I did say that some crappy patents slip through. Still, the fact that you had to go back two months is pretty telling. If IBM patented a stupidly-obvious idea every week, that'd still represent barely 1% of IBM's patent volume. I'd bet that's a better non-obviousness vetting rate than the patent office as a whole achieves. Of course "better than the patent office" isn't saying all that much.
One other thing I think IBM does a good job of is assuring that not only does it obtain a lot of patents, but those patents are generally of high quality, in the sense that they're real innovations that have actual value. IBM provides some significant incentives to employees to encourage patent submission, but those submissions are then vetted by a fairly skeptical evaluation committee before they're turned over to legal. Some crap patents leak through, but they're a tiny minority.
I think one of the best evidences of IBM's success in creating good patents is that IBM earns substantial revenues annually for licensing its patents, and does it without trollish behavior like submarine patents, lots of patent lawsuits, etc. IBM's patent licensees are typically happy to pay the license fees because the patents offer real value, and aren't things that the licensee would likely have independently invented. The result is over a billion dollars annually in patent licensing, which helps to offset a portion of the R&D budget.
Disclosure: I work for IBM, but not in PR and not in research. I also know of plenty of really dumb stuff the company does, but I think this is one area that IBM handles very well.
In the end all I want is something that places the programs somewhere on my screen.
Which programs? Where? On which screen? How do you move them? How do you find and launch the programs you want? How well do all of your programs integrate? How do you find specific files?
What you say is true, but misses the point. There is a huge amount that can be done to make your workflow more efficient than an environment which just requires you to manage everything yourself. I view KDE as something of an ongoing research project in this space, which is also fairly usable. There are some really cool and useful ideas in KDE right now... things like activities which, when fully completed, will allow you to define a set of applications and tools that you use together in particular ways. When you activate an activity, all of the relevant components are started and placed on-screen in the way that you want.
A simpler feature that KDE has long provided -- and which GNOME still doesn't and I don't believe Enlightenment, WindowMaker or XFCE provide -- is the ability to define per-application window settings that affect placement, sizing, etc., so that those apps always act in the defined ways. I use this to make my multiple desktops more efficient. Each of my virtual desktops holds a particular type of application, and each application is assigned to always come up on the appropriate desktop. So I never have to try to figure out which desktop a given app is on.
Comprehensive desktop search to make finding files easy, a good, efficient way to launch programs, seamless integration between applications, both local and on-line -- these are all things that a more sophisticated DE can provide. Oh, and yeah there's also eye candy, some of which has utility, and some of which is just pretty, and I do think aesthetic value is real value as well.
And for those who don't want to be pussies but don't know what action to take, I'll tell you: Complain to your representatives. Bombard your senators and congressmen with letters and phone calls. Make them calm and well-reasoned, but forceful, and make clear that if you don't see them taking action to roll back the useless security theater, you'll not only be voting for the opposing candidate in the next election, you'll be sending them checks.
Replacing hourly wage untrained rent-a-cops at security checkpoints, with employed professionals with actual management, was a good idea. Before 9/11 airport security was designed to be cheap and not impede the paying passengers. Having training, standards, etc. was a real step up.
So, you're saying that expensive ineffective security is better than cheap ineffective security?
This is actually one of the reasons those superzoom cameras work as well as they do.
Interesting. I was actually speaking of superzoom lenses for SLRs, not cameras with integral superzooms. In the latter case I can see that they could just correct it in software. In the former case, the camera you put the lens on may well have been made before the lens was designed, so they can't rely on software.
But the quality of the optics is no better than it used to be.
Actually, that's not true. The field of optical engineering is seeing very rapid advancement, in part due to new lens coatings and in part due to the ability to model complex series of lenses in computers, rather than having to build and test them experimentally.
The result is that today's top-quality lenses are sharper and have less distortion (chromatic abberation, barrel and pincushion distortion, etc.) than older lenses. And the new technology is making some things possible that simply weren't before, like the new crop of superzooms that actually have reasonable performance across very wide zoom ranges.
