It has pretty decent implications for vehicles. Even with a lead-acid battery, a small amp positive trickle charge will go a long way to making your battery last longer before it's in need of replacement.
What are the implications for EVs, which seem to primarily use Li based batteries?
I know the 'record' for these batteries is pretty good so far (not bad enough to make Consumer Reports respond yet, at least), but I have to wonder how much of that is due to ideal environmental factors and how much of it is due to the things not being out long enough, or used enough, to get an accurate measure. How are the first generations of the Prius doing? I've yet to see any reviews or analysis. This is important for the used car market (to assure there is one in 10 years, and automobiles don't become disposable due to the cost of replacement).
A $20-30k vehicle lasting a mere 5 years before needing primary propulsion to be replaced is pretty crazy, especially when it involves a non-trivial amount of Lithium to do so (not exactly the most common of alkali metals).
I'm sure there's probably a technological way to prevent or slow this from happening, but for the time being, we're stuck with the technology we've got. Combine this with the lifecycle degradation of cold and heat on Lithium cells, and they seem to be a really poor broad-application general-purpose power sequestering method.
A device does not 'just work' if it does not meet a socially acceptable definition for 'working', whatever that 'work' may be. "just working" should be, at least, as good as the previous version.
Just like Vista did not work, and Windows 7 did. Just like OSX does, and OSX9 did not.
If a feature bullet is "has PIM" or "integrates contacts" I'd expect that to mean:
* has a PIM which is at least as featureful and adaptive as something made 5 years ago * has the ability to sync gracefully with a number of different calendar/contact management tools
I'm sorry, but if it can't even do basics like IMAP properly (Android, last I looked) then it really has no place being called a "smart phone".
I'm not saying "must work with pine and my exotic contact spreadsheet" or even "must work with Thunderbird" (that'd be nice), but the depth of functionality in this department is dreadful on iPhone and Android. Thanks, but I'd rather not be forced to use use Apple Mail or gmail just to use my phone's functionality.
I would say that less than 50% (as a starting point) of people need to sync with their Outlook calendars, view MS Project, or view a spreadsheet.
Yet anyone who actually needs a smartphone - ie, not for insipid use - needs some basic things:
* well-designed, unified and adaptable PIM which integrates with your desktop/web/phone contacts * email * calendar management
These are things with the iPhone and Android fail at. Microsoft does it well, due to copying Palm years ago, and Palm does well because, well... it's Palm.
If you need 2 of the 3 things above, iPhone or Android phones are probably a poor choice.
Saying the iPhone "just works" is a bit disingenius due to the... lack of actual work applications.
Granted, the same can be said for Android. But consider: "working" on Windows Mobile 6.5 is somewhat more possible than on either Android or iPhone due to good PIM and Office integration, as well as the many other tools which can interface with, and utilize, said functionality. FOr being as fragmented and crusty as WinMo is, it's still more capable than either.
If 'just working' for you is having a unified UI across multiple 'personal' devices slated for different roles so you can check Facebook statuses and your gmail/iCal/whatever and play games, sure. the iPhone 'just works' by those requirements. Just don't expect anything 'complex' (such as anything a common PDA was capable of as recent as uh almost 10 years ago).
Now, the Palm Pre... there's a phone that "just works".
I have the opposite experience: "Apple" programs are invariably frustrating UI experiences, vs. Microsoft ones, which (aside from the Ribbon of Agitation) are quite a bit more sensible.
I should note that I'm not a serious user of either OS; I've been using Linux almost exclusively for 12 years, and have only briefly used one or the other for work (roughly equal proportions).
Considering a development project, I looked at both Xcode and Visual Studio 2008. The ease with which I could start and get a basic app going in VS2k8 was many, many times easier than in Xcode (all previous development has been done on console with vi/vim, cvs, and the like - no GUI stuff).
I'm not even sure how Apple UIs could be considered 'better'. The only think remotely 'superior' about Apple's UI at this point in the game is their control panel, which is fairly minimal on text and clean. (There are also a lot fewer options, which is significant, IMO.)
Properly stored tape will last eons. Hard drives will not.
That's not a problem for backups. It's a problem for archives, which are an entirely different (and immensely more compounded) issue.
apes are more durable.
Durable, compared to what?! Aside from being much more volatile than a drive to ESD, they're also not hermetically sealed. The plastic tape itself stretches, gets kinks, and can get stuck - even without improper physical operation.
Broken tapes are user-serviceable.
This is hardly a benefit, because - unless it's the only copy - you've got another set.
Proper backups entail 2 full revisions/data sets, differentials, and the live copy. At least. Short of your live systems, your daily/etc. differentials, as well as two full week data sets?
