Once upon a time, Apple Computers signed an agreement with Apple Records (of Beatles fame) that allowed them to use the Apple name provided they stayed out of the music business. Then along came iTunes, and the "Computer" was dropped from the name. I just wonder how much they paid the old Apple Records company for the right to broaden their markets?
I found this a fascinating talk. I was particularly intrigued by the their "canary" rollouts to small subsets of users.
Traditionally, the companies I worked for developed and tested locked-down images, taking many months to complete the testing. Rollouts were rare, maybe once a year. Only critical updates got rolled out because of the potential for breaking systems.
Google's "canary" approach seems to be a means of radically reducing the testing time for patches and updates, allowing far more frequent updates than the "traditional" approach to desktop image management.
And I agree whole heartedly: Computers will always fail. One of the tenets of real time systems design is to allow for bad signals, unexpected inputs, and failures of the components you're talking to. Google seems to have simply expanded that philosophy to cover their entire infrastructure.
Every time you have to take your hand off the keyboard to grab the mouse or flick the trackpad, you lose time.
Every time you take your hand off the mouse or trackpad and line it up on the keyboard again, you lose time.
The idiot who came up with the idea of requiring both mouse and keyboard input for one UI metaphor was a complete and utter freakin' MORON with absolutely NO UI design experience worth noting.
"We have the technological reach... so let's do it..."
Apparently this fellow has never heard of this little thing called "priorities".
Like the health care and food issues that face the world, and the tremendous difference that a trillion dollars could make to those problems.
Or investing it in providing actual high speed access to the third world to help them educate themselves so they can crawl OUT of the cesspool of a third-world lifestyle.
Or, or, or. There is a long laundry list of things more important than a ship that serves no purpose other than "build it, and they will come."
I wonder if they're paying license fees to TiVO as well? There had been a patent kerfuffle over the ability to skip ads with their PVR years ago, which I believe Dish ended up losing. I would think this "new feature" would step on that same patent's toes.
Personally I wouldn't trust a system to auto-skip ads. It's not like ads come with a sideband signal to indicate that they're advertising, and I have never seen an image analysis system that would be smart enough to distinguish between the content of the show and advertising.
The old TiVO version was good enough. Hit the 30-second skip 5 times when the ads start, and wait through the remaining 20 seconds or so of ads so you don't miss the return of the show.
You mean senselessly beat and pound on the people who already got the message. As futile and pointless as a preacher screaming to the congregation that they need to "accept Jesus" when they're already in the church.
I must admit I'm surprised by this difference. In Canada, the Charter of Rights portion of our Constitution is considered to be legally applicable to everyone and everything (Corporations are NOT people!) The idea being, of course, that if everyone has rights, everyone has to respect those rights, not just the government.
I would have thought the 4th amendment would have covered this, so all I see is grandstanding politicians pretending they care about the people while ignoring their own Constitution.
People claim new versions of Java break their code all the time, but I've never encountered the issue except when shifting major Java releases (e.g. Running Java 6 SE code on a Java 7 VM.) I've been coding Java since 1.0.
Until I actually encounter the issue myself, I'm inclined to believe the problem is incompetent programmers abusing the API contracts or taking advantage of bugs in the APIs to "save time."
I am primarily concerned with business applications, after all. Typical business applications are not exactly exercising the bleeding edge of technology -- they're usually written by experienced people who use the same tricks and tips they have for several years, and who rely on APIs that are not the "latest and greatest features."
Maybe the problem these companies are having is hiring too many fresh-out-of-school junior programmers who picked up bad habits and design methodologies from their school cohorts rather than a problem with the Java APIs themselves.
The only case where I've found the issue to be valid is complex tool bundles like the Eclipse/Glassfish bundle, and even then, the issue seems to be Eclipse compatability in the stack, not the JVM.
P.S. There probably is an easy way to get the behaviour I'd want. But I wasn't interested in spending any more time learning how to work around the interface. To me, a good interface is natural and makes the work easier; if I have to use keyboard chords and right mouse menus and such to perform common activities, it's a bad interface for my needs.
