Filesystems are sexy. Developing a successful one will make you famous, being a contributor not nearly so much.
Filesystems are specific pieces of software that are carefully optimized and have to be designed with tradeoffs in mind. You can't just go and add bits of ZFS into say, ReiserFS. I'm sure it would have been a far from trivial undertaking, Hans would have disagreed a lot back when he was free, it probably wouldn't be doable elegantly, perhaps performance wouldn't be good, the kernel developers wouldn't accept the resulting mess, etc, etc.
Then not nearly everybody agrees on what is the best filesystem. I for one don't like the approach of sticking everything into ZFS and would prefer to keep my RAID and LVM separate, thank you very much. On a longer term I'd like to see things like Z-RAID integrated into the Linux RAID, and be available to all filesystems.
If it is possible for Nvidia to use binary blobs for their graphics cards, it should be possible to use CDDL code with a compatibility shim in the Linux kernel. All this duplicated effort could instead be focused on one project, thereby having an all around better file system. btrFS has just recently started coming into fruition, would that time not be better spent improving ZFS?
Eww. That sort of thing needs to die, badly. Do you really want a kernel that the kernel developers will refuse to troubleshoot because your filesystem is always loaded, thus always tainted? Are you going to run your servers on that?
There's no rush. I will wait for a GPLd, properly integrated into the kernel, supported by the developers version. ZFS may be cool but cooler things will come after it.
ext2 is nice and simple, but it's neither fast not reliable. It uses a linear search to find directory entries, which means it's very slow on large directories, like Maildir mailboxes. It doesn't do tail packing which means it wastes space and is slower with small files. It's not reliable because without a journal it needs a fsck after a bad shutdown which takes ages on a modern disk, and recovers it worse than a journal would.
Just search for benchmarks, something like reiserfs beats ext2 by huge margins when it comes to important workloads such as a mail server.
There are very good reasons why distributions generally go with ext3, or one of the other filesystems. I haven't seen ext2 as the default option for the root FS in a very long time.
If you go to a restaurant, you don't have to tip your waiter anything. Yet many people would feel bad if the waiter was decent and they left nothing.
No, being a decent waiter is the waiter's job, and they don't deserve any extra for doing what they're supposed to do. If they don't earn enough then they should go and demand more from their employer.
A tip is not a tip if it's mandatory. A tip is used to reward exceptional service. Simply serving food or mechanically pouring a beer or cup of coffee is not exceptional service. Now, when a taxi driver spent a while figuring out how to fit the large CRT I had bought into the car and helped me carry it up, that was very exceptional service and appropiately rewarded.
That's easy, so long it's a normal link and not some weird JS thing, I can middle click the link and open it in a new tab. That way I get to choose whether I want to replace the current page with a new one or not.
And why should the form get erased? My browser (konqueror) certainly doesn't do that.
If you can prove to me why we need door locks as a society, then I can prove to you why we need AV.
AV != door lock. AV == defense inside the house which can only work after somebody got through your door or window. A door is a static defense. If it's any good, it does its job. An antivirus is an active, imperfect defense (like a dog). It may fail to recognize a new virus, or the signatures may not get updated.
The self-healing processing object I was talking about is YOUR BODY. It's a processing unit, and it's self-healing in most instances. Something drastic can break that, but that goes for all self-healing objects. Look at the concept of a cocoon for sudden failures on high speed equipment. Catastrophic failures are different from minute failures.
I was imagining you'd bring this up eventually, but the human body and a computer are completely different things. A computer is a static chunk of plastic, metal and silicon. A human is a constantly changing thing that inhabits a dangerous and unpredictable medium and wouldn't even work without all those self-repair systems.
But still, while it's really good that a broken bone will heal, any plan that results in me breaking a leg is a bad plan to start with.
As for the boot time system check, agreed, it doesn't run at every boot, and I never insinuated that, but by default it IS turned on to run every so many boots. And I disagree that any fs, journaled or not (ext3 and reiser are journaling, no?), never needs to be checked for consistency. I strongly deny that if that is your stance. All filesystems need to be checked at some point. But if you think that your system never runs the potential for corruption, that's just silly. It's based on the influence of magnets and electricity, ain't it?
On a properly made computer, all the possible damage is already accounted for.
The RAM uses ECC and can self-correct when a little bad, or immediately tell the OS it's not good if it's really bad.
The hard disk uses ECC internally and knows when a sector is bad, and tells the OS it is.
The bus between the hard disk and the motherboard uses a CRC to detect transmission errors.
The PCI bus detects errors as well.
Magnets are not an issue. Open a hard drive, there's a very strong magnet in the voice coil, and magnets in the motor. It takes a *very* strong magnet to mess with a hard disk, to the point that erasing hard drives without removing the casing in a domestic environment isn't possible.
The filesystem itself is coded to notice weird data on disk and tell you when something isn't right.
Ok, pass the joint, or the bong, or whatever it is. You want to get rid of self-healing processing units? Ok, then my suggestion is to go find and android case quickly, and move all your permanent data files over. Get rid of that self-healing processing unit that you're using to read this and to type that snarky reply.