For many kinds of shooting, image stabilization is huge, too.
However, none of that enables a tiny P&S lens to match a lens that is four or five times its size. When it comes to optics, size matters. A lot. Bigger glass can gather more light and focus and manage it better across the visible spectrum.
So while your statement isn't true -- optics ARE better, much better, and improving rapidly -- that still doesn't support the GPs claim that a modern P&S is superior to your wife's Pentax in terms of achievable image quality.
I've seen this argument before, and frankly, the cameras we have today (even the point and shoots) are MILES above the top of the line film cameras from even a couple decades ago. Yet, somehow, the photographers from that era (and earlier) managed to get these shots that people think are impossible on lower end models of today.
Nonsense.
I would much rather have a 70s-era SLR with a full set of Zeiss glass and a variety of films than a modern point & shoot camera. Not to mention a Hasselblad...
Beyond your valid point about light sensitivity, there are a lot of other issues, most of which have more to do with optics than with "features". And there's simply no way that, for example, a teeny P&S camera can compete with a 35mm SLR with a set of good lenses. And no way that 35mm can compete with a medium or large format camera.
And that's just looking at issues of sharpness and resolution. Other crucial aspects like the ability to control aperture to manage DOF, and to trade it off against shutter speed and noise/grain can also make the difference between a shot that's worth keeping and one that isn't. Old, purely-manual film cameras provided an ability to do those things that modern P&S cameras often do not, as do DSLRs.
Of course, the biggest difference between a P&S and an SLR (or DSLR) is the obvious one -- interchangeable lenses. And there are many shots you can get with, say, a 10 mm fisheye or a 600 mm telephoto that you simply cannot get with a P&S. And that's not even getting into the more oddball stuff like tilt/shift lenses.
However, a good photographer can, will, and should be able to get the shot regardless of the gear they are using. That's why they are called professionals after all.
Yeah, right. So give a sports photographer a little P&S camera and tell him to go photograph magazine cover shots of fast-moving action with tight DOF and nice bokeh in an indoor arena. He'll just tell you that you can't get those shots with that camera. Now, he might get some nice photos of the game, but they'll be more wide-angle, more static, with greater DOF, etc. Forced to shoot with that equipment, he'll modify the sorts of shots he takes to fit the capability of the equipment.
I have a bunch of digital cameras, ranging from the one on my cellphone, to various sizes of P&S cameras, to a couple of Canon DSLRs (a 350D and a 50D). I use all of them, and (I think) take nice pictures with all of them. I know my cellphone camera doesn't really have the resolution it's rated at, and it adds an odd sort of distortion that gives the images a surreal effect -- so I use it when I like, or at least don't mind, a photo with those limitations.
I carry my small P&S cameras when it's not convenient or feasible to carry a DSLR, but I recognize that low-light shots will not be good without a tripod, that if I want a pleasantly-blurred background I'm going to have to fake it in post, that I don't have filters, that I can't carefully manage the exposure for high dynamic range shots, etc.
Likewise, when choosing between the 350D and 50D for a session, I trade off size against resolution, ISO and frame rate. Sometimes I pick the 350D over the 50D even when size doesn't matter, because at low ISOs the 350D images are cleaner.
But I guess the only reason I can't capture the same images in the same circumstances with my cellphone as I can with my Canon 50D and L-series lenses is because I'm not a professional, right?
Its not the camera that takes great photos, its the photographer. Ive seen great pics taken with a crappy disposable film camera. Ive seen shitty photos taken with a DSLR.
This is why top photographers prefer disposable film cameras over DSLRs.
Oh, wait...
A good photographer can take good pictures with any camera -- but only because he factors the capabilities of the camera into the decision of which shots to take. Many images which could be captured with the flexibility provided by a high-end DSLR with the right lens cannot be captured effectively with a cheap point & shoot. Good equipment provides options. A poor photographer won't know how to use those options, but that doesn't mean a good photographer doesn't need them.
