Not certain it does exactly what you want, but take a look at DRBD. Device replication implemented as a device-mapper layer. I've been using it for about a year as underlying replicated shared storage for xen virtual machines, and found it very useful.
I would have to imagine your SAN is just doing uninformed readaheads.
Yah, if you want to hear me swear about SAN technology drop by one of those discussions some time. I'm not blaming ZFS for that, it's more an example where ZFS's best intentions and assumptions don't quite match reality, and the lack of a multilayer/changeable feature set gets a seriously adverse impact.
This issue with copy-on-write will be entirely sidestepped in a few years by flash storage's
Perhaps. But then again, perhaps not. I was personally stunned by the fact that SAN vendors did what they did with the caches (and let's just say, it's not just readaheads they do. And the cache segments are not small separate blocks, either you have the whole huge stripe or you don't). I'm sure it made sense in the days you were caching tape reels for mainframes, because believe me, that's where the tech is coming from.
So any file and block layers that aspire to be, in any way, generic, simply cannot make assumptions that the underlying layers aren't punchcards. Or people hacking away at stone tablets. Sometimes that's almost exactly what the admins will be stuck with, and playing politics in large organizations to be allowed to buy a thumper or something close to what was intended to be underlying isn't always feasible.
could benefit greatly from some of the features ZFS brings to the table.
Without a doubt, and for Solaris it's a huge step forward. I'm just saying it's not as clear-cut as the marketing:).
Hell, I would even love to have ZFS on my laptop for snapshotting and cloning.
Mmm, if you're looking for something that might be possible to use for that and you're on Linux, take a look at drbd. It's a device-replication layer (implemented as a device-mapper layer:). I'm personally using it for a shared-device semantics layer between machines below virtual machines (I could just go with straight iSCSI, but drbd gives me some advantages in prettiness and controllability), but I've been thinking about using it along the lines of laptop synchronization and/or remote replication. Having run it for about a year I'm impressed and surprised by its stability and capability to deal with me doing various nasties to it without negatively affecting data integrity.
I wouldn't be surprised if we stumble upon a few more neat ideas before we're through.
I'm sure we will, and as long as one is careful not to close doors on the way it's a good idea to transfer such hints and information between layers, but coming to depend on the assumptions is the risk that needs to be avoided.
Sure it is. For the security salesmen it's about convincing politicians and civil servants that they need to buy expensive security systems. Preferably with lots of blinking lights and even better, As Seen in the Movies, with technology that you can claim is sufficiently 'advanced' to justify the hefty pricetag.
For politicians it's another money/power making issue as they can justify sweeping spending and control with it. They're not overly difficult to talk into buying the pointless junk as it's not their money and they make their other gains by keeping people afraid.
People, to a large extent, don't actually give a crap about 'feeling safe' as far as it relates to abstract dangers like transportation security. Most engage in far riskier activities on an everyday basis; heck, terrorism rates about the same as accidental bathtub drownings as a risk over the last ten years. If it weren't for security salesmen and politicians, people would be a bit scared for a month or two and then get on with their lives. Much like other minor scares that we deal with.
Heck, I'd bet you find more people who've quit traveling due to the painfully retarded security, than you'd find not traveling because they're scared of security threats.
RAID-Z uses copy-on-write to avoid RAID 5's required read for every non-cached write.
Of course, the very same copy-on-write will also result in massive file fragmentation, carefully smearing your dbf files over the entire platters, making your SAN caches useless. Over time resulting in horrible read performance.
ZFS is certainly a huge improvement for anyone used to ufs and disksuite, but I have to say that using it in the real world it's not all it's cracked up to be. A more layered approach would have made it easier to switch in and out features that turned out to be misfeatures in certain situations.
Mixing together the features of various layers is, imo, no matter how tempting, simply the wrong approach. Proceed further along that road and you get to record based filesystems or even more special-purpose variants. I mean, there are even more optimizations that you can do if you know the _contents_ of the files. But once you go down that road complexity will grow with every possible different situation you need to handle and you end up either with something far too complex or something unsuitable for many cases. Better then to do the best you can without extra knowledge, code special layers for special features, assume nothing, and let the possibly competent admin add appropriate layers for appropriate data.
I would wonder how well any software on a non-mainframe platform (e.g. Oracle on Linux) would cope with time going backwards - surely this would screw up the logs etc?
