If your device is lost or stolen, data should not be permanently deleted, just locked away until the owner personally comes round to identify herself with a passport or other legal ID of some sort. You can more to permanent delete after some time has passed without a "restore" request.
From an enterprise security point of view, once the device is out of your hands you want the data off it, full stop. If it isn't there then there is no chance that someone can read it. If everything on the device were properly encrypted, then you could just delete any keys and the restore would simply mean putting the keys back on.
I don't see why this should be any problem at all; Apple, Google and all their competitors claim to keep backups, which is effectively the same but with a user-"controlled" restore procedure.
That is the solution, not "not deleting". The off-device backups are your restore point either if you get a new device or that one is returned to you. As long, of course, as the backup account is not compromised at the same time as the device. No matter how securely you store you keys/tokens most phones are unlocked by a four digit pin so you've got not more than two days before someone brute forces that and gets in if they are determined and start when they first get hold of the device (so make sure if you lose the device that all the authentication credentials for the backups are changed ASAP).
Of course most stolen phones just get factory wiped before being fenced anyway, as most thefts of such devices are opportunistic rather than planned, so this is only a concern if someone might specifically target you (such as if others in your company's industry might want to have a peak at some significant trade secret) or if you have something really objectionable on the device (at which point if the thief notices it that can blackmail you)- most people like you or I are unlikely to be targeted in that way.
Having to collect the files together, waiting for that last one of the number you need to recreate the original content because for some reason it hasn't made it to your server's feed yet (or it has expired on your server and you need to wait for a repost), not having the content nicely indexed on the tracker we site, and so on, is more hassle than dealing with torrents.
OK, there are sites that do some of that indexing (but they are potentially subject to takedowns and DoS attacks as much as torrent indexes are), and there are clients that automate the getting of all the parts and unpacking them (as long as the original uploaded has prepared them properly), but at very least you need to research which client to use which is more hassle than just keeping using the torrent client you are already familiar with.
Just because something isn't difficult, that doesn't mean doing it isn't more hassle than what you do now.
Oh, and to get a good fast news feed with decent retention periods on binary groups you are going to have to pay. Not much unless you download silly amounts, but more than the nothing most torrent trackers cost to use.
No, but:
* It is a deliberate attempt to cause offence by bringing up a subject that the recipient is likely to be pretty sensitive about
* Threat of drowning is, and that was contained in another of the tweets, so he wasn't arrested just because of the "dad" tweet
In this case I'm not unhappy that the tweeter's freedom of speech has been trumped by the recipients freedom to not have to put up with that sort of shit from the petulant little prick, and I fail to see the logic of extending this to "when will they come after me?!?!?!?!?!?" (as some people are doing) unless of course you are also in the habit of sending offensive and threatening messages to people you don't even know.
What we have here is yet another instance of http://www.penny-arcade.com/comic/2004/03/19/ - he thought he was safe pretending to be big and clever with his tweets until it became clear that people were going to fight back and that he was as risk of losing his anonymity so peple who know him in real life can see him for what he really is. The threat to Sky News says as much about his world view (i.e. the world should let him do what he wants with no consequences) that much else.
Any website that tries to, yes.
Not any website that doesn't try, obviously. But those that don't try now could start trying at some point either through direct malice, being hacked, or carrying content served by 3rd parties (advertisers, stats collectors, servers providing public copies of common libraries,...) who get hacked.
You have no idea how many people I've talked to who seem to believe RAID is a fancy backup mechanism. I know of one small business owner in particular who had RAID 5 on his office computer, but no backup regimen. He didn't even know that the first drive had failed till he had lost all his business records. People ask about it regularly, with no apparent knowledge of its inherent limitations. They seem to think it's magic (like "the cloud").
Aye. Many expect it all to magically work without monitoring. There are many people out there with Linux software RAID setup (which I use in a few places) but without any monitoring or alerts setup because the recipe they copied didn't mention the need and they didn't think further than "RAID makes me safe". At very least have either smartctl or madam (preferably both) mailing you when there is any sort of issue with the drives and arrays, but people don't think "but how would I know if something has failed? How would I know a drive needs replacing?".
Maybe I'm predisposed to see this particular lunacy everywhere.
That is healthy, especially where us humans are concerned. A pessimist is never disappointed, and more often pleasantly surprised than an optimist!
You might be interested in some of the data integrity features in ZFS
I am. I'm also given to understand that BTRFS is barking up that particular tree.
I believe it is though I've not kept a close eye on it. ZFS is far more mature at this point (IIRC BTRFS is still officially "experimental", not even declared "beta" yet, and ZFS is considered production-ready in some environments (though not Linux)). Last I heard there are licensing issues that stop ZFS proper being ported over to Linux directly, there are various ways to use it (via FUSE is one example I remember though that seems far from ideal IMO) though I don't know how stable they are considered. BTRFS is in part a reaction to the feature set of ZFS (i.e. wanting some of those features without the licensing issues) though I think the Linux fs developers had some of the plans formulating before ZFS hit the scene in a big way.
