you are an idiot.
That is the experience of many with Oracle in comparison to more specialized databases.
Uh, right. Buying a platinum support contract from Oracle is much cheaper than buying a few developers and have them implement the features Oracle has, right after they do a code audit of an open source system
Typical short-sighted thoughts of those in the industry thinking maybe 5 years ahead. In the long-run, it would be cheaper for companies needing a service like Oracle to band together and jointly pay for the development of such, as opposed to everyone paying Oracle continuously. Paying an external company for this software necessarily means they're (everyone who's paying for it cummulatively) paying for more than it costed to develop it, which means they're getting fucked over, compared to if they'd banded together and jointly funded it's production.
Regarding database clustering, I believe there are such enterprise-level features in the kernel-compilation program (menuconfig) for Linux 2.4. I'm on a university computer right now; when I get back to my box, I'll check it out.
Regarding the limiations of FS/OSS: yep, that's exactly why we need to continue using FS/OSS and not be tempted into using proprietary software. So that people feel the need to develop these solutions, because they need to use them.
I'm curious as to what Verizon has to gain here. They can't possibly be doing this out of principle. So I really don't see why they're doing it. Pursuing this is costing them millions, and making them powerful enemies. And it's not like most users have heard of this case -- so they're not gaining themselves a significant number of customers. I really don't see what they have to gain.
A few examples:
On another note regarding Oracle, it is basically slow crap. The executable alone is 18MB, so it naturally has poor performance; specialized database-systems will outperform it. Btw, data assurance from Oracle doesn't come for free. It costs quite a bit. And for that extra money you spend on it, it'd be better just spending that money doing an audit of FS/OSS code to insure that it won't lose data, and creating backup systems. Using journaling file systems like ReiserFS and XFS is also useful.
Enterprise != Personal systems.
Completely correct. The benefits of using FS/OSS at the enterprise level are even greater. Refer to the many research papers and discussions of companies saving millions by using GNU/Linux over Windows-2000/XP/2003. The MITRE study comes to mind: http://www.egovos.org/pdf/dodfoss.pdf This is a study funded by the government to get an objective evaluation; not some crackpot study funded by MS to make them look better.
Your $300 sale from Gateway doesn't mean shit. A $3M sale, does. They don't give a shit about you. Deal with it. Firstly, this is irrelevant to the rest of the discussion. This was simply a personal digression of mine. The point was that you can get excellent technical support for free within a community of intelligent members. If my $300 doesn't mean shit to Gateway, then they and every other OEM should stop their false advertising of "tech-support" -- because all they're doing is reading from a cookbook which we could have found online. Btw, I don't how many customers Gateway has. Let's say they have 1-million home-user customers, and each customer pays $100 for tech-support (these are obviously conservative numbers). That amounts to $100 million in tech support paid to Gateway by home-users. They damn well better care about the quality of tech support they're giving to home-users.
Lets see some open source clusters
Where have you been the last five years? Some of the world's most powerful supercomputers are Beowulf clusters, using GNU/Linux. See an O'Reilly article for an overview. In particular, GNU/Linux Beowulf clusters are being used for:
Yep, this FS/OSS stuff is really useless. It's only made the movie industry more money then from any other movie (see Titanic), assisted in the sequencing of the human genome, and assisted in the prediction of weather patterns, potentially saving lives.
What about SAN support?
Granted, I can not find any FS/OSS implementations at the moment, but there is commercial support available for GNU/Linux:
99% of everything Balmer has said is complete FUD.
.ebuilds, Debian has .debs, Redhat has .rpms...all of this is pretty much automatic installation. Since one corporation is going to use one distribution, that doesn't really matter much. Furthermore, distributions are perfectly capable of sharing install utilities. The one thing I think would be good is if distributions worked on making sure that that could be done more seamlessly. This would only be relevant if a company was moving from one distribution to another, however.
No installation standards? Maybe not one universal standard -- but each distribution has its own standards. Gentoo has
Regarding a consistent GUI, however, he is right. The problem isn't, however, that there are different UI's. That's a good thing. The problem is that Motif applications look out of place in WindowMaker, for example. The problem is that on each individual user end, it is difficult for all applications to look consistent. This could be solved by completely divorcing content from appearance. The programmer should not be deciding how the program looks to the user. The program should specify, "Menu, with such and such items", and "toolbar, with such and such items", and then depending on the user's UI setup, it would display differently. This is sort of the idea for PicoGUI. IOW, an application written by a programmer would look NeXTish on Windowmaker, Motif'ish on a motif-WM, and Windowish on KDE and GNOME.
This would be a good thing, but would require programmers to collaborate on creating some kind of universal translator for menu => appearance, depending on the user preferences. Preferrably, this standard would be lightweight as possible, and allow for maximum integration between apps.
