They are simply telling you the harsh reality that if you want it to happen and they don't have the time and/or inclination, then you either have to make it happen some other way or live without.
Absolutely, and I have zero problems with being told that. It's ludicrous to expect someone else to spend hours of their free time writing code just because I want a particular feature and am not willing to pay for that time. My point was just that the ability to roll your own changes is always trotted out as a big benefit of open source, but often it is of little practical value.
How do you keep an employee from taking that training you just paid for and leaving for what the employee sees as greener pastures?
By keeping your own pastures sufficiently green, of course. Nowadays there seem to be quite a few employers that still don't understand that at-will employment is a two-edged sword. They're quite happy to cut people loose at the drop of a hat when the quarterlies take a dip, but will then turn around and whine when people leave because they've been putting in 50-60 hour weeks for six months straight and the company won't hire more people, or haven't gotten a cost-of-living adjustment in their salary for 5 years, or other similar problem that leads the workers to believe the company doesn't value them. It's not difficult to keep employees, but you do have to be willing to do it instead of displaying the attitude of "don't let the door hit you on the way out" as a large number of companies do today. Loyalty isn't an entitlement - it has to be earned.
Most people don't just change jobs on a whim, but if you come out and demand that your employees spend a few thousand dollars just to keep their current job without offering some kind of incentive to do so, don't be surprised if they walk.
If not then ‘do it yourself’ is a perfectly reasonable response.
It is if you have the time, ability, and willingness to do it yourself. Otherwise it's much more efficient to go buy an existing closed source product that actually does what you want. I'm a coder with a couple of decades of experience across a variety of platforms, so I probably *could* hack on an open-source project to get it to do what I want, but rather than waste God-knows how many hours of my time and then be told to go away when I try to submit a change to the repo, I'm going to save myself a lot of grief, plop down a credit card and buy something made by a company that actually gives a damn about what I need.
"But you could pay someone to do the work for you!" $650 buys me a full copy of Photoshop. That same $650 will buy me 2-3 days of a coder's time, which isn't going to get me very far towards bringing GIMP up to the same functionality.
Right now, OpenSolaris is the only operating system that supports ZFS in the kernel *and* is capable of being a Xen dom0, in addition to offering its own native VM capability via Zones. FreeBSD supports ZFS and recently achieved the ability to run as a domU, but you're SOL if you actually want to host VMs on it under Xen, and I don't see that changing anytime soon.
Linux already supports ZFS via FUSE, but the performance sucks and can't really get much better since it's limited to running in userspace. The GPL and CDDL (which ZFS is licensed under) are fundamentally incompatible with each other, so don't hold your breath waiting for ZFS support in the Linux kernel.
Also note the parent poster's first link is a couple of years old, in addition to your comments.
ZFS *can* run into problems if it's run on cheap hard disks that try to boost their performance numbers by returning immediately from a cache flush request instead of actually writing the data to the platters first, but that's not a problem with ZFS itself. Most of the issues that I've seen regarding ZFS have been the end result of the storage subsystem not honoring flush semantics, or the result of a RAID controller going south and taking everything attached with it. The remainder have been due to flaky drivers or just the result of using code still in development.
On my own hardware (which experiences a pretty fair amount of I/O load), it's been totally solid, and makes LVM look like a toy IMO.
Its open sourced and I would imagine it costs about 1/2 the amount it takes to develop MS Office.
This doesn't do anything to help your argument - it's a number that was pulled out of the air. We don't really know how the development efforts for MS Office and Open Office compare to each other, so you can't really say whether or not it's profitable. Besides, while OOo is a passable office productivity suite and meets my own needs just fine, there's plenty that MS Office can do that OOo can't, and it's not really accurate to say they're comparable products.
Unlike Linux, the best parts of Solaris have never come from outside Sun. Dtrace, ZFS, integrated hardware, all this stuff is where Sun's real value lay.
This is true. Let's hope Oracle works to keep the people that come up with these kinds of innovations. I'm not expecting it to happen though, and as more Sun alumni leave, it's going to be harder and harder for Oracle to continue making Solaris 10/Solaris Next a viable product, much less OpenSolaris.
It's a shame too. I rather like having OpenSolaris on my personal machines, and there's nothing else out there that will give me both ZFS and the ability to run a Xen dom0. It'd be damn hard going back to LVM after having the chance to use ZFS for a while.
Oracle is far and away more profitable than Red Hat will likely ever be, so I'd say they have a pretty good handle on the general process of turning a profit. What they're looking for is a means to turn the open source products they inherited from Sun into something profitable. If they can't, then don't be surprised if those products vanish, or if you have to write a check to get a copy in the future.
