B&N did the same to me a couple years ago with the touchpad fire sale. They even took my money and had to refund it a month later when they discovered they sold a million of the things but only had 5 in stock.
However, not everyone can spend about $6000 for the drive
Vs hiring a bunch of engineers to build an optical jukebox, which is what happened in the article in question...
I don't know, i'm betting you could have bought a really big tape library with a lot of tapes/drives for the price it cost facebook to build that thing using $100k+/year engineers.
T10000 tapes are the same size and store 5TB uncompressed
Now see you missed one of the cool things about the "enterprise" drives (vs LTO). Those T1k carts you purchased last year that stored 5TB, can now magically store 8.5 TB uncompressed with the latest T1kD tape drive.
Spend a few tens of thousands on new drives increases the capacity of your existing media investment by 60%, not bad if you happen to have a exabyte of capacity.
What we need are some judges that realize that just because you didn't find the critical security hole when you tested said product, doesn't mean that the company isn't liable for assuring that the product you purchased remains useful for its intended purpose.
Instead what we have are companies giving you shit, and then trying to charge you to fix their mistakes. Most other industries worked this out years ago, someone discovers your refrigerator's defrost unit catches fire and burns peoples houses down after its been in the market for a few years, the original company is generally on the hook to repair it.
A few years ago, GE sent me a nice notice telling me that I should immediately contact a repair organization to replace the door on my 10 year old dish washer, or alternatively they would give me a $200 discount off the purchase of a new unit... All because it turns out they were catching fire.
Besides the difficulty of dealing with 174 bluray disks instead of 1 tape... You have to wonder about the reliability of those disks sitting around on a shelf for ten years..
Oh, and you can write said tapes at 500+ MByte/sec.
Plus, tape is well understood, and there are tons of media management applications that track whats on the tape, when it expires, where its located, what encryption keys are used to decrypt it.. Basically 40 years of data management infrastructure.
However, Software wise z/OS is now so far behind it gets harder and harder to find people who want to work with it.
I agree. No one expects to have to write assembly to basically roll some logs, but that is exactly what I found myself doing a couple years ago. Hacking up the example hooks to install a piece of functionality that should be part of any sane OS.
Or for that matter spending 2 days installing/configuring z/VM and editing the configuration with text editors that would have been put to shame by the editors found on home computers of the early 80s. All this for a configuration similar to what could be archived on a PC in less than 15 minutes with few mouse clicks and vmware.
Really? Didn't know that, I thought you had to have a POWER compute node in your pureflex system to run IBM i... Do you have a link indicating that it will run on the x86 compute nodes ?
Frankly, i'm skeptical. I've seen the presentations and I don't ever remember seeing anything about IBM i/i5/i5OS/as400 nodes being hosted on x86, only on POWER. Sure you can put them in the same cabinet, but that is really nothing new, just now you can manage them from the same console. With a crapload of limitations of course. It takes about 3 seconds to find all kinds of issues with the way IBM is managing the systems.. I blew some minds when I asked if I could manage vmware and powerVM instances from the same console (the answer was no at the time).
You mean z and i.... p has been struggling due to the onslaught of x86's with Linux.
Sure p/i are the same hardware now and can run linux too. But in the end its far easier to switch from linux/power or aix/power to linux/x86 than it is to switch from IBM i/zos to some other platform.
So, both of those markets are nice, but IBM will continue to try an milk them like HP does with nonstop/openvms/etc. Its only a matter of time before the margins on those start to get squeezed as people decide paying ten million dollars for something they can port to linux and run on $50k a year of hardware is stupid. Or more likely, a 3rd party starts providing similar services as the in house applications and the bean counters decide to outsource the whole shebang (seen that a couple times recently).
Frankly, both IBM i and AIX could be ported to x86 and the whole power hardware division go away in 5 or 6 years while still maintaining the existing software lock.
zOS, probably not so much unless they use one of the emulation technologies they have locked away in the vault.