Very cool idea. Automatically clustering the complaints by location and keyword would make it simple for the city employees to handle a large volume of complaints, too -- which would be important because if it were easy to make such complaints, the volume would rise dramatically. But there should be plenty of information to collapse 100 complaints about the same pothole into one entry in a ticketing system. The number of complaints in a cluster is also a clue to its priority.
Even in certain banks, Linux is now being used or researched
Certain banks? Based on my experience consulting in the financial industry, you'd have a hard time finding a bank that doesn't have Linux deployed somewhere.
"like the Israelis do, well, that works very well... but makes the security screening vastly more manpower-intensive and time-consuming. "...Except for that it takes the same number of people and you get through an Israeli airport in 25 minutes.
Not my experience. In my experience, it takes two screeners per passenger for 15 to 20 minutes, and you had better arrive at the airport at LEAST three hours before your flight. Four is better. Three might be cutting it close.
There's also this thing called the fourth amendment. Something about having the right not to be searched without probable cause.
Unfortunately, the courts long ago decided that the fourth amendment is irrelevant, because by choosing to fly and to enter the security checkpoint you chose to waive your fourth amendment protection. If you want to attack this issue on the basis of rights, you have to use the court-supported but less well-defined Right to Travel, and argue that in many cases air travel is the only reasonable option.
If you remove the security, do you really think terrorists wouldn't notice and exploit it again? Read up on history, once hijackings were extremely common UNTIL rigorous security measures put an end to them.
No, hijackings were extremely common until United Airlines Flight 93 put an end to them.
If we would just profile we wouldn't need half the security we have.
What do you mean by "profile"?
If you mean "apply extra scrutiny to certain ethnic and religious groups", that's completely and utterly useless.
If you mean "put all of the passengers under intense stress and watch their reactions", like the Israelis do, well, that works very well... but makes the security screening vastly more manpower-intensive and time-consuming. And, frankly, much more unpleasant than being briefly groped. I've flown out of Ben-Gurion airport a few times and I'd rather have a prostate exam.
The truth is that we simply don't need half the security we have. We should just roll it all back to pre-9/11 levels, keeping only the cockpit door locks. That plus the passengers' understanding that allowing their plane to be hijacked is likely to get them killed will mean that terrorism on airplanes will be restricted to killing passengers, making planes a low-value target. It's possible that the occasional Bad Thing will happen on an airplane, but it'll still be safer than driving.
Although one thing I've noticed is a serious reliance on "specialty parts". Now that spaceship comes in a kit with a special cockpit, wing, and landing gear piece..
The mindstorm stuff looks really cool though. I _really_ would have had a blast with something like that as a kid.
Lego is shifting away from those specialty parts, partly because of the limited building possibilities and partly because their exploding unique part count was increasing their costs. It's cheaper to produce fewer unique parts.
I'm happy to say that my kids have pretty much lost interest in the pre-designed kits. We recently stumbled over a big plastic bin full of thousands of assorted generic lego parts at a garage sale and they immediately pooled their money to buy it.
I'm telling you, Linus would accept a patch for in-kernel mixing tomorrow because it makes sense.
He would? After consistently rejecting many such patches for years, because audio mixing isn't something a kernel should do? When did he change his mind? And why?
SVN is valid for many groups, teams, around the world and will continue to be so for the foreseeable future.
SVN is particularly excellent when you use GIT as the front-end. You get all of the super-easy local branching and merging, history rewriting and lighting speed of GIT for everyday use, and when you're ready to push your work to the rest of the team, you publish it to SVN.
BTW, yes, SVN branching and merging has gotten much better since merge tracking was added, but it's still too slow and clunky for effective use with the "branch per feature" pattern of development. It can work, but GIT runs circles around it.
Now, I admit that I haven't extensively used most of the tools in the "article", but I haven't needed what they offer above and beyond CVS.
