There are usually trades that are low risk and make money around, in whatever markets still have arbitrage potential. Some of the high frequency apps work that way, finding spots where they can capture a bit of the bid/ask spread for themselves when they see someone made a trade request over a slower network and they wedge themselves in. Unfortunately, this behavior just makes people who would normally be trying to capture bigger moves paranoid and therefore try to increase their own trading speed, and that's where things get risky--the techniques you need to trade in volume and the ones that work really fast are not the same thing.
I was an individual trader with an account just big enough to comfortably day trade. Wasn't trading every five seconds, that was just the minimum amount of smoothing my trading system applied to any decision it made. Bad data from fat-finger trade makes tends to only be a few seconds long.
For the year I was doing this, almost every minute my system wasn't trading, it was running historical backtesting using minute-level data I purchased, to tune the high-level strategies. Learned so much about database performance tuning to speed that up, run more simulations per minute, that I now do that for a living.
When I used to write my own automated trading system software, I wrote some code that ignored bad events until they had persisted for a small period of time. That was motivated by a stop loss order I had in place automatically taking me out of a position at a severe loss when a bad tick (one second) of data from a mistaken trade showed up, the chart was quite similar to today's mess. So it's easy to write something that rejects bad market data for a little bit, waiting for some confirmation before doing something rash. For what I traded, if I saw the same condition for five seconds straight, it was probably real and then I'd have the program act.
Unfortunately, the current situation market includes so many automated systems that try to make money based on high frequency trading that the normal safeguards here are rejected as "adds too much latency". It's yet another one of those situation where optimizing for the normal case, where fast trades are better, causes instability during unexpected situations.
Excellent article, and it doesn't apply to just games either. The "idea guy" who can't actually build a technical product is equally worthless to all development projects. Every good developer in every field has a stack of ideas they're sitting on, waiting only for time+money to build them, and those are always going to be better thought out than ones coming from someone doesn't understand the capabilities of the platform.
Often problems and solutions make it into the bug tracker; however, that's where the pipeline ends. Fixes almost never get checked into mainline.
And in the three years I've been tracking Ubuntu development and its related bugs, I have never seen a fix for an issue I was running into backported to a LTS version. Far as I can tell, there's little beyond security fixes actually backported. This is why all my server deployments remain on RedHat/CentOS, where the bugs I run into are aggressively backported, not just the "fixed in next release" I see even on resolved ubuntu issues.
Hint: if you read something that doesn't many any sense to you, like "Only if she's 5'3", but it might be a joke, do not mod it as "offtopic". Just leave it alone for those us who get pop culture references to evaluate. (That's a "Baby Got Back" reference for you kids who weren't listening to pop radio in 1992)
Re:I nominate this for all-time...
on
Zen Coding
·
· Score: 5, Funny
You should save your vote for when its dupe shows up in a couple of days.
Where would we be if you had to use a restricted format to read normal web pages?
There are plenty of Flash-only web sites out there to demonstrate exactly where we'd be. The best possible outcome from what Apple is doing would be that becoming a less viable plan for businesses to follow. I'm glad they're making this stand because it pushes companies toward real standards instead of Flash, but I'm not falling one minute for the idea that Apple's own products are any less proprietary than Adobe's.
I'm not involved in cutting video but I work with someone who is, and they tell me they like H.264 a lot better than Ogg Theora. So...am I part of the problem? Is the Free Software movement not up to the task of competing with proprietary software?
It's been possible for a couple of years to encode for H.264 on Linux. H.264 is unfortunately encumbered by patent issues which certainly works against its adoption in more free software programs. So in this particular case, free software's ability to compete with proprietary is strangled by the ridiculous software patent situation. This makes it yet another case where corporate and legal douchebaggery is getting in the way of forward progress for free software, and therefore the world at large.
Now, had the argument been "[iMovie|Final Cut] is an easier to use and more stable video editor than anything I can get for Linux", that's both true and not so directly the result of corporate malfeasance. It's just paid software that is better polished than what you can get for free, which does happen. When using Apple's video software, I certainly go out of my way to use open standards whenever possible for what the program outputs though, just to keep myself from being forever dependent on them. I don't trust or let myself become reliant on products from Apple and Adobe, but that's also the case for Microsoft, Google, Sony (shakes fist at now non-Linux running PS3), and most other software vendors. They're all trying to keep your in their respective walled gardens.
The manual is part of the online marketing for a product nowadays. Ditto for the numerous interviews linked to in the blog entry I suggested; there were too many for me to quote individually. I'm sorry you don't understand how on-line marketing words.