I'm a little skeptical of mashing all of this into the file system level, but I'm not sure my fears are rational. It just feels like it ought to be easier to detect and fix problems if there were a separate RAID layer
I think part of the rational for keeping it all together is efficiency. Two separate layers that may not be aware of what each is doing can easily be very sub-optimal. Even if you get everything aligned perfectly and the filesystem is aware of your array's stripe sizes, there are still extra data copy operations going on. ZFS (and BTRFS) does a lot that would traditionally be elsewhere (like dealing with different storage types for performance, like moving things between SSD and spinning metal as needed (or using SSDs as cache rather than separate fast areas that add to the total space available).
RAID does provides some redundancy, but it is NOT a backup method.
I didn't say it was a backup method. RAID provides availability through redundancy (if the array succeeds in protecting you from a single drive failure, it stays up and can be repaired with very little downtime (or no downtime with hot-swap) so you don't jhave to wait how-ever long for your backups to restore (nor do you need to worry about work that might be lost that was done between now and the last backup) before getting back to work. In the last six months I've had drives go back in two locations and I was saved much hassle in both cases by them being members of RAID1 arrays.
RAID is primarily for performance. It only adds a smidgen of reliability.
Only RAID0 is about performance. Other arrangements can offer improved performance for some I/O patterns but availability through redundancy is the main aim of them. In fact for write heavy patterns, particularly for lots of small writes, levels 5 and 6 can impose a significant performance penalty.
I keep waiting for someone to come out with a RAID 6 ECC standard. Yes, It would be slow, but it would be reliable. Sometimes that's desirable. I think some of the expensive RAID cards can do this, though I've never seen one.
You might be interested in some of the data integrity features in ZFS (see http://en.wikipedia.org/wiki/Zfs#Data_Integrity as a starting point, if you are not already aware of it). I've not been in a position to use/test/benchmark it myself, but the descriptions of what it does (checksums of all blocks, data or meta, with self-healing from other array members as needed) certainly make it sound most impressive in that regard.
The key there, though, is it go and try to do a restore from your backup every now and then to make sure you can. There are few things worse than feeling secure in your backup process only to find it was not running correctly all along
A key point that many people miss until it is too late. Depending on your backup solution, automated testing is possible. Some of our Live databases get transferred back from the backup site and restored on another machine, then if that fails or the data in the restored copy seems to old (for that daily backed up DBs the test is that the last audit trail entry is no more than 24 hours after the last one in Live), an email is pinged out to several people so the situation gets looked at. That doesn't protect against everything (it is feasible that corruption in the backup file won't be picked up by the checksums used by the DBMS as it restores, and our last-audited-action check would still pass in most cases where that to happen) of course...
Similarly my home mail/calendar server gets its backup restored to a copy in a VM. I don't have automated checks on that though, I must do something about that omission soon (I currently manually log in occasionally to check things look right).
Your answer: correct.
But... your answer: pointless.
His answer: correct.
Though backups are not your only line of defense, redundancy is one too. For data that you care about you need both. Redundancy (via RAID) protects you from single (or more depending on arrangement) device failure (though not all failure routes so don't go feeling too safe!), backups protect you from catastrophic failure or an error (human or otherwise) that causes lost or corrupt data (if you delete everything, your RAID array will have redundant copies of nothing and your backups are what will save you unless they are stale).
That may make the drive work again, but it won't recover the data that was already on it (which it what they are asking about here).
The reason what you did works in some circumstances is that all drives beyond a certain point have more capacity than they actually let the outside world see. The extra capacity is used to remap back blocks to when they appear: meaning a multi Gbyte drive doesn't fail because a few MByte worth of surface goes bad. This only works until the "hidden" capacity is all used, after which there is nowhere to map bad sectors to. One of the readings usually offered by SMART is how many sectors have been reallocated this way: if that goes up too fast or nears full the drive will report itself as near to failing or in need of immediate replacement. If it is just one or two small spots that have failed due to some impurity in the platter, then a drive won't fail for writing BUT may still fail for reading: if the sector goes completely bad between being written to and read there is nothing the drive can do: it can't read the data to write it elsewhere, it will retry the read a number of times and if that fails you get a permanent error reported (if it succeeds the drive can remap the sector and all you see is an unusually slow read). Your drive will have had several sectors that went bad fast, so could not be remapped before the data was unreadable, by repeated writing you forced the drive to reconsider every sector and remap those ones that were iffy. You were lucky: there were enough sectors going spare for that to work and what-ever caused the bad sectors in the first places was not something that was drive-wide so no more failures occurred.
lt;dr: Yep, that sometimes works, but it doesn't get you any lost data back and I personally would never trust the drive again. The fact that this can work, and sometimes only temporarily as the drive my degrade further, is why I never but a drive from eBay or other second hand sources...