I think, on the whole, it's a moot point and we could all do with less Lennart Pottering-esque conflating of server vs. desktop priorities.
I think boot times matter on server gear too, and for a while so did IBM, back when they were pushing availability numbers because a single hour long reboot, once a year can put the system availability in the three nines category. Sometimes the hardware needs actual reboots be that for firmware or hardware upgrades... None of my machines have physical hotplug memory for example... And none of my PCIe x86's have hotplug PCIe, even though a fair number of the PCI-X machines did.
When my desktop takes 5 minutes to boot, one person is inconvenienced. When my database server takes 30 minutes to boot, potentially thousands are inconvenienced.
BTW: I have a zSeries too, and we don't have good power conditioning on it, and it seems to take a power hit once or twice a year. That thing is a PITA because the "DASD" takes the better part of two hours to come back up after hard power loss.
If you hadn't noticed it seems most cars made in the last few years have custom sized radio's with buttons on the steering wheels USB/ipod ports in the glove compartments/etc. I think the need to design a space that holds a GPS screen and/or a radio was the final blow.
The old standard form factor radio is dead is most cars. Meaning the days of aftermarket stereos are as limited.
This seems to be a common practice in a lot of industry lately. Take a standard and "improve" it in a dozen different ways so that the market is fragmented and every manufacture gets their own way of doing something.
I suspect that this will be the case for most of these "upgrades".
No, crap.. Some of these UEFI servers are amazing...
Amazing that they can take 5+ minutes to get to the grub menu.
Of course the RS6k/pSeries machines I've used are worse. Stack a few IO drawers/FC cards in a machine and be prepared to wait a 30 minutes to boot to AIX.
I didn't realize dell was better. I'm going to have to give them another look. We have HP, IBM, and supermicro.
The supermicros we have boot in 1/10th the time with similar system configurations.
But still, we don't make money off the server itself, so why make them?
If this is what passes for IBM logic these days, then the sooner they die the better.
You sell/manufacture them, so that you can bundle higher margin products and provide a one stop shopping/support experience. How many Dells are running Tivoli software, or better yet how many places are running Dells and hiring IBM to manage them? I'm betting those numbers are far lower than the places with IBM hardware.
So cry me a river, Lenovo is doing just fine with the "low margin" PC's business they purchased from IBM. In fact they are kicking IBM's ass, their revenue numbers have been increasing ~13% per year for the last few years while IBM's revenue numbers have been flat for a decade.
If IBM management had any creativity, they would figure out how to in source some of that manufacturing and streamline it rather than giving foxconn/whoever a 5% cut then crying their isn't any margin left.
The problem with these solutions is that when you want to actually share data (think clipboards, or word documents) you have to poke huge holes in the security model to get them in/out of the container. That is the fundamental issue with windows, either you allow applications the ability to extend the environment (think adding thumbnail viewers for proprietary image formats), share data, or even do some level of application embedding/etc it becomes very difficult to secure that environment from a rogue application.
. Above that amount, gasoline engines start to have issues with Ethanol. Rubber seals, hoses and plastic parts in fuel systems start having reduced lifespan. Above 10% some engines start having other internal issues.
Tell that to my lawn equipment that needs to have the little rubber diaphragms used as vacuum fuel pumps, the little polymer priming bulbs, and even the rubber hoses replaced on a 6 month to 3/4 year schedule when run on the 10% mixture at the gas station. Yes, and some of these devices have been manufactured in the last couple years.
Those idiots in congress think the only gas engines that matter are the ones sold by the automobile manufactures in the last 5 years. Its a lot harder to build a small engine with metal fuel lines and hard gaskets than it is to build a car engine with fixed fuel lines, large fuel pumps and injectors.
The result is that i'm paying ~15 a gallon at the local lawn and tractor supply store to avoid ethanol. Funny thing is I put a couple gallons in my old toyota and it runs _MUCH_ better on that stuff than what is available at the gas pump...
I'm all for building an ethanol fuel economy, just don't fscking do it by putting that shit in the fuel I need for my _gasoline_ engines.