Of course I can. There's a long list of spots where Sony used to refer to this feature at listed on the PlayedStation blog. The writing at Open Platform for PLAYSTATION 3 is certainly ad copy aimed at the market these are being sold to: people who do their research on-line. I don't believe the print or TV ads mentioned the feature, but Sony's on-line ad campaign, such as material on their web site and interviews done with the press, have plenty of spots where it was highlighted.
I don't think you realize exactly what Sony did here. Back when they were fighting the war against HD-DVD, they loved these Linux sales. Every user who bought a PS3 for reasons besides playing games was listed on the headcount of active Bluray players, and therefore served their master plan to kill off their competitor though showing superior market share.
Now that said competitor is gone, they'd prefer not to sell to or support those users, and so they're just killing them off. Sony has finished with using them now and now is actively fucking them over. All of us who leaned toward buying a PS3 due the Linux feature have been intentionally played here.
Al does get permission from the original writers of the songs that he parodies, even though he doesn't have to. There's your citation. Note that part of this is a pragmatic business move; if he didn't have his paperwork in order, the original artist could sue him to collect the songwriter royalties Weird Al collects from his changed version of each song.
The main thing it's appropriate to blame the US government for is the repeal of the Glass–Steagall Act in 1999. There used to be exactly the sort of regulation needed here, so your wondering about if they "could have done more" is the wrong idea--they used to do more, the question is why they stopped. That law had been protecting consumers against their banks speculating using their money for over sixty years. Its abolishment via the same sort of "things have changed!" thinking that the.com industry was foisting on the world in '99 is a large contributor to why consumers were left even caring about these financial firms going under. The consumer banking and investment speculation sides of these businesses should never have been allowed to get coupled together in the first place, such that an investment risk management failure could ripple into the need for a a consumer-oriented bailout.
"Better" is not a concept you can apply to Linux distributions, anymore than you can apply it to (wait for it...) cars. Is a giant Ford truck better than a Prius? Well that really depends on how large the stuff you have to move in the near future is, doesn't it?
The better Linux distribution for you is the one that matches your business or personal priorities more closely. Since those are your priorities, no one else can answer that question for you.
If you want to setup a cloud deployment on, say, Amazon's EC2, it's quite easy, and then once you're up and running you can then decide if you want to buy support for that deployment.
If instead you visit RedHat's cloud page, you'll see no similar guide to getting started. As far as I can tell, this is because you need to have a RHEL license to even get access to their EC2 AMI files. As close as you can get for free is the RightScale AMI for CentOS.
A lot of technical decision are made through the path of least resistance for getting started. Right now, if EC2 is your cloud platform, and you want to deploy a simple setup that you can add support to later, Ubuntu is where you'll end up at. A someone who leans toward RedHat for servers, I've been frustrated lately that I can't do free proof of concept deployments of RHEL on EC2, and then add support to them only once the result has been signed off as working. And for always unpaid setups, people don't really trust the RightScale AMI images the way they do the official Ubuntu ones either.
Plymouth originated as a RedHat technology, so expect to see it there too. Wouldn't be surprised to find it in the next Debian too--it's where everybody else is going. The ability to "degrade" back to simple text mode is supposed to be there. I expect that months from now, part of the standard set of tricks every Linux server admin knows will be how to force Plymouth into text mode. I believe this works:
plymouth-set-default-plugin text
/usr/libexec/plymouth/plymouth-update-initrd
...presuming that you can get your server booted via single user mode or via rescue disk to execute the commands. Not sure if there's a grub-based solution here that always works; adding "nomodeset" is the first thing to try.
I already do something similar on regular hard drives, based on the fact that the logical start of most hard drives now are almost twice as fast as the end. Create a small root partition, which will be on the fast part of the disk, for things that tend to be randomly accessed all the time, and then put all the bigger files on the slow part. There is no reason you need tiered storage at the hardware or OS level for this.
"Our extensive online searches have failed"...are you kidding me? Dump the current URL into archive.org and you'll find V1.2 of the agreement, from June of 2008. Here is the old text for section 3:
Without limitation, services may include the provision of the latest update or download of new release that may include security patches, and new or revised settings and features which may prevent access to pirated games, or use of unauthorized hardware or software in connection with the PS3 system. Some services may change your current settings, cause a loss of data or content, or cause some loss of functionality.
And here's the current one:
Without limitation, services may include the provision of the latest update or download of new release that may include security patches, new technology or revised settings and features which may prevent access to unauthorized or pirated content, or use of unauthorized hardware or software in connection with the PS3 system. Some services may change your current settings, cause a loss of data or content, or cause some loss of functionality.