Then I guess we don't use the same definition of "self-healing". There's nothing in my hardware that heals itself that I'm aware of. Hardware is at best redundant. The only exception to this might be ECC RAM with scrubbing enabled, which isn't common on consumer hardware.
Dude, the whole point of engineering is to build systems that are self-healing. If you think you can design a system that is impervious to damage, let me know, but most engineers try to think of how to make their systems self-healing.
How is a bridge self-healing? They simply build it strong enough to stand the intended load plus a marging. A bridge is fortunately not an assembly that is constantly developing cracks that get filled by strategically placed capsules of materials and maintained by soldering robots, but a mostly static chunk of metal and concrete that was built in the first place to require little maintenance.
Have you seen the work being done in Plastics that are being made to be self-healing?
Yes, it's for very specialized applications. Which consumer hardware isn't.
So you deny that the base install using ext3fs of most modern distros will run a fsck like procedure on a drive every so many boots? Why do you think the modern system designers who are responsible for overseeing these builds continue to include that feature? Do you really think you know more than a good fs designer, or are you one, and you're just not admitting it yet?
ext3 does, reiserfs doesn't. The reason fsck.ext3 doesn't run after every bad shutdown is because the system is capable of dealing with sudden poweroffs. It was designed for that case specifically in fact, because fsck takes ages. The reason it does run ocassionally is not because the filesystem needs it, but because it provides insurance against kernel bugs and disk problems. You could disable it and your system would remain perfectly functional.
You are correct, I did mix metaphors. But the point of locked doors versus open doors is still somewhat valid, as most who don't run any AV still don't run a firewall, regardless of XP SP2, and most who don't run a firewall still run AV, so your analogy of a dog inside is more apt.
No, look, a firewall isn't a protection against viruses. It's a protection against worms maybe, and it's not even needed on a correctly configured system, which doesn't have vulnerable services running to start with.
And I repeat that an antivirus is a completely flawed way of dealing with a problem that shouldn't have existed in the first place.
Ok, what are you smoking? Every MS O/S I've run since Win98SED (except ME) has not had a general problem with dislocating things on it's own.
I guess, but what comes with the OS isn't terribly useful. Things start going downhill pretty fast once you start making the box actually do something.
It's only when you start adding in 3rd party stuff that I've ever had problems with my box. And I'VE (note the subject, not other people, myself) never had to reformat my own box due to Windows corruption. If you're truly IT or sysadmin, then you shouldn't have had to ever reformat your own box either. QED, Windows doesn't break on it's own. If you have had that happen, spell it out for me.
Then you're a competent user who manually takes measures the OS should have made unnecessary, such as not clicking on.scr attachments.
And defrags and disk scans are not the realm of sysadmin's only, or why would they be in XP Home, or Vista Home?
Because the OS isn't good enough to make them unnecessary for a normal user to think about. Why should I worried about defragmenting my hard disk? The OS should be either ensuring it doesn't happen (not possible 100%, I know), or invisibly defragging on the background. Manual intervention should be reserved for special cases such as if a sysadmin really wants to optimize the placement of a database on disk or something of the sort.
Methinks you think the user is too unskilled. And the last I checked, every so often when I boot my system into Linux it runs fsck automatically, (yeah, it can be reconfigured on the number of reboots inbetween) but it does that because you MUST check your fs every so often, no matter what system you're running on. I've seen ext3 have crosslinks on clusters, but only once or twice, and that several years ago on an abused box.
I'm not talking about the user, I'm talking about the system. I believe than any system that needs healing (which I understand as self-repair, as in fixing after something breaks) is flawed and should be replaced by one that doesn't need it.
Your example about ext3 for instance. Ext2 was a fragile filesystem in the style of VFAT which *needed* fsck to fix it after a bad shutdown. Ext3 doesn't and recovers gracefully on its own. fsck is still included for cases such as bad disks, bad RAM and kernel bugs, but it's not really necessary for a system that's operating properly. Most users should never see fsck fix anything on an ext3 disk.
Now, as for your point that viruses and rootkits and worms etc shouldn't exist, I agree, but just because nobody has broken into your house, does that mean you shouldn't have any sort of security system? Where do you live, I would love a place I can walk into without a key and take what I like. Since computers must allow programs to run (hence the reason MS has started both asking for signed applications/drivers and the reason for UAC) then viruses et al are a necessary worry, hence the needs for AV et al.
You're mixing things up. The good security system prevent from things getting into your house in the first place. Antiviruses are on the other hand akin to a dog inside the house that can only act once somebody has already broken in and very possibly caused damage. The intruder shouldn't have been able to get this far in the first place, and a scary dog hidden inside the house isn't a fix for a lack of a lock on the door.
As far as Office including VBA, you've obviously never used it in a business setting. It far outweighs not having VBA available. I for one am happy it's there.
Way to miss the point. There's nothing wrong with VBA in concept. There's something very wrong with the fact that you can write a virus in it, however. VBA should have been a sandboxed language unable to do anything worse than messing with the document it's in.
You're assuming that Windows will break on it's own, but I'm assuming that MS is trying to create a product that is robust and attempts to stand up to outside attacks.
Not assuming, it does. And if it didn't break, it wouldn't need a tool to fix it.