They want support, they want assurance, they want compliance with existing standards
All of which FS/OSS offers at competitve prices, which is much better deal than you can get from MS. Btw, you get no assurance with MS software -- all software licenses explicitly deny any assurance. So that's just fuzzy buzzword thinking on the part of stupid executives who don't really know wtf they're talking about.
As for support, that is purchased at competitive rates which are much better than anything you can get from MS. Furthermore, you'll get better support, precisely because there is competition. On a personal note, I get better support for free from Gentoo Forums than I get from Gateway for $300.
The benefits of using FS/OSS also scale very well, in that the more computers you use an FS/OSS product on, the more money you save, compared to using MS NT/2k/XP/2.003k. Oh yea, and there's also the fact that you don't have to worry about hundred-million dollar extortion-attempts from the BSA. These benefits -- though providing the most savings for large companies -- are extremely crucial for smaller companies.
So I will state flat out that regarding software, I agree with the FSF on the most important end-goals of software (see the four essential freedom's mentioned below...skim to them). If you disagree with me that these four freedom's are essential, then all we have to talk about is the basis for that disagreement. This is not a way to close down meaningful debate on my part, but simply a way to place the debate in meaningful area.
The use of proprietary software for the end-user is no different then that of use of GPL software, yet you describe it as clubbing them over the head.
You obviously refused to read the text of the GPL I provided, and the link describing MS' latest EULA I provided. I cannot generalize between all EULA's, because they are extremely variable; however, the common feature is to deny the end-user various rights. Most of them specifically prevent multiple copies of the same program from being run on different computers, or multiple instances of it from running. They also tend to give MS or the respective copyright holder the right to termiante the user's license to use for any reason at any time.
Could you name maybe one project that "couldn't compete" becuase it was BSD?
I believe *BSD comes to mind, in that I believe WinNT/2k/XP has more server market share than it, despite being an inferior product. I'm not sure about this, however, but I seem to remember Windows having the largest server-share of any OS. Even if there isn't an example right now, the point is it *could* happen: that's bad enough.
I will admit that I was full of shit on guild-socialism being something you made up. However, I from it's definition -- industrial self-government through national worker-controlled guilds -- I fail to see how it relates much to socialism in general (which in practice had nothing to do with slf-government). It most certainly has nothing to do with the FSF or any other Free Software organization. Well at least you admitted it when you said "The FSF specifically chose not to enumerate end-user rights"
Are you trying to purposefully mis the point? Consider all of the user-rights that would have to be enumerated for the list to be considered full. Your talking about hundreds -- possibly thousands -- of rights. And then if one right isn't enumerated, everone will panic and assume that that the user doesn't have that right. Consequently, the GPL specifically states that it places no restrictions on the use of the software by the user -- both for practical and ideological reasons.
Practically, it is impossible to fully enumerate all of the user's rights, partially because we can't even think of them all. Ideologically, if those rights were enumerated in the GPL, that would effectively be implying that any rights not enumerated are specifically denied, something which the FSF does not want to state. The FSF wants to make it clear that the GPL is *not* a EULA, and that they feel *no restrictions* should be placed on the user's use of the program. Thus, because the GPL places no restrictions on the user's use of the program (your example of the output only applies when the output is GPL'ed code, which only applies if redistributing).
You have either chosen not to read the references I gave you, or have patently lied in your response. I specifically referenced an article mentioning many restrictions which EULA's place on you. Since you obviously refused to RTFR, I will quote some relevant points regarding MS' WinXP EULA. For your convenience -- since you appear to have trouble distinguishing between things important to developers and things important to users -- I will enum
Yes, the BSD-license allows proprietary developers to club their users over the head. The GPL does not. One is anarchistic in that regard, the other is not.
If I release something BSD, someone can come along and change it so me or my friends don't have access to it any more?
What I'm saying is that you won't necessarily have access to the source of modifications/additions on your code. Whereas in the GPL, you will. This is why BSD-licensed software simply cannot compete with proprietary software -- because proprietary developers can simply incorporate everything into their own code, and return nothing; thus, always being ahead of FS/OSS developers.
Its guild socialism.
Lets not make up meaningless phrases here. Socialism means that everyone's forced to give what they've worked for and their creations to the community, whether they use community resource X or not. It also means you can't keep anything for yourself, privately. This is not the GPL. Giving aid to a neighbor on the condition that (s)he give aid to another when the other's in need is not socialism; it is simply being a good neighbor, and doing that which will make a good neighborhood. Kantianistic philosophy.
No, it doesn't ensure the end-user of anything.
Wrong. Please turn off your FUD-machine: GNU Public License:
Activities other than copying, distribution and modification are not covered by this License; they are outside its scope. The act of running the Program is not restricted, and the output from the Program is covered only if its contents constitute a work based on the Program (independent of having been made by running the Program). Whether that is true depends on what the Program does.