The main change made since the version they had years ago was expanding "pirated games" into "unauthorized or pirated content"; everything has been minor wording adjustment.
Why do newbie users even need to care about that? If you pick a distribution that has a good set of packages, they should rarely have to leave the ones provided with it. Run whatever front-end for package management you've got, make sure all the optional repositories are enabled, and there should be so many packages there the hard part is sorting through them all--not finding even more. Particularly given that so many things that used to be run as local apps have moved onto web applications nowadays, the main headaches for Linux newbies I see is getting their hardware working and making Flash work.
The Citrix stuff had little to do with it. Th Linux kernel developers favor code that is easy for them to integrate and maintain, and KVM fit better into that model than Xen. There are some situations where it performs quite a bit better too, and frankly few people care about those stuck with processors that don't have the right extensions to use KVM. Some good reading on the background here includes Discover the Linux Kernel Virtual Machine, Linux: KVM Paravirtualization, and The truth about KVM and Xen.
It's not even quite that simple unfortunately. I highlighted the kernel example because FC12 is based on 2.6.31, RHEL6 on 2.6.32, and FC13 on 2.6.33. So in that particular case, they're picking a version that doesn't match any Fedora release.
I wasn't the one who suggested VirtualBox as a Xen replacement. If your position is that "VirtualBox = desktop", that's just further evidence that it's probably not appropriate for the FP here to adopt, which is in line with my suggestion to tread carefully in that direction.
While primarily targeting the desktop, VirtualBox was becoming increasingly useful as a server virtualization solution. My main point was that such improvements are less to continue now, because Oracle already has a Xen based solution for servers they're selling.
The packages mostly match those in Fedora 12, which makes sense as that came out in November and FC13 isn't released yet. However, they have bumped some things. Most notably, the FC12 kernel was 2.6.31, while RHEL6 uses 2.6.32. That's not surprising given a fair number of virtualization and performance features, as well as bug fixes, happened for 2.6.32.
While I've been a fan of VirtualBox for a while too, with the Oracle acquisition I wonder if adopting it now isn't just asking to take a ride onto another abandoned VM platform. Oracle already has Oracle VM, which is Xen based. At this point it looks like Oracle is going to turn VirtualBox into a gateway product used to hook people used to upsell onto Oracle VM. I'm not sure what that bodes for the future of VirtualBox development. I'm guessing that Oracle shifting development focus toward Oracle product compatibility concerns, so that it's easier to move paying customer to their more serious product, isn't a good sign for people who have been expecting VirtualBox to move further toward being more suitable for larger scale business deployments.
Slashdot Mirror
There are usually trades that are low risk and make money around, in whatever markets still have arbitrage potential. Some of the high frequency apps work that way, finding spots where they can capture a bit of the bid/ask spread for themselves when they see someone made a trade request over a slower network and they wedge themselves in. Unfortunately, this behavior just makes people who would normally be trying to capture bigger moves paranoid and therefore try to increase their own trading speed, and that's where things get risky--the techniques you need to trade in volume and the ones that work really fast are not the same thing.
I was an individual trader with an account just big enough to comfortably day trade. Wasn't trading every five seconds, that was just the minimum amount of smoothing my trading system applied to any decision it made. Bad data from fat-finger trade makes tends to only be a few seconds long.
For the year I was doing this, almost every minute my system wasn't trading, it was running historical backtesting using minute-level data I purchased, to tune the high-level strategies. Learned so much about database performance tuning to speed that up, run more simulations per minute, that I now do that for a living.
When I used to write my own automated trading system software, I wrote some code that ignored bad events until they had persisted for a small period of time. That was motivated by a stop loss order I had in place automatically taking me out of a position at a severe loss when a bad tick (one second) of data from a mistaken trade showed up, the chart was quite similar to today's mess. So it's easy to write something that rejects bad market data for a little bit, waiting for some confirmation before doing something rash. For what I traded, if I saw the same condition for five seconds straight, it was probably real and then I'd have the program act.
Unfortunately, the current situation market includes so many automated systems that try to make money based on high frequency trading that the normal safeguards here are rejected as "adds too much latency". It's yet another one of those situation where optimizing for the normal case, where fast trades are better, causes instability during unexpected situations.
That's because the IBM version included that fancy control program that made everything easy.
Excellent article, and it doesn't apply to just games either. The "idea guy" who can't actually build a technical product is equally worthless to all development projects. Every good developer in every field has a stack of ideas they're sitting on, waiting only for time+money to build them, and those are always going to be better thought out than ones coming from someone doesn't understand the capabilities of the platform.