MVS was supported for 34 years, from 1966 to 2000, when it became replaced by z/OS..
Its not really the same thing, MVS went through lots of revisions in that time period, and in many ways zos is just another step in that list of revisions (aka add a big feature and rename the OS). The early versions of zos could run on non 64-bit hardware and were basically identical to mvs.. Its just that support for legacy systems were dropped along the way. Even the applications have sort of moved on and now require the newer zos features.
That said, you could take a system programmer from the 1970's and put them in front of a 3270 emulator connected to a zBC12 and besides the fact its in color they wouldn't know that the machine is in a couple racks rather than taking up 40k sq/feet in the basement.
I keep reading posts on here talking about how god damned expensive RD-RAM was in the early 2000s and I just don't seem to remember it ALWAYS being that expensive, especially in summer '02
The price of RD-RAM came in line with DDR once the P4 got a chipset that could support DDR (which was after the athlon, which was already laying down the smack). Which of course pointed at the underlying agenda... Locking up who could manufacture the technology so the profit margins could be controlled. (see socket 8 licensing fiasco, itanium, etc)
IIRC the price of rambus memory literally halved in a month or two. Plus, people started talking about the latency issues, and the power consumption. By middle of 02, IIRC the price difference was basically non existent, and everyone was looking forward to DDR2.
Code samples are usually in "good" change logs as well.
Code samples for what? The library your selling? I don't think i've ever put a "code sample" in a changelog for human consumption. That is probably because everywhere I've worked, one of the first things I make sure works is a way for the source management, bug tracking, and release management systems to reference each other. That way, the internal change log has references to the bugs and patches. Curious what the code change associated with something is? Click the link and look at the diffs.
The only games in town for server Linux is now RHEL (CentOS or SciLinux if you are cheap) or Debian.
I wonder what environment you run servers in because, from where I stand (multiple platforms/etc) really the only two choices for corporate servers running linux are RHEL and SLES. Probably 1/2 the hardware I run isn't supported using linux with anything else. There is a little support for oracle linux if you live in that world, or maybe even a little xenserver, but debian doesn't even show up on the radar in most cases. I don't think I've ever seen debian running on a production server at a publicly listed company.
For example, using a SAN with qlogic cards? Where are the debian drivers (I see RHEL, SLES and Oracle)?
And a lot of people still have that sick taste in their mouth over the whole SUSE+Microsoft affair.
And does the sick taste extend to the closed source nvidia driver your installing to get opengl to work properly, or maybe the firmware on your wireless card with unpublished specifications, or how about the fact that the NSA is spying on everything you type on the internet? Did you stop using the internet? Do you refuse to use android because MS is getting a license fee (rather than cross licensing the technology) from your purchase?
Patent cross licenses are a fact of life in the computer industry, yes they suck, but in the end as a _USER_ of the OS it helps _ME_ too because I don't have to worry about MS sending a cease and desist to Suse/Attachmate and forcing them to pull support for samba/whatever. Or, for that matter putting them out of business.
Frankly, the whole thing is a pretty weak excuse to ignore Suse, especially in a corporate environment where the agreement actually should be reducing uncertainty in purchasing decisions.
The latest version opensuse actually is the best linux I have ever run, and that counts for a lot having run every major distribution since when the kernel was in the.9x timefame. That also includes all the recent versions of Ubuntu/mint/etc. It falls closer to the "it just works" mantra than any previous version (of course a few things still have hickups).
No one talks about Suse because we are off talking about more exciting things. That is the problem with having a stable sensible distribution that actually works.... Its doesn't have the latest $sexy to ignite peoples fires, or the latest $sucky to piss everyone off.
Personally, I suspect a fair number of people drop suse when they thought KDE jumped the shark a few years back. Now that it turns out its Gnome that jumped the shark no one remembers the one remaining major KDE based distribution.
Finally, there is SLES which is all the goodness of opensuse combined with long term vendor support as good as what is provided by redhat.
but takes 2x as long to completely refresh the screen.