Speaking of shovelware, are they going to default to a desktop that is actually useful for developers, or are they going to stick with the Unity or Gnome 3 "defaults" that are designed for "regular" users? The key issue I have with both Unity and Gnome 3 is the tendency to re-focus an existing window instead of launching a new one; as a developer I often need many copies of the same application running, especially terminals and editors.
Or is there some way to change the default behaviour of Unity and Gnome 3 to open new copies of applications instead of bringing the already running copies to the foreground?
There is also the little issue that regardless of whether someone lied on an application, failure to do your due diligence means that you need more than a bad resume to get rid of someone after they've worked more than a year in most jurisdictions. You need a cause for firing people after a certain amount of time.
Even if an employee hasn't reached "tenure", most companies are very wary of letting someone go who lied during the application process, even if they've proven to be incompetent. I've worked for one company that paid out a full six month contract after realizing three weeks into it that the contractor was a fraud. But even frauds can sue for wrongful dismissal, and they didn't want the media headache of a public lawsuit in order to get rid of the guy.
How much more media furor would there be if you tried to fire a CEO without a clear cut case of incompetence or fraud?
True, he lied on his resume. He should be fired for it if there is any justice in the world. But business, industry, and law are not about justice -- they're about profit and law.
I don't expect Yahoo will try to fire him until/unless they've got a long laundry list of specifically incompetent actions they can use as an excuse to get rid of the CEO. And even then, it's going to cost them, because they've got contractually-obligated "golden parachute" clauses to pay out.
As the bad resume would not be part of the actual signed employment contract or offer letters, the job contract itself cannot be considered invalid because of the bad resume. That means a flaw with the contract or performance as stipulated in the contract have to be found before the contract could be terminated without penalties.
We could spin accountability around and provide people and companies absolutely no insurance coverage nor liability when they have a break-in due to out-of-date software. You'd have to provide an exemption for those who can prove they couldn't upgrade, transferring the liability to the software providers who aren't keeping up to date.
But we'll never be able to do anything to completely prevent attacks, force people to be responsible about upgrades, nor to ensure there are never any holes in software in the first place. Software defects are a fact of life.
Absolutely true. I know a half dozen people who download cams to check out a movie before they're willing to spend $80+ taking the whole family to see it. While that means the parent who did the previewing has to sit through the movie a second time, everyone else in the family is enjoying a new flick. And the parents in question are willing to do that because they're not willing to gamble $80+ on the film of the week not being a stinker, given Hollywood's track record for producing drek.
Considering that the Y2K problem and the 32-bit time problem were known issues raised way back in the early 1980s when I was starting university, I seriously doubt any one company could claim they "discovered" the problem.
Anyone with a functioning brain cell knew there would be a problem with the end of the century due to 2-digit years. Even when they were writing the software that contained the problem; they just decided to postpone dealing with it while technology advanced and storage became cheap enough that you didn't need to do trivial data compression like cutting the size of a date field in half.
You can remotely access and program pretty much any system you'd ever work on in an offshoring relationshing. Your physical location has little or nothing to do with the ability to provide the contracted services.
While there is demand for at least some of the offshore service provider's staff to be working on-site with the customer companies, you wouldn't be able to do that with this ship. You still wouldn't have a visa, so you still wouldn't be allowed to "land" from the ship for such meetings.
In order to be in international waters, the ship would be what, 200 miles out from shore? That's a pretty long ride for any landbound customers to take in order to come meet with you on the ship. Customers don't tend to meet at provider sites; they expect the provider to come to them.
That being the case, what is the actual purpose served by working on this ship?
Or is this like the old Sealand failure? A great idea in concept that has no practical purpose and few real backers?
We're talking about new business models. New data organizations and structures that support a new way of looking at user information.
So what "open standards" are they demanding that these companies use for data export/import? Each of them probably uses dramatically different database designs, and this is not a mature industry that has developed any standards to implement, much less open/shared ones!