Notice how Valve is already running for the exits?
Not quite. They are setting up an extra stall outside, in the expectation that a fair few users will make for the exists. They are increasing their potential total market, not leaving one subset of the market in favour of another.
These people hate and fear change, and so will latch onto nearly any noticeable differences, so I'm thinking in terms of both front end functionality and the look of the interface.
You will not find a perfect Windows interface on Linux, so with that sort of person don't try. They will complain no matter what. They will complain if you move to Windows 8 too, because it is different too.
Instead of trying the impossible (pleasing the terminally unpleased), instead concentrate on picking something that is intuitive enough, that you can easily support (remotely if needed), and that works in your environment more-or-less out of the box.
As others have pointed out, don't change everything at once. If the tools you intend to use are cross-platform, introduce them under Windows first. That way you will take a few relatively small productivity hits as people get used to the new tools then get used to the new OS running them, instead of one huge productivity hit when everything changes at once.
Yet , those same people will say they love OS X, and they don't realize that its still Linux
It isn't. It is Unix derived (BSD mainly at the core) and uses a fair chunk of the GNU tool chain, but Linux it isn't. Not that the average use knows (or cares about) the difference.
Also what is it with cars feeling the need to overtake. Even if you're bobbling along in traffic at the same speed as the traffic, some lunatic will often try to overtake.
They feel they are losing time if they are behind or otherwise at the same speed as a cyclist. Some people perpetually in a rush, some others are just small-dicked wonders who use speed as a compensating device. In either case I'd actually prefer they get past me than stay behind: in front of me I can see what they are doing better, behind me they are a less well seen danger.
What annoys me sometimes is people slamming their foot down to overtake when I indicate that I intend to turn right (which will involve crossing the main body of the traffic lane) even when there is initially enough space between us and the opposing lane is empty so that they'd not even have to slow down at all as I safely cross their path with them at their original speed. Once one dick does that everyone behind keeps as close as they can so I end up sat by the side of the road for ages waiting for a clear spot in the flow to cross through (or, if I'm feeling petty and I know I can catch them up at the next lights, I'll get in front of the dick and set off very very very slowly when the lights change...).
If you want to be pulled over cycling in Cambridge, go through a red light by Parker's Piece.
Really? I've never seen anyone stopped for running a red light on a bike by the police.
I was once pulled over for this and given an on-the-spot fine. Well away from a police station, though the copper in question was on a cycle himself.
Entirely justified: while I knew I was safe (I know which lights were on green, and there was no traffic coming those ways) and I could have tried to claim "by the time it was Red it wasn't safe to slam on the breaks" (it changed as I passed, I'm that argument sometimes holds water though it really shouldn't as the lights are on amber more than long enough before they hit red) I had broken the letter of the law very clearly.
As a cyclist who likes to think he is generally safe and conscientious, I think more should be pulled over and slapped for what they get up to on the roads and pavements. I've seen many cyclists ride like complete 'king loonies then yell and scream blaming someone else for the near misses that result.
Back to the original topic: I'm not anti-CCTV-in-public-places (and I count taxis as a form of public transport so while on a fair-paying trip I consider myself to not be in a private place) for its stated purposes. A cycling related case in point is the recent case when a bus driver literally used the bus to knock over a cyclist: the case may not have been so conclusive without the CCTV evidence. My problem is the lack of due diligence and other controls with respect to who has access to the recorded results: there are things I might say that I might not want certain other people to find out about (so I'd not want it made public) because it could cause unnecessary hassle/confusion/hurt.
People must be F'in cheap if they aren't willing to spend 99 cents.
Some studies have shown that many people are more likely to pirate something that is cheap. The act feels less damaging because you are depriving someone of 99p not £9.99 op $19.9 or £99.99. I assume the thought it "who is going to miss my single payment of $0.99?" and the situation doesn't register as important enough for other parts of their personality to take note (if they did they might jump up to question "hang on, have we fully considered the implications of that?").
If I see a Kindle book for 99 cents I just grab it;
I know some people wish their kids didn't think this way when they get the phone and credit card bills each month...
(then again, my answer to that is they little brats shouldn't have such free reign yet!)
I'm not wasting time trying to find a free pirate version. (shrug).
A lot of piracy is more passive than that - a real friend or an online contact (it could just be some random person on a forum) passes a copy (or a link there to) and says "this is the thing we talked about the other afternoon" or just "here, try this". They are not hunting the pirate copy, it is handed to them on a platter.
Some people are members of forums specifically for this purpose which is a grey area on this point: they make the general effort (they are a member of and at least occasionally log into that group) but don't make any specific effort to pirate that particular game/book/what-ever (they were on the group anyway, and had the few minutes spare at the time to click download and wait for it to install).