I'd also say it's debatable how 'user servicable' a tape is if it's been damaged, given the speed at which current tape systems move, the sensitivity of the electronics, and the relative density of the tape. It does not take much dirt or oil to damage your tape drive.
Data is far more recoverable from broken tape than from a broken hard drive.
Provided you write as a an unencrypted/uncompressed stream to the tape, yes. But then you're talking about half the data density that's advertised.
The costs you cite are ridiculous for a simple home setup.
Are they? They're not for the data amounts mentioned. an LTO3 tape (400GB) is $35+. An LTO5 tape (1.5TB) is over $110. Both are significantly more than hard drives of similar capacity, which have electronics to detect and correct media errors on the fly, giving you significantly better 'wiggle room'. Let's not forget that you've also got to buy a tape drive: find me one that's new and under $900 and worth putting in the role of 'backup'; I've yet to see one.
I read recently that drive plater errors occur roughly once every 2TB of transfered data. How often does a backup happen where the archive is corrupt and 'off by one'? I'm guessing it's somewhat more frequently than once every 2TB.
Tape backup is an epic fail when you're using large files or tapes smaller than your files; splitting archives across tapes significantly reduces the reliability of your backups.
Are you aware that there are a number of good, free virtualization platforms available for you to use? This will allow you to natively run whatever crap it is that won't run on a 64 bit host, and you can actually make use of more than 2GB of memory effectively as a result.
Steadily improving?! The Vista debacle happened under Ballmer. Why do you attribute it to Bill Gates?
Gates stepped down in June of 2006. Vista came out in November. Do you really think that Vista was not already a fattened cow by the time Gates left? This is an OS which they'd been working on for well over 5 years at that point; the problems that are there, were already there. It took another 3 years after Vista came out to make the damn platform usable (W7). Arguably, the later releases of Vista were invariably better than the first two major ones, and unlike other Windows releases, it actually has improved in almost every regard with each update (vs. the instability, bloat, etc. of everything prior).
Yes, I realize that 'officially' the mantle was passed some time before that. I'm also aware that Gates was likely heavily involved in decisions until his final, actual departure due to the decision making distrust between the two.
I'm not saying that Ballmer is invariably good, mind you. I'm just saying he's better than Gates. Gates was a meddler; Ballmer seems to be much more of a chubby visionary who's not going to shitcan someone for doing something he disagrees with, as long as it works. Gates, on the other hand, seems like he was a real pleasure to work with.
* Stability, scalability, usability - name it, it's improved since Gates left and Balmer took over.
Some examples would be nice.
Not exactly a difficult list.
* Memory management/precache actually works quite well now, on all lines. This, vs. the crap in Vista and the lack of it in XP. * CIFS performance isn't nearly the nightmare it used to be (though there's still a lot of protocol overhead) * Support for more than 2 CPUs, and very, very good multicore/processor scaling (low overhead). Performance improvement is nearly linear in W7 based OSes (W7 and 2k8r2), which, IIRC, rivals even Linux (or at least compares favorably). * W7 will run on very low end systems (512Mb, 1GHz), with applications, even though it is advertised to 'require' 1GB (IIRC). Under such situations it performs roughly as well, if not better than, an up-to-date XP machine. * Power management for laptops on W7 is better than pretty much everything else, at this point. Yes, that means Mac and Linux, too. * I can make W7 look and behave like Maemo 4 (which I love) or like XP if I want. I can make it look fairly similar to OS X, too - albeit with a bit more finessed user interaction. I can properly tile and arrange windows, which was not previously possible. Oh, and I don't have to navigate through inane menus (which I'd personally regard as a step back for those familiar w/ the old arrangement) - I just type what I want and it 'content completes' it for me. (This functionality works much better than it did in Vista.) * While just an example (on stability), I've yet to see a 2k8r2 or W7 machine crash or reboot where it was not the fault of nvidia or ATI drivers due to fullscreen video bugs. Once those minor issues were fixed w/ updated drivers, I've yet to see it happen.
I suppose it might help if I note that this isn't a fanboy list; I don't even have a Windows system at home (though I do have a couple VMs). I'm just stating observations.
Why would you put a 3TB drive in a 5-year-old+ computer? There haven't been 32bit x86 machines on the market (as 'new product') for several+ years now, and only then on netbooks and laptops. Anything capable of running Windows 7 will, in all likelihood, be using a 64 bit processor.
If you're running 32 bit Windows 7, let me refer you to my previous post where I say "It's dead, Jim." There is no reason to do so unless it is archaic hardware (in which case - does it even have a SATA port?)
What I want to know is: how can you justify the cost of tape? And why isn't a raid6 array a valid backup location?