Antiviruses and spyware removers are things that shouldn't have to exist. Things should never get to the point where they're needed in the first place. Take MS Office viruses for instance. Just whose bright idea it was to actually make the language support the features needed for replication?
Disk defragmenters and filesystem checkers should only exist as very specialized tools, invoked by a sysadmin in special cases like recovering data from a disk with bad sectors, not by a normal user as a sort of periodic maintenance, or even by the OS to routinely fix that which should have never been broken (for defrag it should just be an automatic background operation)
NTFS does not remain consistent without needing an external tool to fix it. Hello chkdsk.
chkdsk isn't needed for normal operation. It's not needed to fix the FS after the power goes out.
A filesystem checker isn't guaranteed to fix your FS successfully. Those who used Windows 9x enough know that scandisk often couldn't fix things 100% because there wasn't a way for it to find what the correct state of the system was. For instance, if you get crosslinks (two files that incorrectly share clusters, with the effect of changing one overwriting part of the other), scandisk can only go as far as removing the crosslink, but one or both files still end up with screwed up data, and there's no way for it to fix that.
The only reason for tools like scandisk and chkdsk to exist is to handle those cases the system can't handle automatically by itself, because there's no 100% guarantee of recovery.
The solution to problems like these is not having something that fixes it, but making sure it doesn't happen in the first place.
Wouldn't it be easier if the system as delivered by the manufacturer would automatically heal itself and just allow you to use the computer to "work"?
Certainly not, "healing" implies something broke and now needs fixing. It'd be a lot easier if the system didn't break in the first place.
"healing" sounds to me much like Win98 running scandisk after a bad shutdown: the FS can't deal with it, so it's got to be fixed. The right way is the way NTFS does it: it remains consistent without needing an external tool fix it.
I Download Ubuntu and use it as a desktop system. Chances are that I am going to be abiding by the GPL.
The GPL doesn't even apply to this, as it says absolutely nothing about usage, and is not an EULA. The GPL only comes into play if you're going to redistribute a modified version of Ubuntu.
Picture quality is more about the lens than the resolution. With the tiny lens on all those devices, if you could put a 20 MP pixel, the pictures would still come out horrible.
I heard this claim several times already, but never seen an explanation. As far as I can tell, he's a pretty smart guy and what he says seems to make sense.
Why not? You're the one responding. A troll is only allowed to continue when allowed. And you're allowing it.
A troll is only successful if it gets a good reaction. On my part I'm just relaxedly sipping tea here and waiting for a compile to finish. Got to pass the time somehow.
And, I'd think you would agree that we need a killfile akin to Usenet. There are trolls abound (well, me) and a shared plonka would be the last bastion that separates usenet from these damned forums.
There is one, actually, all I have to do is to mark you as a foe, and problem solved. Additionally it'll lower your score for most people who friended me as well.
I dont think you'd appreciate me doing this over at perlmonks, nor would I. They offer some damned good help:) I know you over there, as a completely different non-trolling nick.
You must be confusing me with somebody else, I haven't been on perlmonks in years and never did anything of importance there.
As a last recommendation, use different usernames on different sites. I also recommend setting up your own mailserver that allows unlimited user creation. vadim_t (Aght) teleline (dought) es looks to be your email, given gentoo is correct. And no, I wouldnt attack your email. That's just too.. uncivilized.
Like I was saying, you must be confusing me with somebody else. There's more than one vadim on the net.
After all, what exactly is a label of "troll"? Someone who you dont agree with?
No, somebody who pretends to hold an opinion in order to provoke a reaction. Most of the time online, because in real life that kind of behavior would get you quickly excluded from the group, fired, or punched in the face.
In that regard, I am doing a community a favor. I'm not filling links to goatse, gnaa, fecaljapan, nor am I trying to overtly hack the html bounds engine (page lengthening posts).
Wow, how utterly gracious of you. Why, you're probably also doing other favours, such as by not farting in public and spitting at random people. The people around you sure must be feeling fortunate.
And I'm pulling in the posts and argument, so I should even get paid for my services. Without argument and diatribe, nobody would stick around, given the bad articles and the stupid editors. I mean, the editors expect that we go through and rate everybody else without verification (have they fixed the overrated "bug"?), metamoderation to fix the bad moderations, and general groupthink. We see these articles that even newspapers have already printed, let alone Digg and the rest.
Ah yes, the "I'm such a wonderful gift to humanity" argument, which every troll makes. My guess is that this is another example of the cognitive dissonance somebody explained in another article. "I think of myself as a decent person" and "I behave like an asshole online" are a bit self-conflicting. The way is of course to imagine you're actually doing people a favor.
Sorry to disappoint you, but the website would do perfectly fine without your helpful services. I belong to another community where haven't had somebody like you in maybe 3 years, and we're doing perfectly fine so far.
Yeah, all those you fell for the obvious troll aren't that clever.
But you're just beyond pathetic, because obviously you're doing this because there's nothing more exciting in your life (my condolences), and because you think think that finding a few people among thousands on the site that fall for your stupidity is something to be proud of.
How about the real question: how well does the iPhone framework work for developing applications? I've heard it's very nice, and very similar to desktop Mac programming so it's an easy transition for Mac developers. How nice is the Android setup? It it easier/harder to make simple applications? More complex things?