Time on Unix platforms basically doesn't go backwards. It's defined as seconds since 1970, and the number of seconds since then goes only one way.
How that information is _presented_ to human consumers is a different matter, and it's at that layer the DST stuff is done. So presentations for human consumption like time stamps in logs can be ambiguous, depending on how the application writing the log decides to format its timestamps, but software never gets confused; things happened when they happened, no matter what that may get called depending on time zone or politics or country.
(Of course, if you changed the actual clock you'd run into the same type of problems on Unix, which is why it's not done that way.)
Actually, the lack of the ability to easily fork may be one of the most deeply ingrained flaws and problems with current political systems. It's a privilege largely reserved for corporations and/or the very rich, to easily change into and out of what political system you currently prefer.
It would be interesting to explore the options of more modular political systems where citizens, when they dislike their unit enough, could reasonably easily disengage and join another unit. A system could be designed on multiple dimensions ranging from geographic protection through healtcare through trade-related aspects, and comprise both low-level units up to world spanning organizations. If nothing else it might at least provide more interesting and intellectually challenging politics.
In the real world, there are single companies running hundreds of these mainframe CPUs. And they run at 80%+ busy 24 hours a day, by the way.
See, here's a perfect example of why people tend to be skeptical. You're saying hundreds like it's a lot. There are single companies running on thousands or tens of thousands of x86 CPUs, consolidated with VMware, Xen and other similar technologies and running at 80% load. It's not exceptional any more.
Then we get marketing from mainframe people about running a hundred thousand machines in a single mainframe. Yes, I could probably hack xen to let me do that too, but I don't expect it to have executed the bootloader for half of them until sometime during the next decade, so how about some serious data? I've had mainframe people say they could consolidate entire specific x86 parks into a piece of the mainframe when I know that one or two of those fully loaded x86 machines moved into the mainframe would mean we need another mainframe. It's like they think they have some magic smoke in their CPU's and that instructions magically get shorter in the mainframe. That light travels faster in the mainframe FC cards. This isn't 1989 anymore. The mainframe of today is as close as you can get to commodity hardware without having clones sold at Egghead.
The boring but wonderful truth is that -- surprise! -- different servers are good at different things!
Without a doubt. But the only way to determine what is good for what is by having actual benchmarks. Got any comparisons on mysql performance on a mainframe running 2 instances/cpu compared to a RHEL/xen setup with the same config? Web servers? IO? lmbench tests? Iozone runs? Simultaneous multiple kernel compilations? Anything?
Most of the very few mainframe related benchmarks items I've found on google have disappeared very quickly, for some reason.
Without any such data it's impossible to have a reasonable discussion about the advantages of any specific platform. The mainframe side keeps saying they're good at some things but without any solid data at all to back it up. Until such data becomes commonly available, I'm afraid the skeptics will go on having a field day on any mainframe-related story.
From what I've seen, weekly IPLs seems to be close to standard practice within the mainframe world. With a reboot frequency like that you're not going to even see most stability issues.
Has it even become possible to switch to and from daylight savings yet without rebooting the mainframe?
almost none of them have real evidence with which to backup their statement.
From what I've seen in a SAN in a mixed environment, it certainly isn't the mainframe that's exposing the FC bottlenecks. It's a long time ago that the mainframe had any special hardware.
Frankly, I've heard so many sales pitches for so long that there's only one thing that matters. Publish the benchmarks or it's just hot air.
In the case of mainframes I haven't seen any serious benchmarks for more than a decade. I expect performance to be entirely predictable from that point only.
"Riccitiello admitted that he personally doesn't like DRM, as it 'interrupts the user experience.' He also added that 'We would like to get around that. But there is this problem called piracy out there.'"
Of course, for the potential customers who 'would like to get around that', piracy isn't the problem, it's the solution.
Personally I have bought a number of EA games, but since they started using SecuROM they're permanently off the list. I have better things to do with my time and money.
When all else can be held equal, formal modeling can give you very powerful insights into causal mechanism and prediction.
Of course, as the very foundation upon which all the formal modeling is built is flawed, it's no surprise that it falls down. It's not math, it's not science, it's the very axiom of the banking system that's at fault.