SCM is one of those areas of technology where you can easily believe that something is unnecessary until you use it -- at which point you discover that it's impossible to live without it, in spite of the fact that you lived without it just fine for decades.
So what am I supposed to do? Start every one of my X applications from a different Xterm so they have a different controlling TTY (and do not close any of them)?
Umm, no, that would effectively place every program in a different group, which is the same as not having any grouping.
The point of this patch is to group related processes together and schedule them as a unit, rather than individually, effectively putting the whole set on an even footing with other individual processes with regard to scheduling fairness. That way a single very aggressive group can't starve the rest of the processes on the system.
You're right that to some degree this is a nerd feature, since it won't address situations of low responsiveness that "normal" desktop users might suffer -- but those situations aren't caused by a group of aggressively CPU-hungry processes. Generally they're I/O-related. But for people who do the sorts of things this patch addresses, it's a valuable improvement, and it's unlikely to have any negative effects for those who don't.
It is best to provider an interface through which a user can group all of the processes with his UID and leave the Desktop environment do the grouping. Or put something on the dbus which listens and follows who talks to whom to do the same. This will provide much better results than putting yet another simplistic heuristic in the kernel.
Maybe. It's not clear to me how grouping will help in general for desktop usage. Desktop usage tends to be a bunch of single-process programs which all need to be scheduled individually, and fairly, with preference given to interactive processes. I think perhaps something like Con Kolivas' BFS might be a better approach for the desktop, and BFS doesn't have the concept of groups at all. I wonder if BFS couldn't be improved by finding a way to identify and boost interactive tasks, rather than rigid fairness, but perhaps CK is right about focusing on rigid fairness in the scheduler, and we should then have the desktop environment automatically manage nice levels, moving the question of scheduling entirely out of the kernel.
However, I don't see how the DE could usefully make use of grouping. Perhaps as our desktop applications move more to multi-threaded implementations to take advantage of the exploding number of processor cores we will want the DE to automatically group the processes within a single program -- effectively extending the TTY-based grouping notion to pure GUI apps. But I really don't think enough applications are multi-threaded enough yet that we'll see significant benefit from that in the short term.
I think that they've hit on an excellent way of making imperfect freeform linguistics very useful. Basically, it's provided as an alternative to a more precise input language, one that works a lot of the time and when it does requires less effort by the user. When the computer can work out what you mean, it does. But it also always translates what it thought you meant into the precise, formal language and displays that so that you can tell whether it got it right. Further, if there are a few competing interpretations of the vague, freeform input, the tool makes it easy for you to pick the one that you meant.
As long as it gets what you mean right enough of the time, while allowing you to be vague enough that using the freeform input really is saving you mental effort, I can see tremendous value in this approach. And over time it will continue to get better.
I have no doubt that you'll quickly learn that there are some kinds of things which are just better-expressed in formal notation. Heck, we sometimes find that to be the case even when we're communicating with other humans. But there are also a LOT of cases where informal language works just as well and is a lot easier. I'm sure that set will be smaller when working with Mathematica than another person -- but it looks to me based on the article like the set is large enough to be useful.
In particular, I can definitely see using this as an easier route to a lot of the more obscure commands and options, such as the options to tweak graphs in various ways. And if it turns out that using the natural language didn't work, well, your next step was going to be to dig through the manual to figure out the formal notation anyway, so it didn't cost you much time.
I can see this freeform linguistics approach being quite valuable, even though it's quite imperfect.
Slashdot Mirror
I did say that some crappy patents slip through. Still, the fact that you had to go back two months is pretty telling. If IBM patented a stupidly-obvious idea every week, that'd still represent barely 1% of IBM's patent volume. I'd bet that's a better non-obviousness vetting rate than the patent office as a whole achieves. Of course "better than the patent office" isn't saying all that much.