Sometimes true but not always.. most 1080i content is 1080i/60 which means 60 fields per second. Lots of 1080p content is 24 or 30 frames per second. That actually means that a 1080i/60 image is refreshing the screen faster than a 1080p/24 (most bluray movies without 2:3 pulldown). With the 2:3 pulldown (pretty common on bluray) its the same but the pulldown induces "jutter".
The usual problem with "interlaced" content on LCDs is the fact that the LCD is displaying both fields simultaneously (often due to panels with max 30fps refresh). Run it with alternating black lines or some of the motion processing (usually included on 120 or 240 hz tvs) and it can look better than the 24p content. With native 30p content, sending it at 60i is basically indistinguishable..
Basically if you want a clean movie transfer you need an 24p bluray disk/player and a TV with a 24p conversion with black frame insertion.
Yah, all these naysayers need to get their eyes checked. I have crappy eyes and looking at the demo sets at frys at a 6-8' viewing distance its pretty amazing. They have some loop where there is a crowd scene and its possible to actually see what people are wearing rather than seeing "guy with red shirt" its "guy with striped red shirt". Some of the nature scenes are amazing too.
Personally, I think the old 720vs1080 arguments made during the bluray/hddvd vs DVD upgrade were more credible than the current crop of "4k is useless" people.
To me , it sounds like it could self discharge, which contrary to the "couple%" a month folks, the self discharge rates are dependent on how charged the battery is. Its not unheard of for Lithium batteries to lose 10% in the first 24 hours if they are completely charged. Combine the amount of power the car is drawing keeping assorted MCUs running, and temp variations and pretty soon your car probably appears to loose a fair amount of range sitting in the garage.
I'm skeptical that JIT requires a GC (in the general case, though for Java it is clearly required by the JVM). Do you have a reference for this claim?
Well, to point to evidence otherwise, if by JIT your taking about dynamic translation/recompilation for optimization purposes then no. A number of machine emulators (CPU->CPU, see apple Rosetta) perform dynamic translation, and in a number of cases pretty advanced optimization. There have also been a number of native dynamic translators for a given machine architecture see (http://www.hpl.hp.com/techreports/1999/HPL-1999-78.html) for one example. There are also strange hybrids like the Intel Profile Guided Optimizations, where the feedback from a particular run is used to statically recompile the code. This effectively removes the profile/recompile overhead from an application at the expense of the fact that its now statically optimized for a particular kind of dataset.
But, you ask why isn't a JIT part of most compiler packages if it gives you all these advantages. And the answer is three fold. First OOO CPU's kill the majority of the advantage that projects like dynamo found. Secondly, the overhead of the dynamic translation monitoring/recompile is often greater than the performance gain. And finally, the JIT advantages are often isolated to small hot pieces of code (compared with optimizing an application that is hundreds of MB of code). Applications with small performance critical functions are often hand optimized in C/C++ languages in ways that cannot be done by simple dynamic optimization and often yield functions that execute at the maximum a particular piece of hardware is capable of (say limited by memory bandwidth, functional unit throughput, etc).
Yah, I think redbox was the biggie. The business model encourages people to return them in 1 day. People are fairly happy, even if they keep the DVD for longer. Even if you return the DVD after two or three days its not like you feel raped by late fees that cost more for one day late than the previous 3 days of rental.
So, redbox probably gets more per/day/DVD, they rent pretty much the same DVDs (the new release wall at blockbuster was the majority of their rentals), their overhead is way lower, they cater to the same customers (people looking for a new release _right now_), the customers pay less, and they leave feeling happy instead of raped. Is it any surprise blockbuster isn't around anymore?
People looking for rare back catalog have netflix and the one or two local rental chains that are still around in any given town that have 20 years of back catalog rentals crammed into isles 2 feet wide.
Slashdot Mirror
B&N did the same to me a couple years ago with the touchpad fire sale. They even took my money and had to refund it a month later when they discovered they sold a million of the things but only had 5 in stock.