Well, see, for me it doesn't even make sense get to this part. It doesn't matter how nice the SDK might be when the reward for spending a not that small amount of money on the reqired hardware and the subscription, and weeks or months of my time on development could be having my application removed from the store, and Apple actually forbidding me from telling my customers what happened.
Now when Apple stops being stupid, then I will become interested in comparing them on their technical development merits.
It really illustrates how dialog boxes as a warning system are a flawed mechanic, we got this fancy computer with a fancy operating system, why can't it figure out the right thing to do when an application tries to access memory it's not supposed to?
Which other than aborting the application and informing the user would be what, exactly?
Once this point is reached there's absolutely nothing else to do. The other alternative is ignoring it, but it's not going to do any good. Infinite loops and horrible data corruption seem like the most likely consequences of making that choice.
If you or someone else wants to fork Wikipedia, and create a version that allows unverifiable material, go right ahead. Let me know how it works out - considering how much people criticise Wikipedia for any false statement that briefly appears, I can't wait to see what they make of unverifiable-pedia:)
Ok, first of all, this discussion is getting a bit out of track. My original point wasn't that Wikipedia should allow unverifiable material, but that much of what gets deleted is perfectly salvageable. My point wasn't that I want that precise content on Wikipedia, but that I think that most of those articles would be worth having, even the most obscure bands, so long their existence is verifiable of course.
Many of the articles didn't even assert their importance, so we have no reason to think that sources could be found in future.
Aha, another point of disagreement. I simply don't think this is needed. If it exists, is verifiable, and relevant to people outside the immediate circle of family and friends, it should be documented.
And how would such information be verified, in these cases?
The way it currently is.
Furthermore, even when articles have existed and been tagged for months or years, with no sign of improvement, you still seem to oppose deletion.
I will admit that I didn't have the time when I wrote the post to throughly check the entries. In my experience stuff normally gets deleted a lot faster.
The problem is that without speedy delete, Wikipedia would become overwhelmed with completely made up or unverifiable crap. How long should joke pages be kept up, just in case someone might produce reliable sources for? It'd become a laughing stock. As I say, if you want to create unverifiable-pedia, I'll be curious to see how it turns out.
Strawman. I'm not arguing speedy delete should be removed. Though IMO it's applied far too widely. Actively edited and popular pages shouldn't be getting listed there. Pages that are clearly salvageable shouldn't be getting listed there. I recently reloaded the "nonsense pages for deletion" page. Two entries: conversation fillers, which is not nonsense, and a very badly formatted entry that seems to be about a movie, but still doesn't fit in the category.
I'd argue that it would also make things worse for the inclusionists - with more AfDs, there'll be less editors looking at any given AfD, meaning it much more likely that things are deleted for dubious reasons, just because a few people don't like it. What Wikipedia needs is a way to streamline deletion for obviously crap articles, whilst focusing people's time on debating the less obvious cases.
Doesn't follow. Most people who protest against deletionism don't want to get rid of speedy delete per se. They're against excessive deletions period. They want less nominations for deletion, which reduces the size of those queues.
Well, the difference here of course is that you're replying following the current Wikipedia guidelines, while I'm not.
I'm of the "extreme inclusionism" sort. My test for notability would be "Does anybody besides the immediate family, friends, or members of the organization have a reason to look this up?". Thus, if an organization exists that did something the public could know about, it's notable. If an actor exists that participated in a movie the public could watch, it's notable. If a band released music the public could listen, it's notable. And so on.
Yes, that would result in an awful lot of articles. I vastly prefer this, over the current system where an article like Hispalinux get deleted as not notable. This is a convention that had thousands of attendants, got mentioned in spanish newspapers, and included talks by people like Richard Stallman, Marcelo Tossatti and Miguel de Icaza. It'd say it definitely passes the "pokemon notability test"
The other requisite is that an article can be written beyond the point of "X exists". My minimum requirement would be "Information above what you can obtain from a typical mention". Hence all the "expand" qualifications.
AZCOM - You would seriously keep this? It's all very well saying "Should be improved", but who will improve it? It's completely made up for all we know. If you want to improve it, there's nothing stopping you writing the article properly - delete just gets rid of the crap that was put there for now.
Wouldn't keep keep it as-is of course. First would add a notice asking for references, then nominate for deletion if none appear. But for articles like this I'd probably just batch create the whole bunch of them from some public database.
Pianist - agreed it should clearly go, but it's not clear why this is significantly different to the rock bands, for example?
The difference is that no useful information is provided. If somebody is reading an article that says that Joe played the piano at a concert, checks wikipedia, and finds that yep, Joe is a pianist, that's not exactly adding anything new. Now, the bands seem to have precisely the information I'd be interested in when looking up a band -- who are they, who are the members, what they made, what's their music's style.
Dancer - It's all very well saying "assuming it's true", but that's the point, we don't know.
By saying "assuming it's true" I mean I'd add a "citation needed" instead of deleting the article.
Do you think that an article for someone's website should be kept, or is there some reason why it is different?
Generally, no. Following the above criteria of "Information above what you can obtain from a typical mention", a website is its own documentation in a large part. For a website my criteria would be "Can you do better than repeating obvious facts and content from it?"