What, exactly, does tape provide you in terms of archival veracity and longevity that current drives do not? Assuming no significant sunk cost for tape hardware, you're still looking at similar if not greater costs per GB of tape storage as you would be disk, whether you're looking at LTO 3 or LTO 5. Throw in $1,000 to $3,000 for your standalone chassis tape drive... you'd have to burn through hundreds of single-use 1.5TB hard drives to justify tape on cost (and even then, questionably - tape is more expensive per raw GB than drives).
The whole 'raid isn't backup' argument seems a misnomer to me these days. Yes, bad backup practices make tape less reliable, but it's much more difficult to put good practices in place for effective tape backup than it is for hard disks. With filesystems like ZFS (with CoW and a number of other nice features that make tape further irritating in comparison), I don't see the point at all.
No, raid isn't a backup in and of itself - but neither is a single tape. A raid5 live copy of your data with periodic/daily/whatever diffentials to external drive, however, seems like a pretty good backup to me.
Unfortunately, if you're still using Windows XP, don't expect your system to make full use of any 3TB drive (yet).
What do you mean, "yet"? XP Pro was EOL'd in April of last year. It's dead, Jim. There will be no updates or upgrades from that date forward.
Furthermore, there is absolutely 0 (legitimate) reason to be running XP on a machine which will recognize a 3TB drive at the hardware level. None. If you can come up with a reason, there's probably a better way to do it than your proposed approach, long term.
At this point, short of a large uniform environment where there are specific applications to support that only work on XP, running XP is a fool's errand. It's rife with security issues. Sure, keep it around until your older hardware kicks it, but by all means you should be upgrading to something that will actually work with, and be able to utilize, newer hardware.
For all his personal faults, Balmer has done for Microsoft what Bill Gates could not. He's made some very prudent decisions (or lack of decisions, maybe) which have had opposite results to what BG did:
* Windows has been steadily improving since Bill Gates left the helm. The last vestige of Gates' impact was seen in Office 2007 and Windows Vista, both of which were horrible. * Stability, scalability, usability - name it, it's improved since Gates left and Balmer took over. * Xbox was a huge fail; Xbox 360, on the other hand (while having been released under Gates and not doing that well during that time) has seen steady improvements over it's long life - and is still considered 'premiere' by many after 5 long years (since when, the PS3 and Wii have been released - to limited impact). * Sharepoint, while it sucks giant donkey cock and is the bane of my existence, has become quite the beast, seemingly being the preferred choice for any sort of corporate extranet deployment/content management system. It is better than most of its competition at substantially lower prices. * HyperV has become quite the mature, capable product - a far cry from the code it was based on, at this point. Microsoft has helped push virtualization forward through many of its VT-related initiatives - to the benefit of the industry as a whole.
The only thing Balmer seems to have butchered severely is the mobile handset OSes. It had potential 3-4 years ago; today, the MS offerings are has-beens and also-rans not much worth consideration (except for the older WinMo based phones, which have some merit due to Exchange/PIM and decent integration).
* 'soft' degrees seem to have broader applicability than a technical degree. You can get - literally - a degree in comparative studies and do poorly on it and be more successful in your career than a technical type. What's more, you'll be successful and advance cross-fields, regardless of how similar or dissimilar to your previous position. Likewise for an 'artistic' degree. I've seen it quite a few times: these people end up in middle management and make life hell for everyone around them. * An 'engineering' type degree - pretty much anything technical - shoehorns your job opportunities upon graduation. Don't bother applying for something non-technical; you're not qualified for something technical outside your field, but at the same time you're also not seen as qualified for much else - despite any reality.
Arguably, universities are already run like Wikipedia, in many ways:
* On highly politicized topics, the level of meta debate beneath the surface (article/curriculum) is quite heated. * If you disagree with the party line of someone in charge (moderator, professor), you can often be banned/given a bad grade simply due to an opinion/difference in belief despite the absence of fact from either side of the argument. * They have highly entrenched cultures which are hostile to the outside world. * The highly established paragons of the institution are there not because of merit but because of someone they know or tenure. This usually results in the previous instance. * There are a few at the top of the organizations which gain massive amounts of wealth - to little effort on their part - due to the efforts of the masses. * The food is horrible.
Now, there are some good universities and colleges - and I'd argue the smaller private schools are probably better than the common university by a long stretch. I went to a little school called Juniata College in PA. You picked your own major - not from a list, but completely piecemeal. It was supervised for cohesion and comprehension by a faculty member. The school was small enough that you actually did get to have some decent interactions with those in your major, and there was a fair amount of cross-discipline intellectual pollination.