Essentially every single thing that has gone wrong can be traced to, and derived from, fractional reserve banking and the central banks. Everything above that is built on the theory that credit is always expanding. People will always borrow more, allowing the central banks to hold rates below the free market cost of money, enforcing 'investment' (speculation) and spending as the appropriate ways to divest money. Credit expansion is what creates asset inflation, it is what leaves the banks insolvent, heck, it's even what creates demand for oversized extra expensive cars, and it is what leaves the models fundamentally flawed as they never consider the end game.
The end game comes when the last debtor cannot carry his debt any more. The models figure that, but they miss the fact that as the last debtor loses his ability to carry debt, the assets everyone else has purchased suddenly lose his added bid. Asset prices start falling. Then we lose those who could only carry their debt as their assets were appreciating. And asset prices start falling even more. Then we get credit destruction as debt defaults, assets are written down and credit gets constrained by balance sheet requirements (the banks that thought they were solvent suddenly realize they're sitting on a pile of assets that's not worth anywhere near what they were when the last sucker paid his price). When they go over their books they realize that they, and probably everyone else, has nowhere near the money they need to repay their debts.
Suddenly credit isn't expanding anymore. And the central banks lose control over their ability to control interest rates; when credit isn't infinite you pay the market price or you don't get the money. And the whole house of cards comes crashing down.
Of course, it's entirely possible to build a rational model of a fractional reserve system too, but as that would require debt ratings that go down as economic leverage goes up it'd essentially negate the exact thing the FRB system and central banks try to accomplish. It would basically mean that risk rating and cost of money would go up as the central banks tried to lower interest rates and inflate. We would have had the whole economy on a B rating in 1999, and on C last year.
It'd be saner just to move to full reserve banking. I wish. I'm not holding my breath.
Apparently after 400 recharges the battery is down to 80% of its life
Li-Ion batteries are not so much affected by recharges as they are by temperature and sheer age. On the plus side that means you're going to get a lot out of it if you use it a lot.
On the negative side it also means your battery is going to go bad in 18-24 months even if you only charge it once per month. (unless you keep it in the freezer, which is the appropriate storage for unused li-ion batteries)
Then again, you replace your phone every 18 months
You do? I'm on my, hm, fourth since 1994, I think.
The Logitech diNovo Edge keyboard is, as far as I've been able to tell, another device with the same problem.
Of course, that's also the reason I haven't bought one. There's simply no way I'm throwing $200 at a keyboard with the battery longevity of li-ion tech.
So I certainly welcome this directive; it'll save me the trouble of having to avoid certain products I'd otherwise buy.
In a foray off Asus I bought an MSI motherboard (K9N Neo F). When I later wanted to put a new CPU in it, I looked up how to upgrade the BIOS. Turns out you _have_ to have Windows XP, and that the only ways to upgrade the BIOS is with an XP app, or with an ActiveX control.
No boot-from-dos-floppy to flash. No BIOS builtin flasher. No flash from CD. You have to have XP.
Since then I'm not touching MSI hardware (most it is reputedly much better, but the mindset behind a motherboard that cant be flashed outside an OS simply makes me doubt their capabilities and intentions).
That would indeed be a much better way to determine the quality of people you'd be working for. There are semi-legitimate reasons for a company to not want to donate code under BSD licence, but maybe they'd be ok with donating it under GPL? Perhaps there's a middle ground to be reached where their desire to avoid getting abused by comptetitors, and the posters ethical dilemma can both be resolved equitably.
Or, of course, it may turn out they don't want to negotiate at all, and that they're just looking for a free ride. In which case I'd suggest they're probably also looking for a free ride off anyone they employ, and I'd be careful signing anything they put forward.
Sometimes seeing what they want to do to others is a good step in avoiding getting it done to you.
Not sure if it suits your situation, but you could take a look at drbd. Current stable versions only support 2 node mirroring, but future versions are planned with further nodes.
Personally I've used it for shared-device semantics for backing storage on Xen VM's (and prefer it over my previous iSCSI-in-VM config). It is also, however, eminently suitable for remote-site mirroring of block devices. It isn't too difficult to build a stack with backing devices remote-mirrored over drbd, shared out over iscsi in the local location and layered with whatever you want on top of that. Essentially the equivalent of what some storage vendors do with remote-mirrored SAN stuff.