Joe's homepage already says it's Joe's homepage, no need to document that. Adding that Joe posted lots of photos of his cat doesn't help either, anybody can see that as well.
On the other hand, many websites like Slashdot and Kuro5hin are not self-describing, so they get a page.
I do hope you're not suggesting that pages of pets, webpages, random lists don't exist simply because they didn't turn up in the first 10 of a random search!
Certainly not, I'm suggesting that such things aren't all that common, and that much of what gets deleted isn't pages about somebody's cat.
I've got some more hypothetical articles - what would your view be on these?:
Keep, if more information can be had than that, and evidence of it. "X is a programmer and wrote Y" provides no useful information that you couldn't already find from the About box of the application. I certainly wouldn't see anything wrong with you having an article listing y
Of the 10, I'd only delete two of them, one for not containing anything useful (Khatia), and one for being original research. Wouldn't hurry with the deletion of any of them, I'd wait a couple of weeks to see if the article improves.
No articles on some random person, their cat, webpage, bad poetry, or a "List of Slashdot users with the letter 'i' in their name" so far.
Don't even try to touch anything related to Ron Paul, furries (they're by far the worst, there's actually a userbox, essentially a "I have an interest in X subject/belong to X community" template, with no other purpose than to discuss the sexual fetish of the user and this will never ever be deleted despite a crackdown on much more tame userboxes), obscure fantasy, or any other subject that attracts the middle-upper class, white, slightly-nerdy biased editor pool. Oh, or anything that breaks precedent. If there's a List of Warhammer Characters (A-Z) that is unsourced and breaks every guideline, that article will be given a magical exception because it's been there a while and a lot of people like it.
What you don't seem to understand is that all this stuff is precisely the reason why many people like Wikipedia. Huge parts of the population find a wonderfully formatted and researched article about World War II to be terribly boring. But they would love a huge collection of pages on Pokemon.
What I *want* in wikipedia is precisely lots and lots of information on obscure topics. For me this includes the furry fandom (not a purely sexual fetish, btw), otherkin (not a member but find it interesting), obscure fantasy, warhammer, rock bands nobody heard of, and so on.
I don't need a resource that tells me about something popular. If I wanted to know who Hitler was I could go look in the encyclopedia on the shelf. What I want is a resource that explains to me why the heck mudkips are so referenced all over the web all of a sudden, and what is a Commissar in the Warhammer universe. This may not fit with some of the administration's delusions of grandeur, but it's precisely what a large part of the users want.
Not a joke. It's just a pitfall when benchmarking.
Let's say you benchmark how much power is expended during indexing documents, for 10 minutes. This is a bad benchmark.
First you test with a hard disk, which spends a large amount of time seeking, during which the CPU sleeps waiting for data to arrive. Then you test with a SSD, which has near instant seeking, leaving the CPU very little time to sleep.
Then you look at the results: The HDD laptop consumed 100 units of power during 10 minutes. The SSD laptop consumed 120 units of power. But you forget to notice that the HDD laptop processed 10000 files, while the SSD managed to do 20000.
The problem is that you have to benchmark the right thing. If you're going to have a time limit you have to account for the amount of data processed during that time. A much better test would be measuring how much power it takes to perform a given task to completion with a SSD and an HDD, if total power usage is what you're interested in.
This can have perverse results as well. If for some strange reason you have a laptop that is continuously doing background processing whose completion far exceeds any possible battery time, and whose result you're not interested in, then you could well get less battery time on a SSD. For instance this could be the case for a laptop with a very bad spyware infestation.
Slashdot Mirror
Why? Many reasons.
Filesystems are sexy. Developing a successful one will make you famous, being a contributor not nearly so much.
Filesystems are specific pieces of software that are carefully optimized and have to be designed with tradeoffs in mind. You can't just go and add bits of ZFS into say, ReiserFS. I'm sure it would have been a far from trivial undertaking, Hans would have disagreed a lot back when he was free, it probably wouldn't be doable elegantly, perhaps performance wouldn't be good, the kernel developers wouldn't accept the resulting mess, etc, etc.
Then not nearly everybody agrees on what is the best filesystem. I for one don't like the approach of sticking everything into ZFS and would prefer to keep my RAID and LVM separate, thank you very much. On a longer term I'd like to see things like Z-RAID integrated into the Linux RAID, and be available to all filesystems.
Eww. That sort of thing needs to die, badly. Do you really want a kernel that the kernel developers will refuse to troubleshoot because your filesystem is always loaded, thus always tainted? Are you going to run your servers on that?
There's no rush. I will wait for a GPLd, properly integrated into the kernel, supported by the developers version. ZFS may be cool but cooler things will come after it.
I hope you're joking.
ext2 is nice and simple, but it's neither fast not reliable. It uses a linear search to find directory entries, which means it's very slow on large directories, like Maildir mailboxes. It doesn't do tail packing which means it wastes space and is slower with small files. It's not reliable because without a journal it needs a fsck after a bad shutdown which takes ages on a modern disk, and recovers it worse than a journal would.
Just search for benchmarks, something like reiserfs beats ext2 by huge margins when it comes to important workloads such as a mail server.