Well, I guess all we can do is hope the justices give the EU opinion on the matter due consideration, along with the potential impact of putting the entire industry at risk of IP "owners" retroactively changing the open nature of their publicly stated contracts. Such as the fact that Sun endorsed the Apache implementation, and that the Sun Java source was released as open source, with a specific clause that effectively treated the Java API as an LGPL interface so non-GPL software could run on top of the GPL'd code.
If the intent of the license used was to treat the interface as LGPL, then that means anyone is free to use the interface. On either side of the interface contract. There is nothing in neither the GPL nor the LGPL that indicates the interface contract is unidirectional.
I actually wonder if anyone ever considered the possibility of someone re-implementing something that was already under the GPL and LGPL...
What's to prevent a closed-source implementation of the various C/C++ interfaces and specifications that form standards specifications? Absolutely nothing that I can think of, and many examples of both sides of the interface in practical use.
How can you arbitrarily and retroactively change the bidirectional nature of an interface contract?
While I have issues running the 12.04 distribution, most people don't seem to have any problem. From what I've seen of it, the UI has been substantially cleaned up and a lot of configurability has been implemented. From a performance standpoint, I saw up to a 30% improvement in the runtimes of some key utilities I tested over the course of a weekend compared to 10.04.1, with absolutely no investment in hardware upgrades what so ever.
I didn't do enough testing before nuking the partition to determine if the improvements carried over to other key software I use (suggesting kernel-level and library implementation improvements) or application-specific.
But it's the initial LTS release -- it's not surprising third-party products don't run on it until around the.1 release anyhow. Third-party products I tried to install all had issues with installation, configuration, or startup. Nothing I actually need that wasn't baked into the distro would work except Oracle's Java 7 implementation.
Hopefully the updated version of Fedora/KDE 16 fares better when I have time to try installing the third party products. Hardware replacement is not an option at this time, and that's the last distro I'm even going to try to install with.
As happy as I've been with 10.04.1, I may have to switch back to Windows 7 just to keep this hardware useful. It's long in the tooth, but it works, so I'm loathe to just toss it once 10.04.1 support expires.
In an age of cruise missiles and smart missiles launched from conventional fighters and a demand from the public that all weapons be precisely targeted to minimize "collateral damage", what's the freaking PURPOSE of a bomber in the first place? If you want to nuke them, you'd use an ICBM. If you want precise targetting, you'd use a Tomahawk or a drone.
The F-22 kickbacks have been paid off for a weapon that will likely never see service.
The F-35's on their way to suck at the budget teat in Canada and the US both. Planes which just happen to be ill-suited to patrolling the arctic, which is the main reason Canada wanted them in the first place, effectively making them as useless as the F-22s are for the US.
How many BILLIONS are they planning to spend on bombers to attack an "enemy" that shows no signs of military buildup or aggression THIS time?
Just how long is it going to take the world to stop feeding the military-industrial pigs that design this overpriced crap? When are our governments going to realize that you reach a point where no matter how much you've spent to date, you have to CANCEL a project because it will NEVER pay for itself nor deliver what it promised?
Copying the "structure, sequence, and organization" of the Java APIs is the definition of implementing an object-oriented interface, regardless of the specifics of the implementation.
The article is also incorrect when it says Android is the "only" project/product impacted by the decision. There's this little Apache project that wrote the code Android uses, so every product or project which relies on that code is affected by this ruling. They just haven't been sued yet.
The essence of this ruling is that publishing something under open source means nothing if the copyright holder later changes their mind. And that is the biggest blow to the software industry that could have been levelled by any company for any reason, because it affects over 75% of the systems which implement the infrastructure of the internet.
When (not "if") this idea is propagated to the POSIX APIs, the C-library interfaces, the C++ standard libraries, and a host of other open source products and packages, the whole industry is fucked!
After all, the money to pay for F-22s that have never flown a combat mission and cost a year's salary to fly for an hour was FAR more important than trivial things like weather forecasting.
But just you wait and see what gets cut to pay for the F-35s!
Slashdot Mirror
Warning! If the world does not admit that Anonymous is the most powerful organization on the planet, we're going to pout.
And deface some websites.
And run some script-kiddie tools.
But most dangerous of all: we'll put out another robotic-voice video of someone in a Guy Fawkes mask!