But a number of red-light camera systems don't get a good photo of the driver, and the registered owner of the vehicle is charged. Having evidence automatically collected isn't the main problem, it's the fact that sometimes the wrong person is charged, and while the systems that record video are a little better, those that simply record still photos can't offer much in the way of context regarding the alleged offense. For instance, did the accused safely enter a totally clear intersection on a red light in order to avoid being rear-ended? Did he run the red at 3 am after sitting there for several minutes at an obviously malfunctioning light?
I vote for a fine mesh of red LEDs across the entire rear of the vehicle, with a symmetrical but random cluster being lit up on each side every time the brake pedal is pressed. Failing that, I'm on board with you re: hidden lights.
I've seen too many times zipperheads blowing through the light when oncoming traffic is moving and in the intersection
Rather than have automated red-light cameras though, I'd support having more traffic cops watching intersections and writing tickets. It irks me to no end that we have law enforcement ambushing people and writing tickets for people going 10 miles over the limit on large highways where it's perfectly safe to do so, when safety would be much better served by having sworn LEOs stationed at busy intersections. Go ahead and put video cameras on the intersections as well in case the accused fights the ticket, but having an officer personally witnessing the offense and citing the driver at the scene guarantees that the proper person is charged, and largely resolves any due process issues. Given that the fines for the appropriate moving violation are a lot higher, they might even make more money as a result.
I'm not against enforcing the law re: red lights, and I'd actually like to see stronger enforcement. I'm just in strong disagreement that automated cameras are the right way to do it.
I'll freely admit that Orlando is the worst place I've *ever* seen as regards traffic, and I say that having driven through more than half of the States and in two other countries as well. Yes, even worse than Boston. It's just not as noticeable to most people because the city sucks in a variety of other ways too.:-)
Florida's traffic laws are defined in Chapter 316 of the Florida Statutes. Specifically, F.S. 316.007 says that the state is the *sole* authority regarding traffic matters that are covered under Chapter 316, except as explicity delegated otherwise (and those explicit delegations are contained in F.S. 316.008).
I live in Orlando, and I suspect the city is trying to justify their program via 316.008(1) which allows localities the power of "regulating or prohibiting stopping, standing, or parking", even though in context the Legislature clearly meant the city can post "No Stopping" signs on curbs and whatnot. I don't know for sure because I can't get the city to respond to any requests for information. Having said that, it seems clear to me that the state fully intended to regulate interactions with traffic lights completely and totally via F.S. 316.075, and thus the relevant section of the Orlando City Code that authorizes the red-light camera program (Orlando City Code, Title II, Article III)) would seem to me to be clearly outside the city's authority under state law.
Like you, I understand there's a problem with people running lights and causing accidents as a result. However, I'd much rather see the city/county do it the *right* way, and have one or two officers stationed at the problem intersections to write tickets instead of doing a sleazy end run around the law and depriving the accused of their right to a fair trial.
Whether they issue a traffic citation or do it administratively via a code infraction, it's still a blatant violation of F.S. 316.007, which states in part "no local authority shall enact or enforce any ordinance on a matter covered by this chapter unless expressly authorized.". The only reason localities implement these programs administratively is so they can attempt to avoid having the accusation heard by a real court, because they know they'd get their butts handed to them if they did.
There's also the issue of the state losing money as a result of these practices - the distribution of money from each traffic ticket is very explicitly defined in state law, but by taking this route, the local municipalities figure they can get away with not having to pay the amounts legitimately owed to the state.
No, it's not that simple. Florida law says you may not *enter* the intersection when the light is red. It's perfectly legal to enter on a yellow, and to be in the intersection on the following red.
Bah, forget the issues with the short yellow - what torques me is that here in Florida it's illegal for municipalities to legislate this kind of thing, but they do it anyway, and no one says boo.
At a cost of resources, whether it's memory, or CPU time, or whatever. Rolling your own types wouldn't be equivalent to having native support in the language, and that can be a big deal in resource-limited environments. That was my point. Of course, this also assumes that you can extend/create new types. Not all languages support this, and require that you do all kinds of ugly gymnastics in code to support the same functionality.
Having one language, so long as it's turing-complete, shouldn't give you fewer methods to solve a problem.
Sure it can. Compare two Turing-complete languages, identical in every way except that one supports unsigned integers and the other doesn't. Right off the bat, one offers more ways to do some things than the other.
Slashdot Mirror
full of commented-out 'print' statements used for debugging and generally ugly (who needs a debugger?)
:-)
Maybe the guy was forced to choose between print statements and gdb?
While that sets you apart from many of us, it also reveals Windows on your desktop.
Well, not necessarily - PuTTY runs on more than just Windows, but your point still stands that the parent poster is using a GUI *somewhere*...