However, not everyone can spend about $6000 for the drive
Vs hiring a bunch of engineers to build an optical jukebox, which is what happened in the article in question...
I don't know, i'm betting you could have bought a really big tape library with a lot of tapes/drives for the price it cost facebook to build that thing using $100k+/year engineers.
T10000 tapes are the same size and store 5TB uncompressed
Now see you missed one of the cool things about the "enterprise" drives (vs LTO). Those T1k carts you purchased last year that stored 5TB, can now magically store 8.5 TB uncompressed with the latest T1kD tape drive.
Spend a few tens of thousands on new drives increases the capacity of your existing media investment by 60%, not bad if you happen to have a exabyte of capacity.
What we need are some judges that realize that just because you didn't find the critical security hole when you tested said product, doesn't mean that the company isn't liable for assuring that the product you purchased remains useful for its intended purpose.
Instead what we have are companies giving you shit, and then trying to charge you to fix their mistakes. Most other industries worked this out years ago, someone discovers your refrigerator's defrost unit catches fire and burns peoples houses down after its been in the market for a few years, the original company is generally on the hook to repair it.
A few years ago, GE sent me a nice notice telling me that I should immediately contact a repair organization to replace the door on my 10 year old dish washer, or alternatively they would give me a $200 discount off the purchase of a new unit... All because it turns out they were catching fire.
Sounds like what they really want is tape..
Besides the difficulty of dealing with 174 bluray disks instead of 1 tape... You have to wonder about the reliability of those disks sitting around on a shelf for ten years..
Oh, and you can write said tapes at 500+ MByte/sec.
Plus, tape is well understood, and there are tons of media management applications that track whats on the tape, when it expires, where its located, what encryption keys are used to decrypt it.. Basically 40 years of data management infrastructure.
However, Software wise z/OS is now so far behind it gets harder and harder to find people who want to work with it.
I agree. No one expects to have to write assembly to basically roll some logs, but that is exactly what I found myself doing a couple years ago. Hacking up the example hooks to install a piece of functionality that should be part of any sane OS.
Or for that matter spending 2 days installing/configuring z/VM and editing the configuration with text editors that would have been put to shame by the editors found on home computers of the early 80s. All this for a configuration similar to what could be archived on a PC in less than 15 minutes with few mouse clicks and vmware.
System i is already hosted on x86
Really? Didn't know that, I thought you had to have a POWER compute node in your pureflex system to run IBM i... Do you have a link indicating that it will run on the x86 compute nodes ?
Frankly, i'm skeptical. I've seen the presentations and I don't ever remember seeing anything about IBM i/i5/i5OS/as400 nodes being hosted on x86, only on POWER. Sure you can put them in the same cabinet, but that is really nothing new, just now you can manage them from the same console. With a crapload of limitations of course. It takes about 3 seconds to find all kinds of issues with the way IBM is managing the systems.. I blew some minds when I asked if I could manage vmware and powerVM instances from the same console (the answer was no at the time).
You mean z and i.... p has been struggling due to the onslaught of x86's with Linux.
Sure p/i are the same hardware now and can run linux too. But in the end its far easier to switch from linux/power or aix/power to linux/x86 than it is to switch from IBM i/zos to some other platform.
So, both of those markets are nice, but IBM will continue to try an milk them like HP does with nonstop/openvms/etc. Its only a matter of time before the margins on those start to get squeezed as people decide paying ten million dollars for something they can port to linux and run on $50k a year of hardware is stupid. Or more likely, a 3rd party starts providing similar services as the in house applications and the bean counters decide to outsource the whole shebang (seen that a couple times recently).
Frankly, both IBM i and AIX could be ported to x86 and the whole power hardware division go away in 5 or 6 years while still maintaining the existing software lock.
zOS, probably not so much unless they use one of the emulation technologies they have locked away in the vault.
I think, on the whole, it's a moot point and we could all do with less Lennart Pottering-esque conflating of server vs. desktop priorities.