You have been warned!
Love, Anonymous
Thanks. Wikipedia and search engines are great info stores -- if you know what terms to search for!
Once upon a time, Apple Computers signed an agreement with Apple Records (of Beatles fame) that allowed them to use the Apple name provided they stayed out of the music business. Then along came iTunes, and the "Computer" was dropped from the name. I just wonder how much they paid the old Apple Records company for the right to broaden their markets?
Yeah, it's off topic. :)
I found this a fascinating talk. I was particularly intrigued by the their "canary" rollouts to small subsets of users.
Traditionally, the companies I worked for developed and tested locked-down images, taking many months to complete the testing. Rollouts were rare, maybe once a year. Only critical updates got rolled out because of the potential for breaking systems.
Google's "canary" approach seems to be a means of radically reducing the testing time for patches and updates, allowing far more frequent updates than the "traditional" approach to desktop image management.
And I agree whole heartedly: Computers will always fail. One of the tenets of real time systems design is to allow for bad signals, unexpected inputs, and failures of the components you're talking to. Google seems to have simply expanded that philosophy to cover their entire infrastructure.
Every time you have to take your hand off the keyboard to grab the mouse or flick the trackpad, you lose time.
Every time you take your hand off the mouse or trackpad and line it up on the keyboard again, you lose time.
The idiot who came up with the idea of requiring both mouse and keyboard input for one UI metaphor was a complete and utter freakin' MORON with absolutely NO UI design experience worth noting.
"We have the technological reach... so let's do it..."
Apparently this fellow has never heard of this little thing called "priorities".
Like the health care and food issues that face the world, and the tremendous difference that a trillion dollars could make to those problems.
Or investing it in providing actual high speed access to the third world to help them educate themselves so they can crawl OUT of the cesspool of a third-world lifestyle.
Or, or, or. There is a long laundry list of things more important than a ship that serves no purpose other than "build it, and they will come."
I wonder if they're paying license fees to TiVO as well? There had been a patent kerfuffle over the ability to skip ads with their PVR years ago, which I believe Dish ended up losing. I would think this "new feature" would step on that same patent's toes.
Personally I wouldn't trust a system to auto-skip ads. It's not like ads come with a sideband signal to indicate that they're advertising, and I have never seen an image analysis system that would be smart enough to distinguish between the content of the show and advertising.
The old TiVO version was good enough. Hit the 30-second skip 5 times when the ads start, and wait through the remaining 20 seconds or so of ads so you don't miss the return of the show.
You mean senselessly beat and pound on the people who already got the message. As futile and pointless as a preacher screaming to the congregation that they need to "accept Jesus" when they're already in the church.
I must admit I'm surprised by this difference. In Canada, the Charter of Rights portion of our Constitution is considered to be legally applicable to everyone and everything (Corporations are NOT people!) The idea being, of course, that if everyone has rights, everyone has to respect those rights, not just the government.
I would have thought the 4th amendment would have covered this, so all I see is grandstanding politicians pretending they care about the people while ignoring their own Constitution.
People claim new versions of Java break their code all the time, but I've never encountered the issue except when shifting major Java releases (e.g. Running Java 6 SE code on a Java 7 VM.) I've been coding Java since 1.0.
Until I actually encounter the issue myself, I'm inclined to believe the problem is incompetent programmers abusing the API contracts or taking advantage of bugs in the APIs to "save time."
I am primarily concerned with business applications, after all. Typical business applications are not exactly exercising the bleeding edge of technology -- they're usually written by experienced people who use the same tricks and tips they have for several years, and who rely on APIs that are not the "latest and greatest features."
Maybe the problem these companies are having is hiring too many fresh-out-of-school junior programmers who picked up bad habits and design methodologies from their school cohorts rather than a problem with the Java APIs themselves.
The only case where I've found the issue to be valid is complex tool bundles like the Eclipse/Glassfish bundle, and even then, the issue seems to be Eclipse compatability in the stack, not the JVM.