Of course it isn't secure in that sense. Of course your calls can be monitored and recorded.
If that is not the case then Microsoft are in breach of US laws regarding telecommunications (some brought in over recent years under the banner of national security, some that have been around longer).
If calls could not be monitored when they bought Skype, they will have changed that soon after, or if they still haven't sorted that yet they will be actively working towards that goal as we speak. Whether the law is right to expect this is a completely different discussion, MS and legal entities like them do not have the luxury of ignoring that law just because doing so might upset their customers (after all if they could get away with ignoring those regulations, what else might they try the trick with?).
tl;dr: No, it isn't truly secure. I'm, not sure why people thought it might be.
Any chance of providing just a little bit more information than "I know something you don't know"? Perhaps a link to somewhere where a beginner in the area might start to educate himself? Or a link to a device that we might consider when also considering the rPi and other common options?
So presumably these boards came from the gadget-maker - either as rejects, unsellable gadgets or as surplus/bankrupt stock.
Or if we are being a little more generous in our assessment, they could have just bought a small batch from the manufacturer direct as a perfectly normal sale. If so then they can do the same again, perhaps even making a bigger order so they can reduce the price per unit a little.
One of the complaints some people have made about the rPi is that there is no plan to offer wireless comms on any model in the foreseeable future, aside from plugging in your own Wifi or Bluetooth USB devices into the USB port(s) of course.For some environments WiFi is simply more convenient than wired networking. For instance if you want to use them in a robotics project, or hook them up to TVs (that you don't have ethernet cables near) as a cheap simple "dumb terminal" for web browsing, or (unless you want to play back HD video content, for which WiFi is usually not adequate IME) media playing.
Personally I'm favouring the rPi over the gPi here, but I can see why some people might not. There is nothing wrong in having variety in a marketplace!
Of course if you look at the specs you'll see why the board they are using exists: it is obviously originally intended for use in a small tablet device where ethernet isn't really a consideration but WiFi is essentially essential - the specs are identical to some of the many cheap tablets coming out of China right now (except the missing screen and case of course!).
I think what the person you are replying to was pointing out, perhaps too flippantly, is that the previous poster's post was the very definition of unhelpful in that it pointed out help could be given then failed to do so in any way shape or form. A couple of links to devices that already exist at good prices (for the home user buying one or two, not the commercial user buying by the thousand) that could be used for the projects people are planning to play with the Pis on, would have been the ideal way to help educate those of us who are missing the clue that is, in his estimation, obvious.
Of course the less charitable interpretation of what Aeros said is "prove you know something or I'm going to assume you are a blabbler mouth who really knows nothing and is just trying to look important", but if that is the case then the very same little bits of info in the original post would also render such a response irrelevant.
In either case the post by undefinedreference basically amounts to "I know something you don't know". Whether that is simply a matter-of-fact-statement, in invitation to politely ask for enlightenment, or something that in the playground would be followed by "ner ner n' ner ner", is a matter for individual interpretation until more information becomes apparent.
I've just (well, yesterday) taken delivery of an rPi and if my playing goes well might well buy more of the same or similar devices, so I for one would be most interested in hearing some detail of what undefinedreference says he knows and I don't (yet).
Which had their protection broken through hardware intervention, as was my point. Properly implemented UEFI can be protected much better than those old consoles offered. There is less likely to be an easy mod that people at home can perform. It'll not be uncrackable, but it has the potential to be a huge pain in the arse for some time. Also, the legislation protecting such hardware devices is much stronger, and the amounts of money the media collectives have to throw at taking pot shots at the general public (or politicians who don't play ball) much larger.
The key difference is that UEFI is enforced by hardware, not just software. While I doubt it will be uncrackable it is going to be significantly harder and the hacks against it may require physical intervention (not just software changes) which will stop may users replicating the crack.
Actually you could also claim that crown goes to Compaq, being the first to make use of the loophole to make IBM PC clones.
The resulting proliferation of machines that ran DOS (rather than manufacturers making more different machines and having to chose between paying MS specially for a compatible version or using something else (from someone else or developed internally, instead of just telling their users "put this on") is one of the significant things that squeezed possible alternatives out of the market.
That isn't really an apples-to-oranges comparison though. The desktop/laptop markets are different enough from the phone/tablet ones at the current time that you probably shouldn't compare Windows 7 against iOS and Android.
Also iOS and Android are growing fast due to new kit the people previously didn't own - they are not replacing Windows in most cases. MS's install base is rather impressive compared to iOS and Android so even at their current growth rate (which they can't maintain indefinitely - there will be a saturation point in the market somewhere) it'll be much time before they come close to eclipsing Windows.
I'm happy for you to put Microsoft down, but I recommend not using obviously flawed statistics as it just looks like desperation (when such desperation is probably not required).