Often problems and solutions make it into the bug tracker; however, that's where the pipeline ends. Fixes almost never get checked into mainline.
And in the three years I've been tracking Ubuntu development and its related bugs, I have never seen a fix for an issue I was running into backported to a LTS version. Far as I can tell, there's little beyond security fixes actually backported. This is why all my server deployments remain on RedHat/CentOS, where the bugs I run into are aggressively backported, not just the "fixed in next release" I see even on resolved ubuntu issues.
Hint: if you read something that doesn't many any sense to you, like "Only if she's 5'3", but it might be a joke, do not mod it as "offtopic". Just leave it alone for those us who get pop culture references to evaluate. (That's a "Baby Got Back" reference for you kids who weren't listening to pop radio in 1992)
You should save your vote for when its dupe shows up in a couple of days.
Where would we be if you had to use a restricted format to read normal web pages?
There are plenty of Flash-only web sites out there to demonstrate exactly where we'd be. The best possible outcome from what Apple is doing would be that becoming a less viable plan for businesses to follow. I'm glad they're making this stand because it pushes companies toward real standards instead of Flash, but I'm not falling one minute for the idea that Apple's own products are any less proprietary than Adobe's.
I'm not involved in cutting video but I work with someone who is, and they tell me they like H.264 a lot better than Ogg Theora. So...am I part of the problem? Is the Free Software movement not up to the task of competing with proprietary software?
It's been possible for a couple of years to encode for H.264 on Linux. H.264 is unfortunately encumbered by patent issues which certainly works against its adoption in more free software programs. So in this particular case, free software's ability to compete with proprietary is strangled by the ridiculous software patent situation. This makes it yet another case where corporate and legal douchebaggery is getting in the way of forward progress for free software, and therefore the world at large.
Now, had the argument been "[iMovie|Final Cut] is an easier to use and more stable video editor than anything I can get for Linux", that's both true and not so directly the result of corporate malfeasance. It's just paid software that is better polished than what you can get for free, which does happen. When using Apple's video software, I certainly go out of my way to use open standards whenever possible for what the program outputs though, just to keep myself from being forever dependent on them. I don't trust or let myself become reliant on products from Apple and Adobe, but that's also the case for Microsoft, Google, Sony (shakes fist at now non-Linux running PS3), and most other software vendors. They're all trying to keep your in their respective walled gardens.
The manual is part of the online marketing for a product nowadays. Ditto for the numerous interviews linked to in the blog entry I suggested; there were too many for me to quote individually. I'm sorry you don't understand how on-line marketing words.
Of course I can. There's a long list of spots where Sony used to refer to this feature at listed on the PlayedStation blog. The writing at Open Platform for PLAYSTATION 3 is certainly ad copy aimed at the market these are being sold to: people who do their research on-line. I don't believe the print or TV ads mentioned the feature, but Sony's on-line ad campaign, such as material on their web site and interviews done with the press, have plenty of spots where it was highlighted.
I don't think you realize exactly what Sony did here. Back when they were fighting the war against HD-DVD, they loved these Linux sales. Every user who bought a PS3 for reasons besides playing games was listed on the headcount of active Bluray players, and therefore served their master plan to kill off their competitor though showing superior market share.
Now that said competitor is gone, they'd prefer not to sell to or support those users, and so they're just killing them off. Sony has finished with using them now and now is actively fucking them over. All of us who leaned toward buying a PS3 due the Linux feature have been intentionally played here.
Al does get permission from the original writers of the songs that he parodies, even though he doesn't have to. There's your citation. Note that part of this is a pragmatic business move; if he didn't have his paperwork in order, the original artist could sue him to collect the songwriter royalties Weird Al collects from his changed version of each song.
The main thing it's appropriate to blame the US government for is the repeal of the Glass–Steagall Act in 1999. There used to be exactly the sort of regulation needed here, so your wondering about if they "could have done more" is the wrong idea--they used to do more, the question is why they stopped. That law had been protecting consumers against their banks speculating using their money for over sixty years. Its abolishment via the same sort of "things have changed!" thinking that the .com industry was foisting on the world in '99 is a large contributor to why consumers were left even caring about these financial firms going under. The consumer banking and investment speculation sides of these businesses should never have been allowed to get coupled together in the first place, such that an investment risk management failure could ripple into the need for a a consumer-oriented bailout.
What makes one Linux better than another?