P.S. There probably is an easy way to get the behaviour I'd want. But I wasn't interested in spending any more time learning how to work around the interface. To me, a good interface is natural and makes the work easier; if I have to use keyboard chords and right mouse menus and such to perform common activities, it's a bad interface for my needs.
Speaking of shovelware, are they going to default to a desktop that is actually useful for developers, or are they going to stick with the Unity or Gnome 3 "defaults" that are designed for "regular" users? The key issue I have with both Unity and Gnome 3 is the tendency to re-focus an existing window instead of launching a new one; as a developer I often need many copies of the same application running, especially terminals and editors.
Or is there some way to change the default behaviour of Unity and Gnome 3 to open new copies of applications instead of bringing the already running copies to the foreground?
There is also the little issue that regardless of whether someone lied on an application, failure to do your due diligence means that you need more than a bad resume to get rid of someone after they've worked more than a year in most jurisdictions. You need a cause for firing people after a certain amount of time.
Even if an employee hasn't reached "tenure", most companies are very wary of letting someone go who lied during the application process, even if they've proven to be incompetent. I've worked for one company that paid out a full six month contract after realizing three weeks into it that the contractor was a fraud. But even frauds can sue for wrongful dismissal, and they didn't want the media headache of a public lawsuit in order to get rid of the guy.
How much more media furor would there be if you tried to fire a CEO without a clear cut case of incompetence or fraud?
True, he lied on his resume. He should be fired for it if there is any justice in the world. But business, industry, and law are not about justice -- they're about profit and law.
I don't expect Yahoo will try to fire him until/unless they've got a long laundry list of specifically incompetent actions they can use as an excuse to get rid of the CEO. And even then, it's going to cost them, because they've got contractually-obligated "golden parachute" clauses to pay out.
As the bad resume would not be part of the actual signed employment contract or offer letters, the job contract itself cannot be considered invalid because of the bad resume. That means a flaw with the contract or performance as stipulated in the contract have to be found before the contract could be terminated without penalties.
We could spin accountability around and provide people and companies absolutely no insurance coverage nor liability when they have a break-in due to out-of-date software. You'd have to provide an exemption for those who can prove they couldn't upgrade, transferring the liability to the software providers who aren't keeping up to date.
But we'll never be able to do anything to completely prevent attacks, force people to be responsible about upgrades, nor to ensure there are never any holes in software in the first place. Software defects are a fact of life.
Absolutely true. I know a half dozen people who download cams to check out a movie before they're willing to spend $80+ taking the whole family to see it. While that means the parent who did the previewing has to sit through the movie a second time, everyone else in the family is enjoying a new flick. And the parents in question are willing to do that because they're not willing to gamble $80+ on the film of the week not being a stinker, given Hollywood's track record for producing drek.
Considering that the Y2K problem and the 32-bit time problem were known issues raised way back in the early 1980s when I was starting university, I seriously doubt any one company could claim they "discovered" the problem.
Anyone with a functioning brain cell knew there would be a problem with the end of the century due to 2-digit years. Even when they were writing the software that contained the problem; they just decided to postpone dealing with it while technology advanced and storage became cheap enough that you didn't need to do trivial data compression like cutting the size of a date field in half.
You can remotely access and program pretty much any system you'd ever work on in an offshoring relationshing. Your physical location has little or nothing to do with the ability to provide the contracted services.
While there is demand for at least some of the offshore service provider's staff to be working on-site with the customer companies, you wouldn't be able to do that with this ship. You still wouldn't have a visa, so you still wouldn't be allowed to "land" from the ship for such meetings.
In order to be in international waters, the ship would be what, 200 miles out from shore? That's a pretty long ride for any landbound customers to take in order to come meet with you on the ship. Customers don't tend to meet at provider sites; they expect the provider to come to them.
That being the case, what is the actual purpose served by working on this ship?
Or is this like the old Sealand failure? A great idea in concept that has no practical purpose and few real backers?
We're talking about new business models. New data organizations and structures that support a new way of looking at user information.
So what "open standards" are they demanding that these companies use for data export/import? Each of them probably uses dramatically different database designs, and this is not a mature industry that has developed any standards to implement, much less open/shared ones!