They are simply telling you the harsh reality that if you want it to happen and they don't have the time and/or inclination, then you either have to make it happen some other way or live without.
Absolutely, and I have zero problems with being told that. It's ludicrous to expect someone else to spend hours of their free time writing code just because I want a particular feature and am not willing to pay for that time. My point was just that the ability to roll your own changes is always trotted out as a big benefit of open source, but often it is of little practical value.
How do you keep an employee from taking that training you just paid for and leaving for what the employee sees as greener pastures?
By keeping your own pastures sufficiently green, of course. Nowadays there seem to be quite a few employers that still don't understand that at-will employment is a two-edged sword. They're quite happy to cut people loose at the drop of a hat when the quarterlies take a dip, but will then turn around and whine when people leave because they've been putting in 50-60 hour weeks for six months straight and the company won't hire more people, or haven't gotten a cost-of-living adjustment in their salary for 5 years, or other similar problem that leads the workers to believe the company doesn't value them. It's not difficult to keep employees, but you do have to be willing to do it instead of displaying the attitude of "don't let the door hit you on the way out" as a large number of companies do today. Loyalty isn't an entitlement - it has to be earned.
Most people don't just change jobs on a whim, but if you come out and demand that your employees spend a few thousand dollars just to keep their current job without offering some kind of incentive to do so, don't be surprised if they walk.
If not then ‘do it yourself’ is a perfectly reasonable response.
It is if you have the time, ability, and willingness to do it yourself. Otherwise it's much more efficient to go buy an existing closed source product that actually does what you want. I'm a coder with a couple of decades of experience across a variety of platforms, so I probably *could* hack on an open-source project to get it to do what I want, but rather than waste God-knows how many hours of my time and then be told to go away when I try to submit a change to the repo, I'm going to save myself a lot of grief, plop down a credit card and buy something made by a company that actually gives a damn about what I need.
"But you could pay someone to do the work for you!" $650 buys me a full copy of Photoshop. That same $650 will buy me 2-3 days of a coder's time, which isn't going to get me very far towards bringing GIMP up to the same functionality.
Right now, OpenSolaris is the only operating system that supports ZFS in the kernel *and* is capable of being a Xen dom0, in addition to offering its own native VM capability via Zones. FreeBSD supports ZFS and recently achieved the ability to run as a domU, but you're SOL if you actually want to host VMs on it under Xen, and I don't see that changing anytime soon.
Linux already supports ZFS via FUSE, but the performance sucks and can't really get much better since it's limited to running in userspace. The GPL and CDDL (which ZFS is licensed under) are fundamentally incompatible with each other, so don't hold your breath waiting for ZFS support in the Linux kernel.
Also note the parent poster's first link is a couple of years old, in addition to your comments.
ZFS *can* run into problems if it's run on cheap hard disks that try to boost their performance numbers by returning immediately from a cache flush request instead of actually writing the data to the platters first, but that's not a problem with ZFS itself. Most of the issues that I've seen regarding ZFS have been the end result of the storage subsystem not honoring flush semantics, or the result of a RAID controller going south and taking everything attached with it. The remainder have been due to flaky drivers or just the result of using code still in development.
On my own hardware (which experiences a pretty fair amount of I/O load), it's been totally solid, and makes LVM look like a toy IMO.
As far as opensolaris, mysql and the rest of Sun's opensource projects go, well that's just the way the cookie crumbles.
Cue the whining from Monty....
Its open sourced and I would imagine it costs about 1/2 the amount it takes to develop MS Office.
This doesn't do anything to help your argument - it's a number that was pulled out of the air. We don't really know how the development efforts for MS Office and Open Office compare to each other, so you can't really say whether or not it's profitable. Besides, while OOo is a passable office productivity suite and meets my own needs just fine, there's plenty that MS Office can do that OOo can't, and it's not really accurate to say they're comparable products.
Unlike Linux, the best parts of Solaris have never come from outside Sun. Dtrace, ZFS, integrated hardware, all this stuff is where Sun's real value lay.
This is true. Let's hope Oracle works to keep the people that come up with these kinds of innovations. I'm not expecting it to happen though, and as more Sun alumni leave, it's going to be harder and harder for Oracle to continue making Solaris 10/Solaris Next a viable product, much less OpenSolaris.
It's a shame too. I rather like having OpenSolaris on my personal machines, and there's nothing else out there that will give me both ZFS and the ability to run a Xen dom0. It'd be damn hard going back to LVM after having the chance to use ZFS for a while.
Oracle is far and away more profitable than Red Hat will likely ever be, so I'd say they have a pretty good handle on the general process of turning a profit. What they're looking for is a means to turn the open source products they inherited from Sun into something profitable. If they can't, then don't be surprised if those products vanish, or if you have to write a check to get a copy in the future.