Slashdot Mirror
If your device is lost or stolen, data should not be permanently deleted, just locked away until the owner personally comes round to identify herself with a passport or other legal ID of some sort. You can more to permanent delete after some time has passed without a "restore" request.
From an enterprise security point of view, once the device is out of your hands you want the data off it, full stop. If it isn't there then there is no chance that someone can read it. If everything on the device were properly encrypted, then you could just delete any keys and the restore would simply mean putting the keys back on.
I don't see why this should be any problem at all; Apple, Google and all their competitors claim to keep backups, which is effectively the same but with a user-"controlled" restore procedure.
That is the solution, not "not deleting". The off-device backups are your restore point either if you get a new device or that one is returned to you. As long, of course, as the backup account is not compromised at the same time as the device. No matter how securely you store you keys/tokens most phones are unlocked by a four digit pin so you've got not more than two days before someone brute forces that and gets in if they are determined and start when they first get hold of the device (so make sure if you lose the device that all the authentication credentials for the backups are changed ASAP).
Of course most stolen phones just get factory wiped before being fenced anyway, as most thefts of such devices are opportunistic rather than planned, so this is only a concern if someone might specifically target you (such as if others in your company's industry might want to have a peak at some significant trade secret) or if you have something really objectionable on the device (at which point if the thief notices it that can blackmail you)- most people like you or I are unlikely to be targeted in that way.
Having to collect the files together, waiting for that last one of the number you need to recreate the original content because for some reason it hasn't made it to your server's feed yet (or it has expired on your server and you need to wait for a repost), not having the content nicely indexed on the tracker we site, and so on, is more hassle than dealing with torrents.
OK, there are sites that do some of that indexing (but they are potentially subject to takedowns and DoS attacks as much as torrent indexes are), and there are clients that automate the getting of all the parts and unpacking them (as long as the original uploaded has prepared them properly), but at very least you need to research which client to use which is more hassle than just keeping using the torrent client you are already familiar with.
Just because something isn't difficult, that doesn't mean doing it isn't more hassle than what you do now.
Oh, and to get a good fast news feed with decent retention periods on binary groups you are going to have to pay. Not much unless you download silly amounts, but more than the nothing most torrent trackers cost to use.
"You let your dad down" is not a threat.
No, but:
* It is a deliberate attempt to cause offence by bringing up a subject that the recipient is likely to be pretty sensitive about
* Threat of drowning is, and that was contained in another of the tweets, so he wasn't arrested just because of the "dad" tweet
In this case I'm not unhappy that the tweeter's freedom of speech has been trumped by the recipients freedom to not have to put up with that sort of shit from the petulant little prick, and I fail to see the logic of extending this to "when will they come after me?!?!?!?!?!?" (as some people are doing) unless of course you are also in the habit of sending offensive and threatening messages to people you don't even know.
What we have here is yet another instance of http://www.penny-arcade.com/comic/2004/03/19/ - he thought he was safe pretending to be big and clever with his tweets until it became clear that people were going to fight back and that he was as risk of losing his anonymity so peple who know him in real life can see him for what he really is. The threat to Sky News says as much about his world view (i.e. the world should let him do what he wants with no consequences) that much else.
Any website that tries to, yes. ...) who get hacked.
Not any website that doesn't try, obviously. But those that don't try now could start trying at some point either through direct malice, being hacked, or carrying content served by 3rd parties (advertisers, stats collectors, servers providing public copies of common libraries,
You have no idea how many people I've talked to who seem to believe RAID is a fancy backup mechanism. I know of one small business owner in particular who had RAID 5 on his office computer, but no backup regimen. He didn't even know that the first drive had failed till he had lost all his business records. People ask about it regularly, with no apparent knowledge of its inherent limitations. They seem to think it's magic (like "the cloud").
Aye. Many expect it all to magically work without monitoring. There are many people out there with Linux software RAID setup (which I use in a few places) but without any monitoring or alerts setup because the recipe they copied didn't mention the need and they didn't think further than "RAID makes me safe". At very least have either smartctl or madam (preferably both) mailing you when there is any sort of issue with the drives and arrays, but people don't think "but how would I know if something has failed? How would I know a drive needs replacing?".
Maybe I'm predisposed to see this particular lunacy everywhere.
That is healthy, especially where us humans are concerned. A pessimist is never disappointed, and more often pleasantly surprised than an optimist!
You might be interested in some of the data integrity features in ZFS
I am. I'm also given to understand that BTRFS is barking up that particular tree.
I believe it is though I've not kept a close eye on it. ZFS is far more mature at this point (IIRC BTRFS is still officially "experimental", not even declared "beta" yet, and ZFS is considered production-ready in some environments (though not Linux)). Last I heard there are licensing issues that stop ZFS proper being ported over to Linux directly, there are various ways to use it (via FUSE is one example I remember though that seems far from ideal IMO) though I don't know how stable they are considered. BTRFS is in part a reaction to the feature set of ZFS (i.e. wanting some of those features without the licensing issues) though I think the Linux fs developers had some of the plans formulating before ZFS hit the scene in a big way.