And in my experience this far (after about a year of running it) drbd is stable regardless of what I throw at it.
Instead we're faced with the new controversy that every skin cell you shed can be considered an embryo that, with the correct application of medical science, can now become a child.
Scratching yourself will now mean you're killing babies!
Or, perhaps you're trying to create an evil clone army with all those cells?
There's plenty of material and interpretations for anyone who wants to find controversy.
SDK comparison wise there's no competition. The openmoko is basically a linux machine with a touchscreen and a GSM chip. Anything you can do with a Linux machine you can do with the Moko, Qt, gtk, shells, perl, python, etc. If you lack it you can port it. Forget special-purpose limited devices, this is the real deal, a full general purpose computer.
However, if you compare them as phones... well, the openmoko is basically a linux machine with a touchscreen and a GSM chip. After you've figured out how to flash it and decided on what distribution to run you get to debug alsa routing to bluetooth headset connections, configure gps daemons, etc. It's a good thing that it's as networked as it is (you can network over gsm, wlan, bluetooth, and what I use mostly, ethernet over USB (you'll want it in an USB port so it's charged anyway)) because you want a computer with multiple ssh sessions connected for many of the things you'll be doing on it.
Personally I'm not the least interested in getting either an Android based phone or the iPhone (or any other locked down proprietary crap), and for me the Freerunner is among the coolest things I've ever played with (and the first phone I've ever wanted to spend a cent of my own money on). The potential is enormous, and the way the base can make it into ubiquitous devices of all kinds makes me think it can become something that influences the future in a serious way. But it's not an end-user product yet, and I think it'll take a while before it's there.
that cost everyone hundreds of thousands of dollars?
Of course, the RIAA corps actually cost everyone approximately 10 billion dollars per year. Subtract the legitimate expenditure going to artists and composers (which would be the cost if we paid just them and handled distribution over p2p) and we end up with the RIAA corps costing the economy around 8 billion.
Intellectual monopoly taxation effects are a significant burden on the economy these days; handing out monopoly rights isn't exactly free.
The quality difference just isn't there, DVD has reached 99% of the acceptable perceived quality for what a consumer expects to see from a movie.
Considering that most 'high quality' DVD rips or HD transcodes aren't even DVD resolution, I'd say it's even better than that. If it's not even worth the extra space for most people to go beyond 1200kbps xvid with appropriate resolution, then it's hard to see compelling reasons to think most people would pay any form of a premium for HD.
Anything non-government issued? Since last year, no way, no how.
If AIG goes under, its backing becomes worthless
Yes, well, anyone who has actually read what's been going on should have moved out of any and all collateralized obligations. The credit unwind is going ahead and I frankly don't think it's possible to stop. If you want to keep your bonds, why should I pay for your bad investment when it sours?
Even today people are complacent, corporate managers are still sticking money in money market funds to gain a little extra interest until payday, people buy bonds that the banks are hawking instead of FDIC insured accounts, etc. What part of 'they're not safe' are they not understanding? You want the extra interest, _you_ take the hit when it turns out that, oops, Lehmans, AIG, Merril, GS, BAC didn't actually have any money to pay you back because they've been leveraged 25-1 and got wiped since their business model didn't account for the concept that things could go down as well as up.
Subprime is barely an issue, nor has it ever been the main one; grossly overstated asset valuations are the main problem and we still have a long way to go until they reach the historic equilibrium in relation to wages and inflation. Bailout or not, you can expect any asset-backed bonds to lose significant amounts of that backing, and there simply isnt an entity in existence that can make up for that (not even the Fed and treasury, their final backstop is to inflate the debt away).
It's not so much about royalties but about control. The internet allows unlimited unique channels custom created for individual listeners. This diminishes the control the big players have over the transmissions; if there's unlimited channel space, everyone can get playtime. This translates into wider music taste among the audience, room for independent bands, lower profit margins and less efficient marketing. Not in the big players interest.
Slashdot Mirror
namely replication
Not certain it does exactly what you want, but take a look at DRBD. Device replication implemented as a device-mapper layer. I've been using it for about a year as underlying replicated shared storage for xen virtual machines, and found it very useful.
I would have to imagine your SAN is just doing uninformed readaheads.