There are very good reasons why distributions generally go with ext3, or one of the other filesystems. I haven't seen ext2 as the default option for the root FS in a very long time.
Can you give some examples?
No, being a decent waiter is the waiter's job, and they don't deserve any extra for doing what they're supposed to do. If they don't earn enough then they should go and demand more from their employer.
A tip is not a tip if it's mandatory. A tip is used to reward exceptional service. Simply serving food or mechanically pouring a beer or cup of coffee is not exceptional service. Now, when a taxi driver spent a while figuring out how to fit the large CRT I had bought into the car and helped me carry it up, that was very exceptional service and appropiately rewarded.
That's easy, so long it's a normal link and not some weird JS thing, I can middle click the link and open it in a new tab. That way I get to choose whether I want to replace the current page with a new one or not.
And why should the form get erased? My browser (konqueror) certainly doesn't do that.
AV != door lock. AV == defense inside the house which can only work after somebody got through your door or window. A door is a static defense. If it's any good, it does its job. An antivirus is an active, imperfect defense (like a dog). It may fail to recognize a new virus, or the signatures may not get updated.
I was imagining you'd bring this up eventually, but the human body and a computer are completely different things. A computer is a static chunk of plastic, metal and silicon. A human is a constantly changing thing that inhabits a dangerous and unpredictable medium and wouldn't even work without all those self-repair systems.
But still, while it's really good that a broken bone will heal, any plan that results in me breaking a leg is a bad plan to start with.
On a properly made computer, all the possible damage is already accounted for.
The RAM uses ECC and can self-correct when a little bad, or immediately tell the OS it's not good if it's really bad.
The hard disk uses ECC internally and knows when a sector is bad, and tells the OS it is.
The bus between the hard disk and the motherboard uses a CRC to detect transmission errors.
The PCI bus detects errors as well.
Magnets are not an issue. Open a hard drive, there's a very strong magnet in the voice coil, and magnets in the motor. It takes a *very* strong magnet to mess with a hard disk, to the point that erasing hard drives without removing the casing in a domestic environment isn't possible.
The filesystem itself is coded to notice weird data on disk and tell you when something isn't right.
Then I guess we don't use the same definition of "self-healing". There's nothing in my hardware that heals itself that I'm aware of. Hardware is at best redundant. The only exception to this might be ECC RAM with scrubbing enabled, which isn't common on consumer hardware.
How is a bridge self-healing? They simply build it strong enough to stand the intended load plus a marging. A bridge is fortunately not an assembly that is constantly developing cracks that get filled by strategically placed capsules of materials and maintained by soldering robots, but a mostly static chunk of metal and concrete that was built in the first place to require little maintenance.
Yes, it's for very specialized applications. Which consumer hardware isn't.
ext3 does, reiserfs doesn't. The reason fsck.ext3 doesn't run after every bad shutdown is because the system is capable of dealing with sudden poweroffs. It was designed for that case specifically in fact, because fsck takes ages. The reason it does run ocassionally is not because the filesystem needs it, but because it provides insurance against kernel bugs and disk problems. You could disable it and your system would remain perfectly functional.
No, look, a firewall isn't a protection against viruses. It's a protection against worms maybe, and it's not even needed on a correctly configured system, which doesn't have vulnerable services running to start with.
And I repeat that an antivirus is a completely flawed way of dealing with a problem that shouldn't have existed in the first place.
Now SELinux is going in the right direction.
I guess, but what comes with the OS isn't terribly useful. Things start going downhill pretty fast once you start making the box actually do something.
Then you're a competent user who manually takes measures the OS should have made unnecessary, such as not clicking on .scr attachments.
Because the OS isn't good enough to make them unnecessary for a normal user to think about. Why should I worried about defragmenting my hard disk? The OS should be either ensuring it doesn't happen (not possible 100%, I know), or invisibly defragging on the background. Manual intervention should be reserved for special cases such as if a sysadmin really wants to optimize the placement of a database on disk or something of the sort.
I'm not talking about the user, I'm talking about the system. I believe than any system that needs healing (which I understand as self-repair, as in fixing after something breaks) is flawed and should be replaced by one that doesn't need it.
Your example about ext3 for instance. Ext2 was a fragile filesystem in the style of VFAT which *needed* fsck to fix it after a bad shutdown. Ext3 doesn't and recovers gracefully on its own. fsck is still included for cases such as bad disks, bad RAM and kernel bugs, but it's not really necessary for a system that's operating properly. Most users should never see fsck fix anything on an ext3 disk.
You're mixing things up. The good security system prevent from things getting into your house in the first place. Antiviruses are on the other hand akin to a dog inside the house that can only act once somebody has already broken in and very possibly caused damage. The intruder shouldn't have been able to get this far in the first place, and a scary dog hidden inside the house isn't a fix for a lack of a lock on the door.
Way to miss the point. There's nothing wrong with VBA in concept. There's something very wrong with the fact that you can write a virus in it, however. VBA should have been a sandboxed language unable to do anything worse than messing with the document it's in.
Not assuming, it does. And if it didn't break, it wouldn't need a tool to fix it.
Antiviruses and spyware removers are things that shouldn't have to exist. Things should never get to the point where they're needed in the first place. Take MS Office viruses for instance. Just whose bright idea it was to actually make the language support the features needed for replication?