I think boot times matter on server gear too, and for a while so did IBM, back when they were pushing availability numbers because a single hour long reboot, once a year can put the system availability in the three nines category. Sometimes the hardware needs actual reboots be that for firmware or hardware upgrades... None of my machines have physical hotplug memory for example... And none of my PCIe x86's have hotplug PCIe, even though a fair number of the PCI-X machines did.
When my desktop takes 5 minutes to boot, one person is inconvenienced. When my database server takes 30 minutes to boot, potentially thousands are inconvenienced.
BTW: I have a zSeries too, and we don't have good power conditioning on it, and it seems to take a power hit once or twice a year. That thing is a PITA because the "DASD" takes the better part of two hours to come back up after hard power loss.
If you hadn't noticed it seems most cars made in the last few years have custom sized radio's with buttons on the steering wheels USB/ipod ports in the glove compartments/etc. I think the need to design a space that holds a GPS screen and/or a radio was the final blow.
The old standard form factor radio is dead is most cars. Meaning the days of aftermarket stereos are as limited.
This seems to be a common practice in a lot of industry lately. Take a standard and "improve" it in a dozen different ways so that the market is fragmented and every manufacture gets their own way of doing something.
I suspect that this will be the case for most of these "upgrades".
No, crap.. Some of these UEFI servers are amazing...
Amazing that they can take 5+ minutes to get to the grub menu.
Of course the RS6k/pSeries machines I've used are worse. Stack a few IO drawers/FC cards in a machine and be prepared to wait a 30 minutes to boot to AIX.
I didn't realize dell was better. I'm going to have to give them another look. We have HP, IBM, and supermicro.
The supermicros we have boot in 1/10th the time with similar system configurations.
But still, we don't make money off the server itself, so why make them?
If this is what passes for IBM logic these days, then the sooner they die the better.
You sell/manufacture them, so that you can bundle higher margin products and provide a one stop shopping/support experience. How many Dells are running Tivoli software, or better yet how many places are running Dells and hiring IBM to manage them? I'm betting those numbers are far lower than the places with IBM hardware.
So cry me a river, Lenovo is doing just fine with the "low margin" PC's business they purchased from IBM. In fact they are kicking IBM's ass, their revenue numbers have been increasing ~13% per year for the last few years while IBM's revenue numbers have been flat for a decade.
If IBM management had any creativity, they would figure out how to in source some of that manufacturing and streamline it rather than giving foxconn/whoever a 5% cut then crying their isn't any margin left.
I think what your looking for is sandboxie.
Its sort of like solaris containers, AIX's WPARS or LXC.
Others listed here: http://en.wikipedia.org/wiki/Operating_system-level_virtualization
The problem with these solutions is that when you want to actually share data (think clipboards, or word documents) you have to poke huge holes in the security model to get them in/out of the container. That is the fundamental issue with windows, either you allow applications the ability to extend the environment (think adding thumbnail viewers for proprietary image formats), share data, or even do some level of application embedding/etc it becomes very difficult to secure that environment from a rogue application.
. Above that amount, gasoline engines start to have issues with Ethanol. Rubber seals, hoses and plastic parts in fuel systems start having reduced lifespan. Above 10% some engines start having other internal issues.
Tell that to my lawn equipment that needs to have the little rubber diaphragms used as vacuum fuel pumps, the little polymer priming bulbs, and even the rubber hoses replaced on a 6 month to 3/4 year schedule when run on the 10% mixture at the gas station. Yes, and some of these devices have been manufactured in the last couple years.
Those idiots in congress think the only gas engines that matter are the ones sold by the automobile manufactures in the last 5 years. Its a lot harder to build a small engine with metal fuel lines and hard gaskets than it is to build a car engine with fixed fuel lines, large fuel pumps and injectors.
The result is that i'm paying ~15 a gallon at the local lawn and tractor supply store to avoid ethanol. Funny thing is I put a couple gallons in my old toyota and it runs _MUCH_ better on that stuff than what is available at the gas pump...