Gearbox, and Valve, have both benefited greatly from user-generated content. If it wasn't for Counter-Strike, it's questionable whether Steam and other Half-Life/Steam franchise games would have ever come to be, nevermind seen the distribution width that they have. Seems Steam was originally conceived to speed the distribution of things like CounterStrike add-ons and updates.
You're assuming there are people on-call to react in 20-30 minutes, and/or it's on-site, and that there are two people to respond.
* 20-30 minutes to show up on site to figure out what's wrong * 10-30 minutes to get the replacement. * 20-60 (possibly concurrently) pull the dead switch and plug the new one in. This is made more difficult by a full rack and/or ports with VLANs due to the need to be careful. * 5min - reload the port/switch configuration
No, please! The more time he spends in front of cameras, the less time he has to 'fix' things.
Look at the results of his "fixes", so far. Not exactly positive results, to say the least! I certainly don't want more if the same; it'll take generations to dig ourselves out of this one alone, nevermind his predecessors.
Didn't Mythbusters already do this one? Like, 2-3 times, I think.
It'd be great if they put more effort into it this time, but seriously: it's not exactly an unknown scientific history mystery. Scientists and historians have been wondering for a long time about the veracity of the Siege of Syracuse tale. It's right up there with the actual content of Greek Fire.
Almost every single example of "replication attempt" tries to do so in a closed-minded fashion. They think: what do we know the $society was capable of? How can we apply these means to the desired end?
Instead, they should be thinking: how, without very specific technological advances several inventive generations removed, might we accomplish the same ends? Instead, they're sitting there rolling copper and making guestimates about things like troop strength and the like.
This is exactly the problem with Office to OpenOffice migrations, at this point.
Arguably, Word deals with complex Word forms and the like better than OpenOffice, too. When you've got hundred+ page document form templates which autogenerate a lot of information, etc. moving to something else is difficult as well.
I've got a friend who ranches. He's technically savvy, but not what I'd call a genius. All his information is kept in excel worksheets, which auto-generate important data via formulas. He uses WinCE/WinMo devices (a number of them) to scan in and enter data when out in the barn or pasture (a common/daily occurrence) with the Pocket Excel (or w/e it's called). He's then able to examine and generate reports - using the same files. Often, he's accessing the files directly over a VPN on a Windows file share.
With Windows, it's not convoluted to do this. To do this in any other way would be difficult, because there's years of development involved. He's tried OO, and it won't due the trick because there's nothing in PDA form that does the trick short of a custom application.
Any organization small enough to have trouble funding and domain controller Doesn't need one.
Are you saying that the client/server model does not scale below a certain number of hosts?
I can see where a capital expense for an office of 5 or so people might be a bit severe, but where a domain is still necessary. Say, where there's a reason to have cleanly segregated user permissions/access controls (which Samba3 does not provide easily/without significant knowledge of how to set it up/reoccurring user admin costs).
I don't see that reducing a huge amount even when Samba 4 is released - there's a lot of configuration involved to get DHCP, DNS and Samba 4 talking to each other properly.
It takes less than 20 minutes to half an hour if you've already got DDNS running on the network.
Pretty much everything relies upon Active Directory. Unfortunately, the s4 implementation isn't there yet for basic stuff, nevermind being able to use later versions of Exchange (which is, as far as I can tell, one of the only reasons why people are moving to 2k8r2 if they've already got a substantial 2k3 install base).
Pretty much the whole picture seems to be moving towards "linux" in one form or fashion. Popular virtualization technology, aside from the awkward bare metal Hyper-V, all runs on top of a Linux kernel (Xen, XenServer, KVM, VMWare). Aside from that, there's MS's HyperV and Sun/Oracle's offerings, which don't have nearly the incentive to use.
In many cases I've seen, even if someone is reliant upon AD and Exchange, they've already taken steps to marginalize their reliance: they've virtualized their AD and MSSQL servers; they've picked up things like Zimbra to replace Exchange (often transparently); and so on. I doubt it's happening so much in a corporate environment, but it's happening in a very big way in shops where there are 2-3 IT workers who have a little extra time (ie not running their heads off).
I remember going into a computer lab once in school, years ago. Windows 2000 was still the most common desktop OS.
I sat down at a workstation and opened the browser. Odd, I thought: the pixels didn't quite line up. Looking closer (at the browsers user agent string) I noticed it was actually Linux, and in all likelihood a themed ICEWM. THe workstation had Open Office (StarOffice, I think) and Mozilla, and that was about it - it was transparent enough and similar enough to a locked down w2k workstation that I'm sure the common user could not tell the difference.
Slashdot Mirror
It has pretty decent implications for vehicles. Even with a lead-acid battery, a small amp positive trickle charge will go a long way to making your battery last longer before it's in need of replacement.