I'm a little skeptical of mashing all of this into the file system level, but I'm not sure my fears are rational. It just feels like it ought to be easier to detect and fix problems if there were a separate RAID layer
I think part of the rational for keeping it all together is efficiency. Two separate layers that may not be aware of what each is doing can easily be very sub-optimal. Even if you get everything aligned perfectly and the filesystem is aware of your array's stripe sizes, there are still extra data copy operations going on. ZFS (and BTRFS) does a lot that would traditionally be elsewhere (like dealing with different storage types for performance, like moving things between SSD and spinning metal as needed (or using SSDs as cache rather than separate fast areas that add to the total space available).
The people working on ZFS (http://en.wikipedia.org/wiki/Zfs) agree with you (though that isn't applicable to all OSs).
RAID does provides some redundancy, but it is NOT a backup method.
I didn't say it was a backup method. RAID provides availability through redundancy (if the array succeeds in protecting you from a single drive failure, it stays up and can be repaired with very little downtime (or no downtime with hot-swap) so you don't jhave to wait how-ever long for your backups to restore (nor do you need to worry about work that might be lost that was done between now and the last backup) before getting back to work. In the last six months I've had drives go back in two locations and I was saved much hassle in both cases by them being members of RAID1 arrays.
RAID is primarily for performance. It only adds a smidgen of reliability.
Only RAID0 is about performance. Other arrangements can offer improved performance for some I/O patterns but availability through redundancy is the main aim of them. In fact for write heavy patterns, particularly for lots of small writes, levels 5 and 6 can impose a significant performance penalty.
I keep waiting for someone to come out with a RAID 6 ECC standard. Yes, It would be slow, but it would be reliable. Sometimes that's desirable. I think some of the expensive RAID cards can do this, though I've never seen one.
You might be interested in some of the data integrity features in ZFS (see http://en.wikipedia.org/wiki/Zfs#Data_Integrity as a starting point, if you are not already aware of it). I've not been in a position to use/test/benchmark it myself, but the descriptions of what it does (checksums of all blocks, data or meta, with self-healing from other array members as needed) certainly make it sound most impressive in that regard.
The key there, though, is it go and try to do a restore from your backup every now and then to make sure you can. There are few things worse than feeling secure in your backup process only to find it was not running correctly all along
A key point that many people miss until it is too late. Depending on your backup solution, automated testing is possible. Some of our Live databases get transferred back from the backup site and restored on another machine, then if that fails or the data in the restored copy seems to old (for that daily backed up DBs the test is that the last audit trail entry is no more than 24 hours after the last one in Live), an email is pinged out to several people so the situation gets looked at. That doesn't protect against everything (it is feasible that corruption in the backup file won't be picked up by the checksums used by the DBMS as it restores, and our last-audited-action check would still pass in most cases where that to happen) of course...
Similarly my home mail/calendar server gets its backup restored to a copy in a VM. I don't have automated checks on that though, I must do something about that omission soon (I currently manually log in occasionally to check things look right).
Your answer: correct.
But... your answer: pointless.
His answer: correct.
Though backups are not your only line of defense, redundancy is one too. For data that you care about you need both. Redundancy (via RAID) protects you from single (or more depending on arrangement) device failure (though not all failure routes so don't go feeling too safe!), backups protect you from catastrophic failure or an error (human or otherwise) that causes lost or corrupt data (if you delete everything, your RAID array will have redundant copies of nothing and your backups are what will save you unless they are stale).
That may make the drive work again, but it won't recover the data that was already on it (which it what they are asking about here).
The reason what you did works in some circumstances is that all drives beyond a certain point have more capacity than they actually let the outside world see. The extra capacity is used to remap back blocks to when they appear: meaning a multi Gbyte drive doesn't fail because a few MByte worth of surface goes bad. This only works until the "hidden" capacity is all used, after which there is nowhere to map bad sectors to. One of the readings usually offered by SMART is how many sectors have been reallocated this way: if that goes up too fast or nears full the drive will report itself as near to failing or in need of immediate replacement. If it is just one or two small spots that have failed due to some impurity in the platter, then a drive won't fail for writing BUT may still fail for reading: if the sector goes completely bad between being written to and read there is nothing the drive can do: it can't read the data to write it elsewhere, it will retry the read a number of times and if that fails you get a permanent error reported (if it succeeds the drive can remap the sector and all you see is an unusually slow read). Your drive will have had several sectors that went bad fast, so could not be remapped before the data was unreadable, by repeated writing you forced the drive to reconsider every sector and remap those ones that were iffy. You were lucky: there were enough sectors going spare for that to work and what-ever caused the bad sectors in the first places was not something that was drive-wide so no more failures occurred.
lt;dr: Yep, that sometimes works, but it doesn't get you any lost data back and I personally would never trust the drive again. The fact that this can work, and sometimes only temporarily as the drive my degrade further, is why I never but a drive from eBay or other second hand sources...