I'm all for building an ethanol fuel economy, just don't fscking do it by putting that shit in the fuel I need for my _gasoline_ engines.
MVS was supported for 34 years, from 1966 to 2000, when it became replaced by z/OS..
Its not really the same thing, MVS went through lots of revisions in that time period, and in many ways zos is just another step in that list of revisions (aka add a big feature and rename the OS). The early versions of zos could run on non 64-bit hardware and were basically identical to mvs.. Its just that support for legacy systems were dropped along the way. Even the applications have sort of moved on and now require the newer zos features.
That said, you could take a system programmer from the 1970's and put them in front of a 3270 emulator connected to a zBC12 and besides the fact its in color they wouldn't know that the machine is in a couple racks rather than taking up 40k sq/feet in the basement.
I keep reading posts on here talking about how god damned expensive RD-RAM was in the early 2000s and I just don't seem to remember it ALWAYS being that expensive, especially in summer '02
The price of RD-RAM came in line with DDR once the P4 got a chipset that could support DDR (which was after the athlon, which was already laying down the smack). Which of course pointed at the underlying agenda... Locking up who could manufacture the technology so the profit margins could be controlled. (see socket 8 licensing fiasco, itanium, etc)
IIRC the price of rambus memory literally halved in a month or two. Plus, people started talking about the latency issues, and the power consumption. By middle of 02, IIRC the price difference was basically non existent, and everyone was looking forward to DDR2.
Code samples are usually in "good" change logs as well.
Code samples for what? The library your selling? I don't think i've ever put a "code sample" in a changelog for human consumption. That is probably because everywhere I've worked, one of the first things I make sure works is a way for the source management, bug tracking, and release management systems to reference each other. That way, the internal change log has references to the bugs and patches. Curious what the code change associated with something is? Click the link and look at the diffs.
The only games in town for server Linux is now RHEL (CentOS or SciLinux if you are cheap) or Debian.
I wonder what environment you run servers in because, from where I stand (multiple platforms/etc) really the only two choices for corporate servers running linux are RHEL and SLES. Probably 1/2 the hardware I run isn't supported using linux with anything else. There is a little support for oracle linux if you live in that world, or maybe even a little xenserver, but debian doesn't even show up on the radar in most cases. I don't think I've ever seen debian running on a production server at a publicly listed company.
For example, using a SAN with qlogic cards?
Where are the debian drivers (I see RHEL, SLES and Oracle)?
http://driverdownloads.qlogic.com/QLogicDriverDownloads_UI/DefaultNewSearch.aspx
And a lot of people still have that sick taste in their mouth over the whole SUSE+Microsoft affair.
And does the sick taste extend to the closed source nvidia driver your installing to get opengl to work properly, or maybe the firmware on your wireless card with unpublished specifications, or how about the fact that the NSA is spying on everything you type on the internet? Did you stop using the internet? Do you refuse to use android because MS is getting a license fee (rather than cross licensing the technology) from your purchase?
Patent cross licenses are a fact of life in the computer industry, yes they suck, but in the end as a _USER_ of the OS it helps _ME_ too because I don't have to worry about MS sending a cease and desist to Suse/Attachmate and forcing them to pull support for samba/whatever. Or, for that matter putting them out of business.
Frankly, the whole thing is a pretty weak excuse to ignore Suse, especially in a corporate environment where the agreement actually should be reducing uncertainty in purchasing decisions.
The latest version opensuse actually is the best linux I have ever run, and that counts for a lot having run every major distribution since when the kernel was in the .9x timefame. That also includes all the recent versions of Ubuntu/mint/etc. It falls closer to the "it just works" mantra than any previous version (of course a few things still have hickups).
No one talks about Suse because we are off talking about more exciting things. That is the problem with having a stable sensible distribution that actually works.... Its doesn't have the latest $sexy to ignite peoples fires, or the latest $sucky to piss everyone off.
Personally, I suspect a fair number of people drop suse when they thought KDE jumped the shark a few years back. Now that it turns out its Gnome that jumped the shark no one remembers the one remaining major KDE based distribution.