One other thing I think IBM does a good job of is assuring that not only does it obtain a lot of patents, but those patents are generally of high quality, in the sense that they're real innovations that have actual value. IBM provides some significant incentives to employees to encourage patent submission, but those submissions are then vetted by a fairly skeptical evaluation committee before they're turned over to legal. Some crap patents leak through, but they're a tiny minority.
I think one of the best evidences of IBM's success in creating good patents is that IBM earns substantial revenues annually for licensing its patents, and does it without trollish behavior like submarine patents, lots of patent lawsuits, etc. IBM's patent licensees are typically happy to pay the license fees because the patents offer real value, and aren't things that the licensee would likely have independently invented. The result is over a billion dollars annually in patent licensing, which helps to offset a portion of the R&D budget.
Disclosure: I work for IBM, but not in PR and not in research. I also know of plenty of really dumb stuff the company does, but I think this is one area that IBM handles very well.
So all you have to do is swap that blacklist for a whitelist. Problem solved.
Unless the installer automatically adds the plugin to the whitelist.
You also need to find a way to secure the whitelist against alteration by anything other than the browser.
In the end all I want is something that places the programs somewhere on my screen.
Which programs? Where? On which screen? How do you move them? How do you find and launch the programs you want? How well do all of your programs integrate? How do you find specific files?
What you say is true, but misses the point. There is a huge amount that can be done to make your workflow more efficient than an environment which just requires you to manage everything yourself. I view KDE as something of an ongoing research project in this space, which is also fairly usable. There are some really cool and useful ideas in KDE right now... things like activities which, when fully completed, will allow you to define a set of applications and tools that you use together in particular ways. When you activate an activity, all of the relevant components are started and placed on-screen in the way that you want.
A simpler feature that KDE has long provided -- and which GNOME still doesn't and I don't believe Enlightenment, WindowMaker or XFCE provide -- is the ability to define per-application window settings that affect placement, sizing, etc., so that those apps always act in the defined ways. I use this to make my multiple desktops more efficient. Each of my virtual desktops holds a particular type of application, and each application is assigned to always come up on the appropriate desktop. So I never have to try to figure out which desktop a given app is on.
Comprehensive desktop search to make finding files easy, a good, efficient way to launch programs, seamless integration between applications, both local and on-line -- these are all things that a more sophisticated DE can provide. Oh, and yeah there's also eye candy, some of which has utility, and some of which is just pretty, and I do think aesthetic value is real value as well.
You're all pussies unless you take action.
And for those who don't want to be pussies but don't know what action to take, I'll tell you: Complain to your representatives. Bombard your senators and congressmen with letters and phone calls. Make them calm and well-reasoned, but forceful, and make clear that if you don't see them taking action to roll back the useless security theater, you'll not only be voting for the opposing candidate in the next election, you'll be sending them checks.
Replacing hourly wage untrained rent-a-cops at security checkpoints, with employed professionals with actual management, was a good idea. Before 9/11 airport security was designed to be cheap and not impede the paying passengers. Having training, standards, etc. was a real step up.
So, you're saying that expensive ineffective security is better than cheap ineffective security?
This is actually one of the reasons those superzoom cameras work as well as they do.
Interesting. I was actually speaking of superzoom lenses for SLRs, not cameras with integral superzooms. In the latter case I can see that they could just correct it in software. In the former case, the camera you put the lens on may well have been made before the lens was designed, so they can't rely on software.
But the quality of the optics is no better than it used to be.
Actually, that's not true. The field of optical engineering is seeing very rapid advancement, in part due to new lens coatings and in part due to the ability to model complex series of lenses in computers, rather than having to build and test them experimentally.
The result is that today's top-quality lenses are sharper and have less distortion (chromatic abberation, barrel and pincushion distortion, etc.) than older lenses. And the new technology is making some things possible that simply weren't before, like the new crop of superzooms that actually have reasonable performance across very wide zoom ranges.
For many kinds of shooting, image stabilization is huge, too.