What are the implications for EVs, which seem to primarily use Li based batteries?
I know the 'record' for these batteries is pretty good so far (not bad enough to make Consumer Reports respond yet, at least), but I have to wonder how much of that is due to ideal environmental factors and how much of it is due to the things not being out long enough, or used enough, to get an accurate measure. How are the first generations of the Prius doing? I've yet to see any reviews or analysis. This is important for the used car market (to assure there is one in 10 years, and automobiles don't become disposable due to the cost of replacement).
A $20-30k vehicle lasting a mere 5 years before needing primary propulsion to be replaced is pretty crazy, especially when it involves a non-trivial amount of Lithium to do so (not exactly the most common of alkali metals).
I'm sure there's probably a technological way to prevent or slow this from happening, but for the time being, we're stuck with the technology we've got. Combine this with the lifecycle degradation of cold and heat on Lithium cells, and they seem to be a really poor broad-application general-purpose power sequestering method.
A device does not 'just work' if it does not meet a socially acceptable definition for 'working', whatever that 'work' may be. "just working" should be, at least, as good as the previous version.
Just like Vista did not work, and Windows 7 did. Just like OSX does, and OSX9 did not.
If a feature bullet is "has PIM" or "integrates contacts" I'd expect that to mean:
* has a PIM which is at least as featureful and adaptive as something made 5 years ago
* has the ability to sync gracefully with a number of different calendar/contact management tools
I'm sorry, but if it can't even do basics like IMAP properly (Android, last I looked) then it really has no place being called a "smart phone".
I'm not saying "must work with pine and my exotic contact spreadsheet" or even "must work with Thunderbird" (that'd be nice), but the depth of functionality in this department is dreadful on iPhone and Android. Thanks, but I'd rather not be forced to use use Apple Mail or gmail just to use my phone's functionality.
I would say that less than 50% (as a starting point) of people need to sync with their Outlook calendars, view MS Project, or view a spreadsheet.
Yet anyone who actually needs a smartphone - ie, not for insipid use - needs some basic things:
* well-designed, unified and adaptable PIM which integrates with your desktop/web/phone contacts
* email
* calendar management
These are things with the iPhone and Android fail at. Microsoft does it well, due to copying Palm years ago, and Palm does well because, well... it's Palm.
If you need 2 of the 3 things above, iPhone or Android phones are probably a poor choice.
Saying the iPhone "just works" is a bit disingenius due to the... lack of actual work applications.
Granted, the same can be said for Android. But consider: "working" on Windows Mobile 6.5 is somewhat more possible than on either Android or iPhone due to good PIM and Office integration, as well as the many other tools which can interface with, and utilize, said functionality. FOr being as fragmented and crusty as WinMo is, it's still more capable than either.
If 'just working' for you is having a unified UI across multiple 'personal' devices slated for different roles so you can check Facebook statuses and your gmail/iCal/whatever and play games, sure. the iPhone 'just works' by those requirements. Just don't expect anything 'complex' (such as anything a common PDA was capable of as recent as uh almost 10 years ago).
Now, the Palm Pre... there's a phone that "just works".
I have the opposite experience: "Apple" programs are invariably frustrating UI experiences, vs. Microsoft ones, which (aside from the Ribbon of Agitation) are quite a bit more sensible.
I should note that I'm not a serious user of either OS; I've been using Linux almost exclusively for 12 years, and have only briefly used one or the other for work (roughly equal proportions).
Considering a development project, I looked at both Xcode and Visual Studio 2008. The ease with which I could start and get a basic app going in VS2k8 was many, many times easier than in Xcode (all previous development has been done on console with vi/vim, cvs, and the like - no GUI stuff).
I'm not even sure how Apple UIs could be considered 'better'. The only think remotely 'superior' about Apple's UI at this point in the game is their control panel, which is fairly minimal on text and clean. (There are also a lot fewer options, which is significant, IMO.)
Properly stored tape will last eons.
Hard drives will not.
That's not a problem for backups. It's a problem for archives, which are an entirely different (and immensely more compounded) issue.
apes are more durable.
Durable, compared to what?! Aside from being much more volatile than a drive to ESD, they're also not hermetically sealed. The plastic tape itself stretches, gets kinks, and can get stuck - even without improper physical operation.
Broken tapes are user-serviceable.
This is hardly a benefit, because - unless it's the only copy - you've got another set.
Proper backups entail 2 full revisions/data sets, differentials, and the live copy. At least. Short of your live systems, your daily/etc. differentials, as well as two full week data sets?