Yah, if you want to hear me swear about SAN technology drop by one of those discussions some time. I'm not blaming ZFS for that, it's more an example where ZFS's best intentions and assumptions don't quite match reality, and the lack of a multilayer/changeable feature set gets a seriously adverse impact.
This issue with copy-on-write will be entirely sidestepped in a few years by flash storage's
Perhaps. But then again, perhaps not. I was personally stunned by the fact that SAN vendors did what they did with the caches (and let's just say, it's not just readaheads they do. And the cache segments are not small separate blocks, either you have the whole huge stripe or you don't). I'm sure it made sense in the days you were caching tape reels for mainframes, because believe me, that's where the tech is coming from.
So any file and block layers that aspire to be, in any way, generic, simply cannot make assumptions that the underlying layers aren't punchcards. Or people hacking away at stone tablets. Sometimes that's almost exactly what the admins will be stuck with, and playing politics in large organizations to be allowed to buy a thumper or something close to what was intended to be underlying isn't always feasible.
could benefit greatly from some of the features ZFS brings to the table.
Without a doubt, and for Solaris it's a huge step forward. I'm just saying it's not as clear-cut as the marketing :).
Hell, I would even love to have ZFS on my laptop for snapshotting and cloning.
Mmm, if you're looking for something that might be possible to use for that and you're on Linux, take a look at drbd. It's a device-replication layer (implemented as a device-mapper layer :). I'm personally using it for a shared-device semantics layer between machines below virtual machines (I could just go with straight iSCSI, but drbd gives me some advantages in prettiness and controllability), but I've been thinking about using it along the lines of laptop synchronization and/or remote replication. Having run it for about a year I'm impressed and surprised by its stability and capability to deal with me doing various nasties to it without negatively affecting data integrity.
I wouldn't be surprised if we stumble upon a few more neat ideas before we're through.
I'm sure we will, and as long as one is careful not to close doors on the way it's a good idea to transfer such hints and information between layers, but coming to depend on the assumptions is the risk that needs to be avoided.
It's not about money.
Sure it is. For the security salesmen it's about convincing politicians and civil servants that they need to buy expensive security systems. Preferably with lots of blinking lights and even better, As Seen in the Movies, with technology that you can claim is sufficiently 'advanced' to justify the hefty pricetag.
For politicians it's another money/power making issue as they can justify sweeping spending and control with it. They're not overly difficult to talk into buying the pointless junk as it's not their money and they make their other gains by keeping people afraid.
People, to a large extent, don't actually give a crap about 'feeling safe' as far as it relates to abstract dangers like transportation security. Most engage in far riskier activities on an everyday basis; heck, terrorism rates about the same as accidental bathtub drownings as a risk over the last ten years. If it weren't for security salesmen and politicians, people would be a bit scared for a month or two and then get on with their lives. Much like other minor scares that we deal with.
Heck, I'd bet you find more people who've quit traveling due to the painfully retarded security, than you'd find not traveling because they're scared of security threats.
Whether it can be bought or not is perhaps besides the point.
Because it can certainly be sold.
RAID-Z uses copy-on-write to avoid RAID 5's required read for every non-cached write.
Of course, the very same copy-on-write will also result in massive file fragmentation, carefully smearing your dbf files over the entire platters, making your SAN caches useless. Over time resulting in horrible read performance.
ZFS is certainly a huge improvement for anyone used to ufs and disksuite, but I have to say that using it in the real world it's not all it's cracked up to be. A more layered approach would have made it easier to switch in and out features that turned out to be misfeatures in certain situations.
Mixing together the features of various layers is, imo, no matter how tempting, simply the wrong approach. Proceed further along that road and you get to record based filesystems or even more special-purpose variants. I mean, there are even more optimizations that you can do if you know the _contents_ of the files. But once you go down that road complexity will grow with every possible different situation you need to handle and you end up either with something far too complex or something unsuitable for many cases. Better then to do the best you can without extra knowledge, code special layers for special features, assume nothing, and let the possibly competent admin add appropriate layers for appropriate data.
I would wonder how well any software on a non-mainframe platform (e.g. Oracle on Linux) would cope with time going backwards - surely this would screw up the logs etc?
Time on Unix platforms basically doesn't go backwards. It's defined as seconds since 1970, and the number of seconds since then goes only one way.