However, none of that enables a tiny P&S lens to match a lens that is four or five times its size. When it comes to optics, size matters. A lot. Bigger glass can gather more light and focus and manage it better across the visible spectrum.
So while your statement isn't true -- optics ARE better, much better, and improving rapidly -- that still doesn't support the GPs claim that a modern P&S is superior to your wife's Pentax in terms of achievable image quality.
I've seen this argument before, and frankly, the cameras we have today (even the point and shoots) are MILES above the top of the line film cameras from even a couple decades ago. Yet, somehow, the photographers from that era (and earlier) managed to get these shots that people think are impossible on lower end models of today.
Nonsense.
I would much rather have a 70s-era SLR with a full set of Zeiss glass and a variety of films than a modern point & shoot camera. Not to mention a Hasselblad...
Beyond your valid point about light sensitivity, there are a lot of other issues, most of which have more to do with optics than with "features". And there's simply no way that, for example, a teeny P&S camera can compete with a 35mm SLR with a set of good lenses. And no way that 35mm can compete with a medium or large format camera.
And that's just looking at issues of sharpness and resolution. Other crucial aspects like the ability to control aperture to manage DOF, and to trade it off against shutter speed and noise/grain can also make the difference between a shot that's worth keeping and one that isn't. Old, purely-manual film cameras provided an ability to do those things that modern P&S cameras often do not, as do DSLRs.
Of course, the biggest difference between a P&S and an SLR (or DSLR) is the obvious one -- interchangeable lenses. And there are many shots you can get with, say, a 10 mm fisheye or a 600 mm telephoto that you simply cannot get with a P&S. And that's not even getting into the more oddball stuff like tilt/shift lenses.
However, a good photographer can, will, and should be able to get the shot regardless of the gear they are using. That's why they are called professionals after all.
Yeah, right. So give a sports photographer a little P&S camera and tell him to go photograph magazine cover shots of fast-moving action with tight DOF and nice bokeh in an indoor arena. He'll just tell you that you can't get those shots with that camera. Now, he might get some nice photos of the game, but they'll be more wide-angle, more static, with greater DOF, etc. Forced to shoot with that equipment, he'll modify the sorts of shots he takes to fit the capability of the equipment.
I have a bunch of digital cameras, ranging from the one on my cellphone, to various sizes of P&S cameras, to a couple of Canon DSLRs (a 350D and a 50D). I use all of them, and (I think) take nice pictures with all of them. I know my cellphone camera doesn't really have the resolution it's rated at, and it adds an odd sort of distortion that gives the images a surreal effect -- so I use it when I like, or at least don't mind, a photo with those limitations.
I carry my small P&S cameras when it's not convenient or feasible to carry a DSLR, but I recognize that low-light shots will not be good without a tripod, that if I want a pleasantly-blurred background I'm going to have to fake it in post, that I don't have filters, that I can't carefully manage the exposure for high dynamic range shots, etc.
Likewise, when choosing between the 350D and 50D for a session, I trade off size against resolution, ISO and frame rate. Sometimes I pick the 350D over the 50D even when size doesn't matter, because at low ISOs the 350D images are cleaner.
But I guess the only reason I can't capture the same images in the same circumstances with my cellphone as I can with my Canon 50D and L-series lenses is because I'm not a professional, right?
Its not the camera that takes great photos, its the photographer. Ive seen great pics taken with a crappy disposable film camera. Ive seen shitty photos taken with a DSLR.
This is why top photographers prefer disposable film cameras over DSLRs.
Oh, wait...
A good photographer can take good pictures with any camera -- but only because he factors the capabilities of the camera into the decision of which shots to take. Many images which could be captured with the flexibility provided by a high-end DSLR with the right lens cannot be captured effectively with a cheap point & shoot. Good equipment provides options. A poor photographer won't know how to use those options, but that doesn't mean a good photographer doesn't need them.