I'd also say it's debatable how 'user servicable' a tape is if it's been damaged, given the speed at which current tape systems move, the sensitivity of the electronics, and the relative density of the tape. It does not take much dirt or oil to damage your tape drive.
Data is far more recoverable from broken tape than from a broken hard drive.
Provided you write as a an unencrypted/uncompressed stream to the tape, yes. But then you're talking about half the data density that's advertised.
The costs you cite are ridiculous for a simple home setup.
Are they? They're not for the data amounts mentioned. an LTO3 tape (400GB) is $35+. An LTO5 tape (1.5TB) is over $110. Both are significantly more than hard drives of similar capacity, which have electronics to detect and correct media errors on the fly, giving you significantly better 'wiggle room'. Let's not forget that you've also got to buy a tape drive: find me one that's new and under $900 and worth putting in the role of 'backup'; I've yet to see one.
I read recently that drive plater errors occur roughly once every 2TB of transfered data. How often does a backup happen where the archive is corrupt and 'off by one'? I'm guessing it's somewhat more frequently than once every 2TB.
Tape backup is an epic fail when you're using large files or tapes smaller than your files; splitting archives across tapes significantly reduces the reliability of your backups.
Are you aware that there are a number of good, free virtualization platforms available for you to use? This will allow you to natively run whatever crap it is that won't run on a 64 bit host, and you can actually make use of more than 2GB of memory effectively as a result.
Steadily improving?! The Vista debacle happened under Ballmer. Why do you attribute it to Bill Gates?
Gates stepped down in June of 2006. Vista came out in November. Do you really think that Vista was not already a fattened cow by the time Gates left? This is an OS which they'd been working on for well over 5 years at that point; the problems that are there, were already there. It took another 3 years after Vista came out to make the damn platform usable (W7). Arguably, the later releases of Vista were invariably better than the first two major ones, and unlike other Windows releases, it actually has improved in almost every regard with each update (vs. the instability, bloat, etc. of everything prior).
Yes, I realize that 'officially' the mantle was passed some time before that. I'm also aware that Gates was likely heavily involved in decisions until his final, actual departure due to the decision making distrust between the two.
I'm not saying that Ballmer is invariably good, mind you. I'm just saying he's better than Gates. Gates was a meddler; Ballmer seems to be much more of a chubby visionary who's not going to shitcan someone for doing something he disagrees with, as long as it works. Gates, on the other hand, seems like he was a real pleasure to work with.
* Stability, scalability, usability - name it, it's improved since Gates left and Balmer took over.
Some examples would be nice.
Not exactly a difficult list.
* Memory management/precache actually works quite well now, on all lines. This, vs. the crap in Vista and the lack of it in XP.
* CIFS performance isn't nearly the nightmare it used to be (though there's still a lot of protocol overhead)
* Support for more than 2 CPUs, and very, very good multicore/processor scaling (low overhead). Performance improvement is nearly linear in W7 based OSes (W7 and 2k8r2), which, IIRC, rivals even Linux (or at least compares favorably).
* W7 will run on very low end systems (512Mb, 1GHz), with applications, even though it is advertised to 'require' 1GB (IIRC). Under such situations it performs roughly as well, if not better than, an up-to-date XP machine.
* Power management for laptops on W7 is better than pretty much everything else, at this point. Yes, that means Mac and Linux, too.
* I can make W7 look and behave like Maemo 4 (which I love) or like XP if I want. I can make it look fairly similar to OS X, too - albeit with a bit more finessed user interaction. I can properly tile and arrange windows, which was not previously possible. Oh, and I don't have to navigate through inane menus (which I'd personally regard as a step back for those familiar w/ the old arrangement) - I just type what I want and it 'content completes' it for me. (This functionality works much better than it did in Vista.)
* While just an example (on stability), I've yet to see a 2k8r2 or W7 machine crash or reboot where it was not the fault of nvidia or ATI drivers due to fullscreen video bugs. Once those minor issues were fixed w/ updated drivers, I've yet to see it happen.
I suppose it might help if I note that this isn't a fanboy list; I don't even have a Windows system at home (though I do have a couple VMs). I'm just stating observations.
I suspect not. Why does it matter?
Why would you put a 3TB drive in a 5-year-old+ computer? There haven't been 32bit x86 machines on the market (as 'new product') for several+ years now, and only then on netbooks and laptops. Anything capable of running Windows 7 will, in all likelihood, be using a 64 bit processor.
If you're running 32 bit Windows 7, let me refer you to my previous post where I say "It's dead, Jim." There is no reason to do so unless it is archaic hardware (in which case - does it even have a SATA port?)
What I want to know is: how can you justify the cost of tape? And why isn't a raid6 array a valid backup location?