Notice how Valve is already running for the exits?
Not quite. They are setting up an extra stall outside, in the expectation that a fair few users will make for the exists. They are increasing their potential total market, not leaving one subset of the market in favour of another.
These people hate and fear change, and so will latch onto nearly any noticeable differences, so I'm thinking in terms of both front end functionality and the look of the interface.
You will not find a perfect Windows interface on Linux, so with that sort of person don't try. They will complain no matter what. They will complain if you move to Windows 8 too, because it is different too.
Instead of trying the impossible (pleasing the terminally unpleased), instead concentrate on picking something that is intuitive enough, that you can easily support (remotely if needed), and that works in your environment more-or-less out of the box.
As others have pointed out, don't change everything at once. If the tools you intend to use are cross-platform, introduce them under Windows first. That way you will take a few relatively small productivity hits as people get used to the new tools then get used to the new OS running them, instead of one huge productivity hit when everything changes at once.
Yet , those same people will say they love OS X, and they don't realize that its still Linux
It isn't. It is Unix derived (BSD mainly at the core) and uses a fair chunk of the GNU tool chain, but Linux it isn't. Not that the average use knows (or cares about) the difference.
Also what is it with cars feeling the need to overtake. Even if you're bobbling along in traffic at the same speed as the traffic, some lunatic will often try to overtake.
They feel they are losing time if they are behind or otherwise at the same speed as a cyclist. Some people perpetually in a rush, some others are just small-dicked wonders who use speed as a compensating device. In either case I'd actually prefer they get past me than stay behind: in front of me I can see what they are doing better, behind me they are a less well seen danger.
What annoys me sometimes is people slamming their foot down to overtake when I indicate that I intend to turn right (which will involve crossing the main body of the traffic lane) even when there is initially enough space between us and the opposing lane is empty so that they'd not even have to slow down at all as I safely cross their path with them at their original speed. Once one dick does that everyone behind keeps as close as they can so I end up sat by the side of the road for ages waiting for a clear spot in the flow to cross through (or, if I'm feeling petty and I know I can catch them up at the next lights, I'll get in front of the dick and set off very very very slowly when the lights change...).
If you want to be pulled over cycling in Cambridge, go through a red light by Parker's Piece.
Really? I've never seen anyone stopped for running a red light on a bike by the police.
I was once pulled over for this and given an on-the-spot fine. Well away from a police station, though the copper in question was on a cycle himself.
Entirely justified: while I knew I was safe (I know which lights were on green, and there was no traffic coming those ways) and I could have tried to claim "by the time it was Red it wasn't safe to slam on the breaks" (it changed as I passed, I'm that argument sometimes holds water though it really shouldn't as the lights are on amber more than long enough before they hit red) I had broken the letter of the law very clearly.
As a cyclist who likes to think he is generally safe and conscientious, I think more should be pulled over and slapped for what they get up to on the roads and pavements. I've seen many cyclists ride like complete 'king loonies then yell and scream blaming someone else for the near misses that result.
Back to the original topic: I'm not anti-CCTV-in-public-places (and I count taxis as a form of public transport so while on a fair-paying trip I consider myself to not be in a private place) for its stated purposes. A cycling related case in point is the recent case when a bus driver literally used the bus to knock over a cyclist: the case may not have been so conclusive without the CCTV evidence. My problem is the lack of due diligence and other controls with respect to who has access to the recorded results: there are things I might say that I might not want certain other people to find out about (so I'd not want it made public) because it could cause unnecessary hassle/confusion/hurt.
People must be F'in cheap if they aren't willing to spend 99 cents.
Some studies have shown that many people are more likely to pirate something that is cheap. The act feels less damaging because you are depriving someone of 99p not £9.99 op $19.9 or £99.99. I assume the thought it "who is going to miss my single payment of $0.99?" and the situation doesn't register as important enough for other parts of their personality to take note (if they did they might jump up to question "hang on, have we fully considered the implications of that?").
If I see a Kindle book for 99 cents I just grab it;
I know some people wish their kids didn't think this way when they get the phone and credit card bills each month...
(then again, my answer to that is they little brats shouldn't have such free reign yet!)
I'm not wasting time trying to find a free pirate version. (shrug).
A lot of piracy is more passive than that - a real friend or an online contact (it could just be some random person on a forum) passes a copy (or a link there to) and says "this is the thing we talked about the other afternoon" or just "here, try this". They are not hunting the pirate copy, it is handed to them on a platter.