Very cool idea. Automatically clustering the complaints by location and keyword would make it simple for the city employees to handle a large volume of complaints, too -- which would be important because if it were easy to make such complaints, the volume would rise dramatically. But there should be plenty of information to collapse 100 complaints about the same pothole into one entry in a ticketing system. The number of complaints in a cluster is also a clue to its priority.
Even in certain banks, Linux is now being used or researched
Certain banks? Based on my experience consulting in the financial industry, you'd have a hard time finding a bank that doesn't have Linux deployed somewhere.
"like the Israelis do, well, that works very well... but makes the security screening vastly more manpower-intensive and time-consuming. " ...Except for that it takes the same number of people and you get through an Israeli airport in 25 minutes.
Not my experience. In my experience, it takes two screeners per passenger for 15 to 20 minutes, and you had better arrive at the airport at LEAST three hours before your flight. Four is better. Three might be cutting it close.
There's also this thing called the fourth amendment. Something about having the right not to be searched without probable cause.
Unfortunately, the courts long ago decided that the fourth amendment is irrelevant, because by choosing to fly and to enter the security checkpoint you chose to waive your fourth amendment protection. If you want to attack this issue on the basis of rights, you have to use the court-supported but less well-defined Right to Travel, and argue that in many cases air travel is the only reasonable option.
If you remove the security, do you really think terrorists wouldn't notice and exploit it again? Read up on history, once hijackings were extremely common UNTIL rigorous security measures put an end to them.
No, hijackings were extremely common until United Airlines Flight 93 put an end to them.
If we would just profile we wouldn't need half the security we have.
What do you mean by "profile"?
If you mean "apply extra scrutiny to certain ethnic and religious groups", that's completely and utterly useless.
If you mean "put all of the passengers under intense stress and watch their reactions", like the Israelis do, well, that works very well... but makes the security screening vastly more manpower-intensive and time-consuming. And, frankly, much more unpleasant than being briefly groped. I've flown out of Ben-Gurion airport a few times and I'd rather have a prostate exam.
The truth is that we simply don't need half the security we have. We should just roll it all back to pre-9/11 levels, keeping only the cockpit door locks. That plus the passengers' understanding that allowing their plane to be hijacked is likely to get them killed will mean that terrorism on airplanes will be restricted to killing passengers, making planes a low-value target. It's possible that the occasional Bad Thing will happen on an airplane, but it'll still be safer than driving.
Giant refrigerator sized cardboard boxes too.
My nine year-old made himself a cardboard coffin out of a big box. He actually sleeps in it most nights.
I'm trying to figure out if I should be impressed with his creativity, or concerned about his morbid bent.
Indeed.
Although one thing I've noticed is a serious reliance on "specialty parts". Now that spaceship comes in a kit with a special cockpit, wing, and landing gear piece..
The mindstorm stuff looks really cool though. I _really_ would have had a blast with something like that as a kid.
Lego is shifting away from those specialty parts, partly because of the limited building possibilities and partly because their exploding unique part count was increasing their costs. It's cheaper to produce fewer unique parts.
I'm happy to say that my kids have pretty much lost interest in the pre-designed kits. We recently stumbled over a big plastic bin full of thousands of assorted generic lego parts at a garage sale and they immediately pooled their money to buy it.
I'm telling you, Linus would accept a patch for in-kernel mixing tomorrow because it makes sense.
He would? After consistently rejecting many such patches for years, because audio mixing isn't something a kernel should do? When did he change his mind? And why?
I will do some research on GIT for fronting SVN in the future when I have time to play outside in the OpenSource domain again.
This will get you started: http://www.kernel.org/pub/software/scm/git/docs/git-svn.html
SVN is valid for many groups, teams, around the world and will continue to be so for the foreseeable future.
SVN is particularly excellent when you use GIT as the front-end. You get all of the super-easy local branching and merging, history rewriting and lighting speed of GIT for everyday use, and when you're ready to push your work to the rest of the team, you publish it to SVN.