I'd watch "A Clockwork Orange" like that.
Yes, I have actually. :-)
But a number of red-light camera systems don't get a good photo of the driver, and the registered owner of the vehicle is charged. Having evidence automatically collected isn't the main problem, it's the fact that sometimes the wrong person is charged, and while the systems that record video are a little better, those that simply record still photos can't offer much in the way of context regarding the alleged offense. For instance, did the accused safely enter a totally clear intersection on a red light in order to avoid being rear-ended? Did he run the red at 3 am after sitting there for several minutes at an obviously malfunctioning light?
I vote for a fine mesh of red LEDs across the entire rear of the vehicle, with a symmetrical but random cluster being lit up on each side every time the brake pedal is pressed. Failing that, I'm on board with you re: hidden lights.
Airbags have been an option on Honda's GoldWing (GL1800) for four years now.
I've seen too many times zipperheads blowing through the light when oncoming traffic is moving and in the intersection
Rather than have automated red-light cameras though, I'd support having more traffic cops watching intersections and writing tickets. It irks me to no end that we have law enforcement ambushing people and writing tickets for people going 10 miles over the limit on large highways where it's perfectly safe to do so, when safety would be much better served by having sworn LEOs stationed at busy intersections. Go ahead and put video cameras on the intersections as well in case the accused fights the ticket, but having an officer personally witnessing the offense and citing the driver at the scene guarantees that the proper person is charged, and largely resolves any due process issues. Given that the fines for the appropriate moving violation are a lot higher, they might even make more money as a result.
I'm not against enforcing the law re: red lights, and I'd actually like to see stronger enforcement. I'm just in strong disagreement that automated cameras are the right way to do it.
I'll freely admit that Orlando is the worst place I've *ever* seen as regards traffic, and I say that having driven through more than half of the States and in two other countries as well. Yes, even worse than Boston. It's just not as noticeable to most people because the city sucks in a variety of other ways too. :-)
I'm not sure what state-wide regulation there is
Florida's traffic laws are defined in Chapter 316 of the Florida Statutes. Specifically, F.S. 316.007 says that the state is the *sole* authority regarding traffic matters that are covered under Chapter 316, except as explicity delegated otherwise (and those explicit delegations are contained in F.S. 316.008).
I live in Orlando, and I suspect the city is trying to justify their program via 316.008(1) which allows localities the power of "regulating or prohibiting stopping, standing, or parking", even though in context the Legislature clearly meant the city can post "No Stopping" signs on curbs and whatnot. I don't know for sure because I can't get the city to respond to any requests for information. Having said that, it seems clear to me that the state fully intended to regulate interactions with traffic lights completely and totally via F.S. 316.075, and thus the relevant section of the Orlando City Code that authorizes the red-light camera program (Orlando City Code, Title II, Article III)) would seem to me to be clearly outside the city's authority under state law.
Like you, I understand there's a problem with people running lights and causing accidents as a result. However, I'd much rather see the city/county do it the *right* way, and have one or two officers stationed at the problem intersections to write tickets instead of doing a sleazy end run around the law and depriving the accused of their right to a fair trial.
Whether they issue a traffic citation or do it administratively via a code infraction, it's still a blatant violation of F.S. 316.007, which states in part "no local authority shall enact or enforce any ordinance on a matter covered by this chapter unless expressly authorized.". The only reason localities implement these programs administratively is so they can attempt to avoid having the accusation heard by a real court, because they know they'd get their butts handed to them if they did.
There's also the issue of the state losing money as a result of these practices - the distribution of money from each traffic ticket is very explicitly defined in state law, but by taking this route, the local municipalities figure they can get away with not having to pay the amounts legitimately owed to the state.
No, it's not that simple. Florida law says you may not *enter* the intersection when the light is red. It's perfectly legal to enter on a yellow, and to be in the intersection on the following red.
//not a lawyer, not legal advice, etc.
Bah, forget the issues with the short yellow - what torques me is that here in Florida it's illegal for municipalities to legislate this kind of thing, but they do it anyway, and no one says boo.
It was a general "for instance" that I gave most people credit for being able to pick up on by themselves.
At a cost of resources, whether it's memory, or CPU time, or whatever. Rolling your own types wouldn't be equivalent to having native support in the language, and that can be a big deal in resource-limited environments. That was my point. Of course, this also assumes that you can extend/create new types. Not all languages support this, and require that you do all kinds of ugly gymnastics in code to support the same functionality.
Having one language, so long as it's turing-complete, shouldn't give you fewer methods to solve a problem.
Sure it can. Compare two Turing-complete languages, identical in every way except that one supports unsigned integers and the other doesn't. Right off the bat, one offers more ways to do some things than the other.