Some people are members of forums specifically for this purpose which is a grey area on this point: they make the general effort (they are a member of and at least occasionally log into that group) but don't make any specific effort to pirate that particular game/book/what-ever (they were on the group anyway, and had the few minutes spare at the time to click download and wait for it to install).
Of course it isn't secure in that sense. Of course your calls can be monitored and recorded.
If that is not the case then Microsoft are in breach of US laws regarding telecommunications (some brought in over recent years under the banner of national security, some that have been around longer).
If calls could not be monitored when they bought Skype, they will have changed that soon after, or if they still haven't sorted that yet they will be actively working towards that goal as we speak. Whether the law is right to expect this is a completely different discussion, MS and legal entities like them do not have the luxury of ignoring that law just because doing so might upset their customers (after all if they could get away with ignoring those regulations, what else might they try the trick with?).
tl;dr: No, it isn't truly secure. I'm, not sure why people thought it might be.
Any chance of providing just a little bit more information than "I know something you don't know"? Perhaps a link to somewhere where a beginner in the area might start to educate himself? Or a link to a device that we might consider when also considering the rPi and other common options?
Or if we are being a little more generous in our assessment, they could have just bought a small batch from the manufacturer direct as a perfectly normal sale. If so then they can do the same again, perhaps even making a bigger order so they can reduce the price per unit a little.
One of the complaints some people have made about the rPi is that there is no plan to offer wireless comms on any model in the foreseeable future, aside from plugging in your own Wifi or Bluetooth USB devices into the USB port(s) of course.For some environments WiFi is simply more convenient than wired networking. For instance if you want to use them in a robotics project, or hook them up to TVs (that you don't have ethernet cables near) as a cheap simple "dumb terminal" for web browsing, or (unless you want to play back HD video content, for which WiFi is usually not adequate IME) media playing.
Personally I'm favouring the rPi over the gPi here, but I can see why some people might not. There is nothing wrong in having variety in a marketplace!
Of course if you look at the specs you'll see why the board they are using exists: it is obviously originally intended for use in a small tablet device where ethernet isn't really a consideration but WiFi is essentially essential - the specs are identical to some of the many cheap tablets coming out of China right now (except the missing screen and case of course!).
I think what the person you are replying to was pointing out, perhaps too flippantly, is that the previous poster's post was the very definition of unhelpful in that it pointed out help could be given then failed to do so in any way shape or form. A couple of links to devices that already exist at good prices (for the home user buying one or two, not the commercial user buying by the thousand) that could be used for the projects people are planning to play with the Pis on, would have been the ideal way to help educate those of us who are missing the clue that is, in his estimation, obvious.
Of course the less charitable interpretation of what Aeros said is "prove you know something or I'm going to assume you are a blabbler mouth who really knows nothing and is just trying to look important", but if that is the case then the very same little bits of info in the original post would also render such a response irrelevant.
In either case the post by undefinedreference basically amounts to "I know something you don't know". Whether that is simply a matter-of-fact-statement, in invitation to politely ask for enlightenment, or something that in the playground would be followed by "ner ner n' ner ner", is a matter for individual interpretation until more information becomes apparent.
I've just (well, yesterday) taken delivery of an rPi and if my playing goes well might well buy more of the same or similar devices, so I for one would be most interested in hearing some detail of what undefinedreference says he knows and I don't (yet).
Which had their protection broken through hardware intervention, as was my point. Properly implemented UEFI can be protected much better than those old consoles offered. There is less likely to be an easy mod that people at home can perform. It'll not be uncrackable, but it has the potential to be a huge pain in the arse for some time. Also, the legislation protecting such hardware devices is much stronger, and the amounts of money the media collectives have to throw at taking pot shots at the general public (or politicians who don't play ball) much larger.
The key difference is that UEFI is enforced by hardware, not just software. While I doubt it will be uncrackable it is going to be significantly harder and the hacks against it may require physical intervention (not just software changes) which will stop may users replicating the crack.
Actually you could also claim that crown goes to Compaq, being the first to make use of the loophole to make IBM PC clones.
The resulting proliferation of machines that ran DOS (rather than manufacturers making more different machines and having to chose between paying MS specially for a compatible version or using something else (from someone else or developed internally, instead of just telling their users "put this on") is one of the significant things that squeezed possible alternatives out of the market.
That isn't really an apples-to-oranges comparison though. The desktop/laptop markets are different enough from the phone/tablet ones at the current time that you probably shouldn't compare Windows 7 against iOS and Android.
Also iOS and Android are growing fast due to new kit the people previously didn't own - they are not replacing Windows in most cases. MS's install base is rather impressive compared to iOS and Android so even at their current growth rate (which they can't maintain indefinitely - there will be a saturation point in the market somewhere) it'll be much time before they come close to eclipsing Windows.
I'm happy for you to put Microsoft down, but I recommend not using obviously flawed statistics as it just looks like desperation (when such desperation is probably not required).