Disk defragmenters and filesystem checkers should only exist as very specialized tools, invoked by a sysadmin in special cases like recovering data from a disk with bad sectors, not by a normal user as a sort of periodic maintenance, or even by the OS to routinely fix that which should have never been broken (for defrag it should just be an automatic background operation)
chkdsk isn't needed for normal operation. It's not needed to fix the FS after the power goes out.
A filesystem checker isn't guaranteed to fix your FS successfully. Those who used Windows 9x enough know that scandisk often couldn't fix things 100% because there wasn't a way for it to find what the correct state of the system was. For instance, if you get crosslinks (two files that incorrectly share clusters, with the effect of changing one overwriting part of the other), scandisk can only go as far as removing the crosslink, but one or both files still end up with screwed up data, and there's no way for it to fix that.
The only reason for tools like scandisk and chkdsk to exist is to handle those cases the system can't handle automatically by itself, because there's no 100% guarantee of recovery.
The solution to problems like these is not having something that fixes it, but making sure it doesn't happen in the first place.
Certainly not, "healing" implies something broke and now needs fixing. It'd be a lot easier if the system didn't break in the first place.
"healing" sounds to me much like Win98 running scandisk after a bad shutdown: the FS can't deal with it, so it's got to be fixed.
The right way is the way NTFS does it: it remains consistent without needing an external tool fix it.
The GPL doesn't even apply to this, as it says absolutely nothing about usage, and is not an EULA. The GPL only comes into play if you're going to redistribute a modified version of Ubuntu.
Fortunately, replacement ones are available.
Picture quality is more about the lens than the resolution. With the tiny lens on all those devices, if you could put a 20 MP pixel, the pictures would still come out horrible.
Why?
I heard this claim several times already, but never seen an explanation. As far as I can tell, he's a pretty smart guy and what he says seems to make sense.
So what's the problem with him?
A troll is only successful if it gets a good reaction. On my part I'm just relaxedly sipping tea here and waiting for a compile to finish. Got to pass the time somehow.
There is one, actually, all I have to do is to mark you as a foe, and problem solved. Additionally it'll lower your score for most people who friended me as well.
You must be confusing me with somebody else, I haven't been on perlmonks in years and never did anything of importance there.
Like I was saying, you must be confusing me with somebody else. There's more than one vadim on the net.
No, somebody who pretends to hold an opinion in order to provoke a reaction. Most of the time online, because in real life that kind of behavior would get you quickly excluded from the group, fired, or punched in the face.
Wow, how utterly gracious of you. Why, you're probably also doing other favours, such as by not farting in public and spitting at random people. The people around you sure must be feeling fortunate.
Ah yes, the "I'm such a wonderful gift to humanity" argument, which every troll makes. My guess is that this is another example of the cognitive dissonance somebody explained in another article. "I think of myself as a decent person" and "I behave like an asshole online" are a bit self-conflicting. The way is of course to imagine you're actually doing people a favor.
Sorry to disappoint you, but the website would do perfectly fine without your helpful services. I belong to another community where haven't had somebody like you in maybe 3 years, and we're doing perfectly fine so far.
I have no clue what are you so proud of.
Yeah, all those you fell for the obvious troll aren't that clever.
But you're just beyond pathetic, because obviously you're doing this because there's nothing more exciting in your life (my condolences), and because you think think that finding a few people among thousands on the site that fall for your stupidity is something to be proud of.
Well, see, for me it doesn't even make sense get to this part. It doesn't matter how nice the SDK might be when the reward for spending a not that small amount of money on the reqired hardware and the subscription, and weeks or months of my time on development could be having my application removed from the store, and Apple actually forbidding me from telling my customers what happened.
Now when Apple stops being stupid, then I will become interested in comparing them on their technical development merits.
Which other than aborting the application and informing the user would be what, exactly?
Once this point is reached there's absolutely nothing else to do. The other alternative is ignoring it, but it's not going to do any good. Infinite loops and horrible data corruption seem like the most likely consequences of making that choice.
Ok, first of all, this discussion is getting a bit out of track. My original point wasn't that Wikipedia should allow unverifiable material, but that much of what gets deleted is perfectly salvageable. My point wasn't that I want that precise content on Wikipedia, but that I think that most of those articles would be worth having, even the most obscure bands, so long their existence is verifiable of course.
Aha, another point of disagreement. I simply don't think this is needed. If it exists, is verifiable, and relevant to people outside the immediate circle of family and friends, it should be documented.
The way it currently is.
I will admit that I didn't have the time when I wrote the post to throughly check the entries. In my experience stuff normally gets deleted a lot faster.
Strawman. I'm not arguing speedy delete should be removed. Though IMO it's applied far too widely. Actively edited and popular pages shouldn't be getting listed there. Pages that are clearly salvageable shouldn't be getting listed there. I recently reloaded the "nonsense pages for deletion" page. Two entries: conversation fillers, which is not nonsense, and a very badly formatted entry that seems to be about a movie, but still doesn't fit in the category.
Doesn't follow. Most people who protest against deletionism don't want to get rid of speedy delete per se. They're against excessive deletions period. They want less nominations for deletion, which reduces the size of those queues.