How that information is _presented_ to human consumers is a different matter, and it's at that layer the DST stuff is done. So presentations for human consumption like time stamps in logs can be ambiguous, depending on how the application writing the log decides to format its timestamps, but software never gets confused; things happened when they happened, no matter what that may get called depending on time zone or politics or country.
(Of course, if you changed the actual clock you'd run into the same type of problems on Unix, which is why it's not done that way.)
Actually, the lack of the ability to easily fork may be one of the most deeply ingrained flaws and problems with current political systems. It's a privilege largely reserved for corporations and/or the very rich, to easily change into and out of what political system you currently prefer.
It would be interesting to explore the options of more modular political systems where citizens, when they dislike their unit enough, could reasonably easily disengage and join another unit. A system could be designed on multiple dimensions ranging from geographic protection through healtcare through trade-related aspects, and comprise both low-level units up to world spanning organizations. If nothing else it might at least provide more interesting and intellectually challenging politics.
In the real world, there are single companies running hundreds of these mainframe CPUs. And they run at 80%+ busy 24 hours a day, by the way.
See, here's a perfect example of why people tend to be skeptical. You're saying hundreds like it's a lot. There are single companies running on thousands or tens of thousands of x86 CPUs, consolidated with VMware, Xen and other similar technologies and running at 80% load. It's not exceptional any more.
Then we get marketing from mainframe people about running a hundred thousand machines in a single mainframe. Yes, I could probably hack xen to let me do that too, but I don't expect it to have executed the bootloader for half of them until sometime during the next decade, so how about some serious data? I've had mainframe people say they could consolidate entire specific x86 parks into a piece of the mainframe when I know that one or two of those fully loaded x86 machines moved into the mainframe would mean we need another mainframe. It's like they think they have some magic smoke in their CPU's and that instructions magically get shorter in the mainframe. That light travels faster in the mainframe FC cards. This isn't 1989 anymore. The mainframe of today is as close as you can get to commodity hardware without having clones sold at Egghead.
The boring but wonderful truth is that -- surprise! -- different servers are good at different things!
Without a doubt. But the only way to determine what is good for what is by having actual benchmarks. Got any comparisons on mysql performance on a mainframe running 2 instances/cpu compared to a RHEL/xen setup with the same config? Web servers? IO? lmbench tests? Iozone runs? Simultaneous multiple kernel compilations? Anything?
Most of the very few mainframe related benchmarks items I've found on google have disappeared very quickly, for some reason.
Without any such data it's impossible to have a reasonable discussion about the advantages of any specific platform. The mainframe side keeps saying they're good at some things but without any solid data at all to back it up. Until such data becomes commonly available, I'm afraid the skeptics will go on having a field day on any mainframe-related story.
"Hey, can we improve the SLA (Service Level Agreement) here?"
Usually, the 'scheduled' IPL's don't count into the SLA uptime measurements, so it doesn't seem like there's much pressure to avoid them.
From what I've seen, weekly IPLs seems to be close to standard practice within the mainframe world. With a reboot frequency like that you're not going to even see most stability issues.
Has it even become possible to switch to and from daylight savings yet without rebooting the mainframe?
almost none of them have real evidence with which to backup their statement.
Yes, well, rather like mainframe benchmarks, eh?
From what I've seen in a SAN in a mixed environment, it certainly isn't the mainframe that's exposing the FC bottlenecks. It's a long time ago that the mainframe had any special hardware.
Frankly, I've heard so many sales pitches for so long that there's only one thing that matters. Publish the benchmarks or it's just hot air.
In the case of mainframes I haven't seen any serious benchmarks for more than a decade. I expect performance to be entirely predictable from that point only.
of course more gamers don't like it
Heck, even he himself doesn't like it:
"Riccitiello admitted that he personally doesn't like DRM, as it 'interrupts the user experience.' He also added that 'We would like to get around that. But there is this problem called piracy out there.'"
Of course, for the potential customers who 'would like to get around that', piracy isn't the problem, it's the solution.
Personally I have bought a number of EA games, but since they started using SecuROM they're permanently off the list. I have better things to do with my time and money.
When all else can be held equal, formal modeling can give you very powerful insights into causal mechanism and prediction.