What, exactly, does tape provide you in terms of archival veracity and longevity that current drives do not? Assuming no significant sunk cost for tape hardware, you're still looking at similar if not greater costs per GB of tape storage as you would be disk, whether you're looking at LTO 3 or LTO 5. Throw in $1,000 to $3,000 for your standalone chassis tape drive... you'd have to burn through hundreds of single-use 1.5TB hard drives to justify tape on cost (and even then, questionably - tape is more expensive per raw GB than drives).
The whole 'raid isn't backup' argument seems a misnomer to me these days. Yes, bad backup practices make tape less reliable, but it's much more difficult to put good practices in place for effective tape backup than it is for hard disks. With filesystems like ZFS (with CoW and a number of other nice features that make tape further irritating in comparison), I don't see the point at all.
No, raid isn't a backup in and of itself - but neither is a single tape. A raid5 live copy of your data with periodic/daily/whatever diffentials to external drive, however, seems like a pretty good backup to me.
Unfortunately, if you're still using Windows XP, don't expect your system to make full use of any 3TB drive (yet).
What do you mean, "yet"? XP Pro was EOL'd in April of last year. It's dead, Jim. There will be no updates or upgrades from that date forward.
Furthermore, there is absolutely 0 (legitimate) reason to be running XP on a machine which will recognize a 3TB drive at the hardware level. None. If you can come up with a reason, there's probably a better way to do it than your proposed approach, long term.
At this point, short of a large uniform environment where there are specific applications to support that only work on XP, running XP is a fool's errand. It's rife with security issues. Sure, keep it around until your older hardware kicks it, but by all means you should be upgrading to something that will actually work with, and be able to utilize, newer hardware.
Don't be silly.
For all his personal faults, Balmer has done for Microsoft what Bill Gates could not. He's made some very prudent decisions (or lack of decisions, maybe) which have had opposite results to what BG did:
* Windows has been steadily improving since Bill Gates left the helm. The last vestige of Gates' impact was seen in Office 2007 and Windows Vista, both of which were horrible.
* Stability, scalability, usability - name it, it's improved since Gates left and Balmer took over.
* Xbox was a huge fail; Xbox 360, on the other hand (while having been released under Gates and not doing that well during that time) has seen steady improvements over it's long life - and is still considered 'premiere' by many after 5 long years (since when, the PS3 and Wii have been released - to limited impact).
* Sharepoint, while it sucks giant donkey cock and is the bane of my existence, has become quite the beast, seemingly being the preferred choice for any sort of corporate extranet deployment/content management system. It is better than most of its competition at substantially lower prices.
* HyperV has become quite the mature, capable product - a far cry from the code it was based on, at this point. Microsoft has helped push virtualization forward through many of its VT-related initiatives - to the benefit of the industry as a whole.
The only thing Balmer seems to have butchered severely is the mobile handset OSes. It had potential 3-4 years ago; today, the MS offerings are has-beens and also-rans not much worth consideration (except for the older WinMo based phones, which have some merit due to Exchange/PIM and decent integration).
Interestingly enough, I've noticed two things:
* 'soft' degrees seem to have broader applicability than a technical degree. You can get - literally - a degree in comparative studies and do poorly on it and be more successful in your career than a technical type. What's more, you'll be successful and advance cross-fields, regardless of how similar or dissimilar to your previous position. Likewise for an 'artistic' degree. I've seen it quite a few times: these people end up in middle management and make life hell for everyone around them.
* An 'engineering' type degree - pretty much anything technical - shoehorns your job opportunities upon graduation. Don't bother applying for something non-technical; you're not qualified for something technical outside your field, but at the same time you're also not seen as qualified for much else - despite any reality.
Arguably, universities are already run like Wikipedia, in many ways:
* On highly politicized topics, the level of meta debate beneath the surface (article/curriculum) is quite heated.
* If you disagree with the party line of someone in charge (moderator, professor), you can often be banned/given a bad grade simply due to an opinion/difference in belief despite the absence of fact from either side of the argument.
* They have highly entrenched cultures which are hostile to the outside world.
* The highly established paragons of the institution are there not because of merit but because of someone they know or tenure. This usually results in the previous instance.
* There are a few at the top of the organizations which gain massive amounts of wealth - to little effort on their part - due to the efforts of the masses.
* The food is horrible.
Now, there are some good universities and colleges - and I'd argue the smaller private schools are probably better than the common university by a long stretch. I went to a little school called Juniata College in PA. You picked your own major - not from a list, but completely piecemeal. It was supervised for cohesion and comprehension by a faculty member. The school was small enough that you actually did get to have some decent interactions with those in your major, and there was a fair amount of cross-discipline intellectual pollination.