Well, the difference here of course is that you're replying following the current Wikipedia guidelines, while I'm not.
I'm of the "extreme inclusionism" sort. My test for notability would be "Does anybody besides the immediate family, friends, or members of the organization have a reason to look this up?". Thus, if an organization exists that did something the public could know about, it's notable. If an actor exists that participated in a movie the public could watch, it's notable. If a band released music the public could listen, it's notable. And so on.
Yes, that would result in an awful lot of articles. I vastly prefer this, over the current system where an article like Hispalinux get deleted as not notable. This is a convention that had thousands of attendants, got mentioned in spanish newspapers, and included talks by people like Richard Stallman, Marcelo Tossatti and Miguel de Icaza. It'd say it definitely passes the "pokemon notability test"
The other requisite is that an article can be written beyond the point of "X exists". My minimum requirement would be "Information above what you can obtain from a typical mention". Hence all the "expand" qualifications.
Wouldn't keep keep it as-is of course. First would add a notice asking for references, then nominate for deletion if none appear. But for articles like this I'd probably just batch create the whole bunch of them from some public database.
The difference is that no useful information is provided. If somebody is reading an article that says that Joe played the piano at a concert, checks wikipedia, and finds that yep, Joe is a pianist, that's not exactly adding anything new. Now, the bands seem to have precisely the information I'd be interested in when looking up a band -- who are they, who are the members, what they made, what's their music's style.
By saying "assuming it's true" I mean I'd add a "citation needed" instead of deleting the article.
Generally, no. Following the above criteria of "Information above what you can obtain from a typical mention", a website is its own documentation in a large part. For a website my criteria would be "Can you do better than repeating obvious facts and content from it?"
Joe's homepage already says it's Joe's homepage, no need to document that. Adding that Joe posted lots of photos of his cat doesn't help either, anybody can see that as well.
On the other hand, many websites like Slashdot and Kuro5hin are not self-describing, so they get a page.
Certainly not, I'm suggesting that such things aren't all that common, and that much of what gets deleted isn't pages about somebody's cat.
Keep, if more information can be had than that, and evidence of it. "X is a programmer and wrote Y" provides no useful information that you couldn't already find from the About box of the application. I certainly wouldn't see anything wrong with you having an article listing y
Sounds like a nice test. I decided to try. Here are my opinions based on 10 loads of the link:
Club. Would keep.
Actor. Would expand and keep.
Company. Would expand and keep.
Rock band. Would keep it.
Some company. Should be improved, but would keep it.
Rock group. Would keep it.
Association. Would keep it. If such an association exists I don't see why not document it, however fringe it might be.
Self care. Would agree with deletion based on original research, though the concept is probably documented somewhere and could be fleshed out.
Pianist, with no useful data. Probably least worthy of keeping of the bunch.
Dancer. Lacks sources, but assuming the real world claims are true, I'd keep it
Of the 10, I'd only delete two of them, one for not containing anything useful (Khatia), and one for being original research. Wouldn't hurry with the deletion of any of them, I'd wait a couple of weeks to see if the article improves.
No articles on some random person, their cat, webpage, bad poetry, or a "List of Slashdot users with the letter 'i' in their name" so far.
What you don't seem to understand is that all this stuff is precisely the reason why many people like Wikipedia. Huge parts of the population find a wonderfully formatted and researched article about World War II to be terribly boring. But they would love a huge collection of pages on Pokemon.
What I *want* in wikipedia is precisely lots and lots of information on obscure topics. For me this includes the furry fandom (not a purely sexual fetish, btw), otherkin (not a member but find it interesting), obscure fantasy, warhammer, rock bands nobody heard of, and so on.
I don't need a resource that tells me about something popular. If I wanted to know who Hitler was I could go look in the encyclopedia on the shelf. What I want is a resource that explains to me why the heck mudkips are so referenced all over the web all of a sudden, and what is a Commissar in the Warhammer universe. This may not fit with some of the administration's delusions of grandeur, but it's precisely what a large part of the users want.
Now that will be AWESOME, even though a yiffy yak sounds like a very scary thing.
Not a joke. It's just a pitfall when benchmarking.
Let's say you benchmark how much power is expended during indexing documents, for 10 minutes. This is a bad benchmark.
First you test with a hard disk, which spends a large amount of time seeking, during which the CPU sleeps waiting for data to arrive.
Then you test with a SSD, which has near instant seeking, leaving the CPU very little time to sleep.
Then you look at the results: The HDD laptop consumed 100 units of power during 10 minutes. The SSD laptop consumed 120 units of power. But you forget to notice that the HDD laptop processed 10000 files, while the SSD managed to do 20000.
The problem is that you have to benchmark the right thing. If you're going to have a time limit you have to account for the amount of data processed during that time. A much better test would be measuring how much power it takes to perform a given task to completion with a SSD and an HDD, if total power usage is what you're interested in.
This can have perverse results as well. If for some strange reason you have a laptop that is continuously doing background processing whose completion far exceeds any possible battery time, and whose result you're not interested in, then you could well get less battery time on a SSD. For instance this could be the case for a laptop with a very bad spyware infestation.