Of course, as the very foundation upon which all the formal modeling is built is flawed, it's no surprise that it falls down. It's not math, it's not science, it's the very axiom of the banking system that's at fault.
Essentially every single thing that has gone wrong can be traced to, and derived from, fractional reserve banking and the central banks. Everything above that is built on the theory that credit is always expanding. People will always borrow more, allowing the central banks to hold rates below the free market cost of money, enforcing 'investment' (speculation) and spending as the appropriate ways to divest money. Credit expansion is what creates asset inflation, it is what leaves the banks insolvent, heck, it's even what creates demand for oversized extra expensive cars, and it is what leaves the models fundamentally flawed as they never consider the end game.
The end game comes when the last debtor cannot carry his debt any more. The models figure that, but they miss the fact that as the last debtor loses his ability to carry debt, the assets everyone else has purchased suddenly lose his added bid. Asset prices start falling. Then we lose those who could only carry their debt as their assets were appreciating. And asset prices start falling even more. Then we get credit destruction as debt defaults, assets are written down and credit gets constrained by balance sheet requirements (the banks that thought they were solvent suddenly realize they're sitting on a pile of assets that's not worth anywhere near what they were when the last sucker paid his price). When they go over their books they realize that they, and probably everyone else, has nowhere near the money they need to repay their debts.
Suddenly credit isn't expanding anymore. And the central banks lose control over their ability to control interest rates; when credit isn't infinite you pay the market price or you don't get the money. And the whole house of cards comes crashing down.
Of course, it's entirely possible to build a rational model of a fractional reserve system too, but as that would require debt ratings that go down as economic leverage goes up it'd essentially negate the exact thing the FRB system and central banks try to accomplish. It would basically mean that risk rating and cost of money would go up as the central banks tried to lower interest rates and inflate. We would have had the whole economy on a B rating in 1999, and on C last year.
It'd be saner just to move to full reserve banking. I wish. I'm not holding my breath.
Apparently after 400 recharges the battery is down to 80% of its life
Li-Ion batteries are not so much affected by recharges as they are by temperature and sheer age. On the plus side that means you're going to get a lot out of it if you use it a lot.
On the negative side it also means your battery is going to go bad in 18-24 months even if you only charge it once per month. (unless you keep it in the freezer, which is the appropriate storage for unused li-ion batteries)
Then again, you replace your phone every 18 months
You do? I'm on my, hm, fourth since 1994, I think.
The Logitech diNovo Edge keyboard is, as far as I've been able to tell, another device with the same problem.
Of course, that's also the reason I haven't bought one. There's simply no way I'm throwing $200 at a keyboard with the battery longevity of li-ion tech.
So I certainly welcome this directive; it'll save me the trouble of having to avoid certain products I'd otherwise buy.
... and then they forget to deduct the jobs that wouldn't exist in other industries because money is spent on MAFIAA monopoly rights.
I doubt it would cost them anything.
In a foray off Asus I bought an MSI motherboard (K9N Neo F). When I later wanted to put a new CPU in it, I looked up how to upgrade the BIOS. Turns out you _have_ to have Windows XP, and that the only ways to upgrade the BIOS is with an XP app, or with an ActiveX control.
No boot-from-dos-floppy to flash. No BIOS builtin flasher. No flash from CD. You have to have XP.
Since then I'm not touching MSI hardware (most it is reputedly much better, but the mindset behind a motherboard that cant be flashed outside an OS simply makes me doubt their capabilities and intentions).
That would indeed be a much better way to determine the quality of people you'd be working for. There are semi-legitimate reasons for a company to not want to donate code under BSD licence, but maybe they'd be ok with donating it under GPL? Perhaps there's a middle ground to be reached where their desire to avoid getting abused by comptetitors, and the posters ethical dilemma can both be resolved equitably.
Or, of course, it may turn out they don't want to negotiate at all, and that they're just looking for a free ride. In which case I'd suggest they're probably also looking for a free ride off anyone they employ, and I'd be careful signing anything they put forward.
Sometimes seeing what they want to do to others is a good step in avoiding getting it done to you.
Not sure if it suits your situation, but you could take a look at drbd. Current stable versions only support 2 node mirroring, but future versions are planned with further nodes.