Gearbox, and Valve, have both benefited greatly from user-generated content. If it wasn't for Counter-Strike, it's questionable whether Steam and other Half-Life/Steam franchise games would have ever come to be, nevermind seen the distribution width that they have. Seems Steam was originally conceived to speed the distribution of things like CounterStrike add-ons and updates.
You're assuming there are people on-call to react in 20-30 minutes, and/or it's on-site, and that there are two people to respond.
* 20-30 minutes to show up on site to figure out what's wrong
* 10-30 minutes to get the replacement.
* 20-60 (possibly concurrently) pull the dead switch and plug the new one in. This is made more difficult by a full rack and/or ports with VLANs due to the need to be careful.
* 5min - reload the port/switch configuration
I was just looking at worst case scenario.
No, please! The more time he spends in front of cameras, the less time he has to 'fix' things.
Look at the results of his "fixes", so far. Not exactly positive results, to say the least! I certainly don't want more if the same; it'll take generations to dig ourselves out of this one alone, nevermind his predecessors.
Didn't Mythbusters already do this one? Like, 2-3 times, I think.
It'd be great if they put more effort into it this time, but seriously: it's not exactly an unknown scientific history mystery. Scientists and historians have been wondering for a long time about the veracity of the Siege of Syracuse tale. It's right up there with the actual content of Greek Fire.
Almost every single example of "replication attempt" tries to do so in a closed-minded fashion. They think: what do we know the $society was capable of? How can we apply these means to the desired end?
Instead, they should be thinking: how, without very specific technological advances several inventive generations removed, might we accomplish the same ends? Instead, they're sitting there rolling copper and making guestimates about things like troop strength and the like.
This is exactly the problem with Office to OpenOffice migrations, at this point.
Arguably, Word deals with complex Word forms and the like better than OpenOffice, too. When you've got hundred+ page document form templates which autogenerate a lot of information, etc. moving to something else is difficult as well.
I've got a friend who ranches. He's technically savvy, but not what I'd call a genius. All his information is kept in excel worksheets, which auto-generate important data via formulas. He uses WinCE/WinMo devices (a number of them) to scan in and enter data when out in the barn or pasture (a common/daily occurrence) with the Pocket Excel (or w/e it's called). He's then able to examine and generate reports - using the same files. Often, he's accessing the files directly over a VPN on a Windows file share.
With Windows, it's not convoluted to do this. To do this in any other way would be difficult, because there's years of development involved. He's tried OO, and it won't due the trick because there's nothing in PDA form that does the trick short of a custom application.
Any organization small enough to have trouble funding and domain controller Doesn't need one.
Are you saying that the client/server model does not scale below a certain number of hosts?
I can see where a capital expense for an office of 5 or so people might be a bit severe, but where a domain is still necessary. Say, where there's a reason to have cleanly segregated user permissions/access controls (which Samba3 does not provide easily/without significant knowledge of how to set it up/reoccurring user admin costs).
I don't see that reducing a huge amount even when Samba 4 is released - there's a lot of configuration involved to get DHCP, DNS and Samba 4 talking to each other properly.
It takes less than 20 minutes to half an hour if you've already got DDNS running on the network.
To a large degree, yes, AD is the missing piece.
Pretty much everything relies upon Active Directory. Unfortunately, the s4 implementation isn't there yet for basic stuff, nevermind being able to use later versions of Exchange (which is, as far as I can tell, one of the only reasons why people are moving to 2k8r2 if they've already got a substantial 2k3 install base).
Pretty much the whole picture seems to be moving towards "linux" in one form or fashion. Popular virtualization technology, aside from the awkward bare metal Hyper-V, all runs on top of a Linux kernel (Xen, XenServer, KVM, VMWare). Aside from that, there's MS's HyperV and Sun/Oracle's offerings, which don't have nearly the incentive to use.
In many cases I've seen, even if someone is reliant upon AD and Exchange, they've already taken steps to marginalize their reliance: they've virtualized their AD and MSSQL servers; they've picked up things like Zimbra to replace Exchange (often transparently); and so on. I doubt it's happening so much in a corporate environment, but it's happening in a very big way in shops where there are 2-3 IT workers who have a little extra time (ie not running their heads off).
I remember going into a computer lab once in school, years ago. Windows 2000 was still the most common desktop OS.
I sat down at a workstation and opened the browser. Odd, I thought: the pixels didn't quite line up. Looking closer (at the browsers user agent string) I noticed it was actually Linux, and in all likelihood a themed ICEWM. THe workstation had Open Office (StarOffice, I think) and Mozilla, and that was about it - it was transparent enough and similar enough to a locked down w2k workstation that I'm sure the common user could not tell the difference.