BTW, yes, SVN branching and merging has gotten much better since merge tracking was added, but it's still too slow and clunky for effective use with the "branch per feature" pattern of development. It can work, but GIT runs circles around it.
Now, I admit that I haven't extensively used most of the tools in the "article", but I haven't needed what they offer above and beyond CVS.
SCM is one of those areas of technology where you can easily believe that something is unnecessary until you use it -- at which point you discover that it's impossible to live without it, in spite of the fact that you lived without it just fine for decades.
So what am I supposed to do? Start every one of my X applications from a different Xterm so they have a different controlling TTY (and do not close any of them)?
Umm, no, that would effectively place every program in a different group, which is the same as not having any grouping.
The point of this patch is to group related processes together and schedule them as a unit, rather than individually, effectively putting the whole set on an even footing with other individual processes with regard to scheduling fairness. That way a single very aggressive group can't starve the rest of the processes on the system.
You're right that to some degree this is a nerd feature, since it won't address situations of low responsiveness that "normal" desktop users might suffer -- but those situations aren't caused by a group of aggressively CPU-hungry processes. Generally they're I/O-related. But for people who do the sorts of things this patch addresses, it's a valuable improvement, and it's unlikely to have any negative effects for those who don't.
It is best to provider an interface through which a user can group all of the processes with his UID and leave the Desktop environment do the grouping. Or put something on the dbus which listens and follows who talks to whom to do the same. This will provide much better results than putting yet another simplistic heuristic in the kernel.
Maybe. It's not clear to me how grouping will help in general for desktop usage. Desktop usage tends to be a bunch of single-process programs which all need to be scheduled individually, and fairly, with preference given to interactive processes. I think perhaps something like Con Kolivas' BFS might be a better approach for the desktop, and BFS doesn't have the concept of groups at all. I wonder if BFS couldn't be improved by finding a way to identify and boost interactive tasks, rather than rigid fairness, but perhaps CK is right about focusing on rigid fairness in the scheduler, and we should then have the desktop environment automatically manage nice levels, moving the question of scheduling entirely out of the kernel.
However, I don't see how the DE could usefully make use of grouping. Perhaps as our desktop applications move more to multi-threaded implementations to take advantage of the exploding number of processor cores we will want the DE to automatically group the processes within a single program -- effectively extending the TTY-based grouping notion to pure GUI apps. But I really don't think enough applications are multi-threaded enough yet that we'll see significant benefit from that in the short term.
I could not yet try Mathematica 8 out, but I hope one will be able to turn the feature on and off.
It's only used when you explicitly invoke it by beginning your input with "=" or "==".
Did you read the article?
I think that they've hit on an excellent way of making imperfect freeform linguistics very useful. Basically, it's provided as an alternative to a more precise input language, one that works a lot of the time and when it does requires less effort by the user. When the computer can work out what you mean, it does. But it also always translates what it thought you meant into the precise, formal language and displays that so that you can tell whether it got it right. Further, if there are a few competing interpretations of the vague, freeform input, the tool makes it easy for you to pick the one that you meant.
As long as it gets what you mean right enough of the time, while allowing you to be vague enough that using the freeform input really is saving you mental effort, I can see tremendous value in this approach. And over time it will continue to get better.
I have no doubt that you'll quickly learn that there are some kinds of things which are just better-expressed in formal notation. Heck, we sometimes find that to be the case even when we're communicating with other humans. But there are also a LOT of cases where informal language works just as well and is a lot easier. I'm sure that set will be smaller when working with Mathematica than another person -- but it looks to me based on the article like the set is large enough to be useful.
In particular, I can definitely see using this as an easier route to a lot of the more obscure commands and options, such as the options to tweak graphs in various ways. And if it turns out that using the natural language didn't work, well, your next step was going to be to dig through the manual to figure out the formal notation anyway, so it didn't cost you much time.
I can see this freeform linguistics approach being quite valuable, even though it's quite imperfect.