In fact, users aren't even required to accept the GPL
5. You are not required to accept this License, since you have not signed it. However, nothing else grants you permission to modify or distribute the Program or its derivative works.
In other words, by default under the GPL, the end-user has every user right conceivable; no restrictions are placed on the use of GPL'ed software. The GPL guarantees that no restricitons be placed on the use of the software by including such a phrase. The FSF specifically chose not to enumerate end-user rights because that would require many more clauses, and because it would imply that the user does not have the right to use the software unless that right is granted. Also see clause #5: no-one is even required to accept the GPL.
Does the GPL guarantee that you get the software for free? No, it does not (in theory). So what. The FSF and FS advocates are not concerned with getting something for free. That is the slashdot gimme gimme gimme attitude. In practice, however, GPL'ed software tends to be either free or priced at very competitive and affordable prices. The GPL grants end-users many rights by default (that is, it does not place any restrictions on the user of the software) that EULAs deny (refer to /.'s reference to MS' EULA vs. the GPL.
zealot...RMS worshiper...upset your ideologies
Why is it that any time an individual has firm beliefs, (s)he's accused of being a zealot? Simply because I agree with RMS on a few issues does not mean I worship him. I disagree with him on a few issues too; see the Open Software License, and patent-arguments. I form my own opinions on various issues, and change them accordingly as new information and cases present themselves.
Please see The Tanenbaum-Torvalds Debate:
/ ap pa.html
http://www.oreilly.com/catalog/opensources/book
Quite frankly, it has everything to do with monolithic vs. micro kernel.
Aside from that, you have the whole thing turned upside down. What makes Linux valuable is all the software around it -- a whole lot of which is GNU software. Without all that software around it, Linux is just a kernel: not particularly useful in and of itself.
The FSF and RMS were and are working on many things; not just the kernel. Linus and his group mainly just works on the kernel. Thus, it is quite easy to see how he could get this done quickly initially. Once he had a working prototype up, developers flocked to it, because it was there first. This is part of the reason why HURD has never taken off -- because developers went to something that worked, and didn't see the need to start over again.
But, according to you, the FSF is just a bunch of people so concerned with philosophy that their coders aren't actually productive. To say this ignores the fact that the FSF effectively built a house from the top down, rather rapidly. Tons of useful software -- I'll single out gcc and screen, the basis of ratpoison -- was developed by FSF developers.
Oh please. When you're done worshiping worshipping Linux Torvalds, please call us so that we can actually engage in an intelligent discussion.
No-one is perfect. No-one's even close. And everyone is fallible in some way or another. I hardly see how RMS, BP, or ESR have been corrupted. That is the most outrageous thing I've ever heard.
Is Stallman eccentric and bizarre and even annoying at times: Yes. Is Linus often times indifferent and dismissive of important political issues: Yes. Perens and Raymond have their own shortcomings, as well. So do I (I've been known to be a little hot-headed at times), and so do you, and so does everyone else.
That's why there is no "one true" leader of the OSS and FS movements (which are broadly overlapping).
As for the free/freedom/liberty thing, simply being a long-time member of Slashdot does not make you any more capable of forming judgements on issues than anyone else. Some would argue that there is an inverse relationship between time spent on Slashdot and intelligent thought.
I think we need to face up to the reality that there is no such thing as absolute freedom for human beings. The closest we can ever come to absolute freedom is if we're in total isolation, and can do whatever we want, and are limited only by physical and natural factors (even then, we're not totally free).
Once we're in social groups, simply having no restrictions on what we can do does not grant us absolute freedom. It's called anarchy.
It grants us freedom, until someone decides to come along and chain us to a stone while they rape our wives and children.
Freedom, until someone decides to cut us up into several dismembered pieces.
Freedom, until someone decides to steal the food we have, and our other posessions.
In short, any freedom granted by anarchy is temporary. I think that the potential for a similar situation exists with BSD-licensed software. Any proprietary developer can come along and make modifications and release the modified version under a proprietary license, which denies the users many rights. Because there is no requirement by the BSD for others to contribute source back, proprietary corporations could take from BSD projects and give nothing back: meaning that there's no way the FS / OSS project could *ever* compete with the proprietary one. This effectively means that the FS / OSS project will die off if any proprietary project decides to take such a tact.
The GPL prevents that. In preventing that, it denies developers the freedom to proprietize modifications/additions to GPL'ed code; but, by doing so, it ensure that end-users and other developers have the freedoms granted by the GPL. So, what we are really talking about here is rights. The GPL ensures that certain rights will be in place, and prevents anyone from taking away those rights (if they base their code off of GPL'ed code).
The entire purpose of the GPL is to ensure the end-user and developers certain freedoms (which RMS deemed important), regarding the use and distribution of GPL and GPL-based software. This requries restricting one freedom of developers, the freedom of developers to proprietize the source code on modifications; which RMS viewed as more of a harmful power, than a freedom.