Finally, there is SLES which is all the goodness of opensuse combined with long term vendor support as good as what is provided by redhat.
but takes 2x as long to completely refresh the screen.
Sometimes true but not always.. most 1080i content is 1080i/60 which means 60 fields per second. Lots of 1080p content is 24 or 30 frames per second. That actually means that a 1080i/60 image is refreshing the screen faster than a 1080p/24 (most bluray movies without 2:3 pulldown). With the 2:3 pulldown (pretty common on bluray) its the same but the pulldown induces "jutter".
The usual problem with "interlaced" content on LCDs is the fact that the LCD is displaying both fields simultaneously (often due to panels with max 30fps refresh). Run it with alternating black lines or some of the motion processing (usually included on 120 or 240 hz tvs) and it can look better than the 24p content. With native 30p content, sending it at 60i is basically indistinguishable..
Basically if you want a clean movie transfer you need an 24p bluray disk/player and a TV with a 24p conversion with black frame insertion.
Yah, all these naysayers need to get their eyes checked. I have crappy eyes and looking at the demo sets at frys at a 6-8' viewing distance its pretty amazing. They have some loop where there is a crowd scene and its possible to actually see what people are wearing rather than seeing "guy with red shirt" its "guy with striped red shirt". Some of the nature scenes are amazing too.
Personally, I think the old 720vs1080 arguments made during the bluray/hddvd vs DVD upgrade were more credible than the current crop of "4k is useless" people.
To me , it sounds like it could self discharge, which contrary to the "couple%" a month folks, the self discharge rates are dependent on how charged the battery is. Its not unheard of for Lithium batteries to lose 10% in the first 24 hours if they are completely charged. Combine the amount of power the car is drawing keeping assorted MCUs running, and temp variations and pretty soon your car probably appears to loose a fair amount of range sitting in the garage.
I'm skeptical that JIT requires a GC (in the general case, though for Java it is clearly required by the JVM). Do you have a reference for this claim?
Well, to point to evidence otherwise, if by JIT your taking about dynamic translation/recompilation for optimization purposes then no. A number of machine emulators (CPU->CPU, see apple Rosetta) perform dynamic translation, and in a number of cases pretty advanced optimization. There have also been a number of native dynamic translators for a given machine architecture see (http://www.hpl.hp.com/techreports/1999/HPL-1999-78.html) for one example. There are also strange hybrids like the Intel Profile Guided Optimizations, where the feedback from a particular run is used to statically recompile the code. This effectively removes the profile/recompile overhead from an application at the expense of the fact that its now statically optimized for a particular kind of dataset.
But, you ask why isn't a JIT part of most compiler packages if it gives you all these advantages. And the answer is three fold. First OOO CPU's kill the majority of the advantage that projects like dynamo found. Secondly, the overhead of the dynamic translation monitoring/recompile is often greater than the performance gain. And finally, the JIT advantages are often isolated to small hot pieces of code (compared with optimizing an application that is hundreds of MB of code). Applications with small performance critical functions are often hand optimized in C/C++ languages in ways that cannot be done by simple dynamic optimization and often yield functions that execute at the maximum a particular piece of hardware is capable of (say limited by memory bandwidth, functional unit throughput, etc).
Yah, I think redbox was the biggie. The business model encourages people to return them in 1 day. People are fairly happy, even if they keep the DVD for longer. Even if you return the DVD after two or three days its not like you feel raped by late fees that cost more for one day late than the previous 3 days of rental.
So, redbox probably gets more per/day/DVD, they rent pretty much the same DVDs (the new release wall at blockbuster was the majority of their rentals), their overhead is way lower, they cater to the same customers (people looking for a new release _right now_), the customers pay less, and they leave feeling happy instead of raped. Is it any surprise blockbuster isn't around anymore?
People looking for rare back catalog have netflix and the one or two local rental chains that are still around in any given town that have 20 years of back catalog rentals crammed into isles 2 feet wide.