Well, I guess all we can do is hope the justices give the EU opinion on the matter due consideration, along with the potential impact of putting the entire industry at risk of IP "owners" retroactively changing the open nature of their publicly stated contracts. Such as the fact that Sun endorsed the Apache implementation, and that the Sun Java source was released as open source, with a specific clause that effectively treated the Java API as an LGPL interface so non-GPL software could run on top of the GPL'd code.
If the intent of the license used was to treat the interface as LGPL, then that means anyone is free to use the interface. On either side of the interface contract. There is nothing in neither the GPL nor the LGPL that indicates the interface contract is unidirectional.
I actually wonder if anyone ever considered the possibility of someone re-implementing something that was already under the GPL and LGPL...
What's to prevent a closed-source implementation of the various C/C++ interfaces and specifications that form standards specifications? Absolutely nothing that I can think of, and many examples of both sides of the interface in practical use.
How can you arbitrarily and retroactively change the bidirectional nature of an interface contract?
While I have issues running the 12.04 distribution, most people don't seem to have any problem. From what I've seen of it, the UI has been substantially cleaned up and a lot of configurability has been implemented. From a performance standpoint, I saw up to a 30% improvement in the runtimes of some key utilities I tested over the course of a weekend compared to 10.04.1, with absolutely no investment in hardware upgrades what so ever.
I didn't do enough testing before nuking the partition to determine if the improvements carried over to other key software I use (suggesting kernel-level and library implementation improvements) or application-specific.
But it's the initial LTS release -- it's not surprising third-party products don't run on it until around the .1 release anyhow. Third-party products I tried to install all had issues with installation, configuration, or startup. Nothing I actually need that wasn't baked into the distro would work except Oracle's Java 7 implementation.
Hopefully the updated version of Fedora/KDE 16 fares better when I have time to try installing the third party products. Hardware replacement is not an option at this time, and that's the last distro I'm even going to try to install with.
As happy as I've been with 10.04.1, I may have to switch back to Windows 7 just to keep this hardware useful. It's long in the tooth, but it works, so I'm loathe to just toss it once 10.04.1 support expires.
In an age of cruise missiles and smart missiles launched from conventional fighters and a demand from the public that all weapons be precisely targeted to minimize "collateral damage", what's the freaking PURPOSE of a bomber in the first place? If you want to nuke them, you'd use an ICBM. If you want precise targetting, you'd use a Tomahawk or a drone.
Just what are these idiots planning to bomb?
The F-22 kickbacks have been paid off for a weapon that will likely never see service.
The F-35's on their way to suck at the budget teat in Canada and the US both. Planes which just happen to be ill-suited to patrolling the arctic, which is the main reason Canada wanted them in the first place, effectively making them as useless as the F-22s are for the US.
How many BILLIONS are they planning to spend on bombers to attack an "enemy" that shows no signs of military buildup or aggression THIS time?
Just how long is it going to take the world to stop feeding the military-industrial pigs that design this overpriced crap? When are our governments going to realize that you reach a point where no matter how much you've spent to date, you have to CANCEL a project because it will NEVER pay for itself nor deliver what it promised?
Copying the "structure, sequence, and organization" of the Java APIs is the definition of implementing an object-oriented interface, regardless of the specifics of the implementation.
The article is also incorrect when it says Android is the "only" project/product impacted by the decision. There's this little Apache project that wrote the code Android uses, so every product or project which relies on that code is affected by this ruling. They just haven't been sued yet.
The essence of this ruling is that publishing something under open source means nothing if the copyright holder later changes their mind. And that is the biggest blow to the software industry that could have been levelled by any company for any reason, because it affects over 75% of the systems which implement the infrastructure of the internet.
When (not "if") this idea is propagated to the POSIX APIs, the C-library interfaces, the C++ standard libraries, and a host of other open source products and packages, the whole industry is fucked!
After all, the money to pay for F-22s that have never flown a combat mission and cost a year's salary to fly for an hour was FAR more important than trivial things like weather forecasting.
But just you wait and see what gets cut to pay for the F-35s!