"Better" is not a concept you can apply to Linux distributions, anymore than you can apply it to (wait for it...) cars. Is a giant Ford truck better than a Prius? Well that really depends on how large the stuff you have to move in the near future is, doesn't it?
The better Linux distribution for you is the one that matches your business or personal priorities more closely. Since those are your priorities, no one else can answer that question for you.
If you want to setup a cloud deployment on, say, Amazon's EC2, it's quite easy, and then once you're up and running you can then decide if you want to buy support for that deployment.
If instead you visit RedHat's cloud page, you'll see no similar guide to getting started. As far as I can tell, this is because you need to have a RHEL license to even get access to their EC2 AMI files. As close as you can get for free is the RightScale AMI for CentOS.
A lot of technical decision are made through the path of least resistance for getting started. Right now, if EC2 is your cloud platform, and you want to deploy a simple setup that you can add support to later, Ubuntu is where you'll end up at. A someone who leans toward RedHat for servers, I've been frustrated lately that I can't do free proof of concept deployments of RHEL on EC2, and then add support to them only once the result has been signed off as working. And for always unpaid setups, people don't really trust the RightScale AMI images the way they do the official Ubuntu ones either.
Plymouth originated as a RedHat technology, so expect to see it there too. Wouldn't be surprised to find it in the next Debian too--it's where everybody else is going. The ability to "degrade" back to simple text mode is supposed to be there. I expect that months from now, part of the standard set of tricks every Linux server admin knows will be how to force Plymouth into text mode. I believe this works:
plymouth-set-default-plugin text
I already do something similar on regular hard drives, based on the fact that the logical start of most hard drives now are almost twice as fast as the end. Create a small root partition, which will be on the fast part of the disk, for things that tend to be randomly accessed all the time, and then put all the bigger files on the slow part. There is no reason you need tiered storage at the hardware or OS level for this.
"Our extensive online searches have failed"...are you kidding me? Dump the current URL into archive.org and you'll find V1.2 of the agreement, from June of 2008. Here is the old text for section 3:
And here's the current one:
The main change made since the version they had years ago was expanding "pirated games" into "unauthorized or pirated content"; everything has been minor wording adjustment.
Why do newbie users even need to care about that? If you pick a distribution that has a good set of packages, they should rarely have to leave the ones provided with it. Run whatever front-end for package management you've got, make sure all the optional repositories are enabled, and there should be so many packages there the hard part is sorting through them all--not finding even more. Particularly given that so many things that used to be run as local apps have moved onto web applications nowadays, the main headaches for Linux newbies I see is getting their hardware working and making Flash work.
The Citrix stuff had little to do with it. Th Linux kernel developers favor code that is easy for them to integrate and maintain, and KVM fit better into that model than Xen. There are some situations where it performs quite a bit better too, and frankly few people care about those stuck with processors that don't have the right extensions to use KVM. Some good reading on the background here includes Discover the Linux Kernel Virtual Machine, Linux: KVM Paravirtualization, and The truth about KVM and Xen.
It's not even quite that simple unfortunately. I highlighted the kernel example because FC12 is based on 2.6.31, RHEL6 on 2.6.32, and FC13 on 2.6.33. So in that particular case, they're picking a version that doesn't match any Fedora release.
I wasn't the one who suggested VirtualBox as a Xen replacement. If your position is that "VirtualBox = desktop", that's just further evidence that it's probably not appropriate for the FP here to adopt, which is in line with my suggestion to tread carefully in that direction.
While primarily targeting the desktop, VirtualBox was becoming increasingly useful as a server virtualization solution. My main point was that such improvements are less to continue now, because Oracle already has a Xen based solution for servers they're selling.
The packages mostly match those in Fedora 12, which makes sense as that came out in November and FC13 isn't released yet. However, they have bumped some things. Most notably, the FC12 kernel was 2.6.31, while RHEL6 uses 2.6.32. That's not surprising given a fair number of virtualization and performance features, as well as bug fixes, happened for 2.6.32.
While I've been a fan of VirtualBox for a while too, with the Oracle acquisition I wonder if adopting it now isn't just asking to take a ride onto another abandoned VM platform. Oracle already has Oracle VM, which is Xen based. At this point it looks like Oracle is going to turn VirtualBox into a gateway product used to hook people used to upsell onto Oracle VM. I'm not sure what that bodes for the future of VirtualBox development. I'm guessing that Oracle shifting development focus toward Oracle product compatibility concerns, so that it's easier to move paying customer to their more serious product, isn't a good sign for people who have been expecting VirtualBox to move further toward being more suitable for larger scale business deployments.