Personally I've used it for shared-device semantics for backing storage on Xen VM's (and prefer it over my previous iSCSI-in-VM config). It is also, however, eminently suitable for remote-site mirroring of block devices. It isn't too difficult to build a stack with backing devices remote-mirrored over drbd, shared out over iscsi in the local location and layered with whatever you want on top of that. Essentially the equivalent of what some storage vendors do with remote-mirrored SAN stuff.
And in my experience this far (after about a year of running it) drbd is stable regardless of what I throw at it.
Instead we're faced with the new controversy that every skin cell you shed can be considered an embryo that, with the correct application of medical science, can now become a child.
Scratching yourself will now mean you're killing babies!
Or, perhaps you're trying to create an evil clone army with all those cells?
There's plenty of material and interpretations for anyone who wants to find controversy.
Well, I'm playing with my Freerunner right now. :)
SDK comparison wise there's no competition. The openmoko is basically a linux machine with a touchscreen and a GSM chip. Anything you can do with a Linux machine you can do with the Moko, Qt, gtk, shells, perl, python, etc. If you lack it you can port it. Forget special-purpose limited devices, this is the real deal, a full general purpose computer.
However, if you compare them as phones... well, the openmoko is basically a linux machine with a touchscreen and a GSM chip. After you've figured out how to flash it and decided on what distribution to run you get to debug alsa routing to bluetooth headset connections, configure gps daemons, etc. It's a good thing that it's as networked as it is (you can network over gsm, wlan, bluetooth, and what I use mostly, ethernet over USB (you'll want it in an USB port so it's charged anyway)) because you want a computer with multiple ssh sessions connected for many of the things you'll be doing on it.
Personally I'm not the least interested in getting either an Android based phone or the iPhone (or any other locked down proprietary crap), and for me the Freerunner is among the coolest things I've ever played with (and the first phone I've ever wanted to spend a cent of my own money on). The potential is enormous, and the way the base can make it into ubiquitous devices of all kinds makes me think it can become something that influences the future in a serious way. But it's not an end-user product yet, and I think it'll take a while before it's there.
that cost everyone hundreds of thousands of dollars?
Of course, the RIAA corps actually cost everyone approximately 10 billion dollars per year. Subtract the legitimate expenditure going to artists and composers (which would be the cost if we paid just them and handled distribution over p2p) and we end up with the RIAA corps costing the economy around 8 billion.
Intellectual monopoly taxation effects are a significant burden on the economy these days; handing out monopoly rights isn't exactly free.
The quality difference just isn't there, DVD has reached 99% of the acceptable perceived quality for what a consumer expects to see from a movie.
Considering that most 'high quality' DVD rips or HD transcodes aren't even DVD resolution, I'd say it's even better than that. If it's not even worth the extra space for most people to go beyond 1200kbps xvid with appropriate resolution, then it's hard to see compelling reasons to think most people would pay any form of a premium for HD.
Do you own any bonds?
Anything non-government issued? Since last year, no way, no how.
If AIG goes under, its backing becomes worthless
Yes, well, anyone who has actually read what's been going on should have moved out of any and all collateralized obligations. The credit unwind is going ahead and I frankly don't think it's possible to stop. If you want to keep your bonds, why should I pay for your bad investment when it sours?
Even today people are complacent, corporate managers are still sticking money in money market funds to gain a little extra interest until payday, people buy bonds that the banks are hawking instead of FDIC insured accounts, etc. What part of 'they're not safe' are they not understanding? You want the extra interest, _you_ take the hit when it turns out that, oops, Lehmans, AIG, Merril, GS, BAC didn't actually have any money to pay you back because they've been leveraged 25-1 and got wiped since their business model didn't account for the concept that things could go down as well as up.
Subprime is barely an issue, nor has it ever been the main one; grossly overstated asset valuations are the main problem and we still have a long way to go until they reach the historic equilibrium in relation to wages and inflation. Bailout or not, you can expect any asset-backed bonds to lose significant amounts of that backing, and there simply isnt an entity in existence that can make up for that (not even the Fed and treasury, their final backstop is to inflate the debt away).
It's not so much about royalties but about control. The internet allows unlimited unique channels custom created for individual listeners. This diminishes the control the big players have over the transmissions; if there's unlimited channel space, everyone can get playtime. This translates into wider music taste among the audience, room for independent bands, lower profit margins and less efficient marketing. Not in the big players interest.