Followup note: the reason why this is a big deal is that there is actually an unexamined problem in the whole definition of OO, in that *functions need to be objects in their own right* and they currently aren't - even prototype OO still has two fundamental entities, 'objects' and 'functions'. So many OO languages just don't grasp the concept that you could assign functions as values to slots at runtime. Prototype OO with functional features does, but it only takes that concept so far. There's still a lot of manual 'glue code' needed to build a GUI; ideally, you should be able to describe a set of widgets and bound events purely as a data structure and let the system figure out what if any 'code' needs to be generated. Dataflow engines like Flapjax are taking the next step in that direction; there's still a few more needed to completely close the loop and get to a really humane programming environment.
"I will also add, that I don't like prototyping as a concept and prefer OO over it. "
Huh? That makes no sense. Prototyping *is* OO. What it is not is *class-based* OO. It's in fact a much more flexible and natural way of doing OO and allows for better modelling of real-world situations.
The point of prototype-based OO is that you don't have two separate entities for 'class' and 'object': you only have one, and then you have the 'prototype' relation which is runtime-assignable to implement inheritance. Logically, this means you have much cleaner semantics (you don't need, ie, a meta-object protocol to define what kind of object a 'class' is); pragmatically, if you do it right, it means you can *change the class of an object at runtime*, which is in fact something that happens in The Real World.
(A Person starts out as a Baby, becomes a Child, then an Adult, all over the lifecycle of one object. Try doing *that* in Java with fixed class hierarchies. That's a bug in the Platonic idea of immutable 'classes', not a bug in the Real World.)
The other case where prototype OO rocks the house is where you have a heavily data-driven model with lots of singleton objects. In prototype OO, there's nothing at all weird about putting special-purpose methods or slots onto instance objects - in class-based OO, methods can only be put on classes, and instance objects can't have special-case slots at all.
Where this is a *huge* win is in GUI code, where you have all these special-case object instances representing widgets *with bound events*.
If you can't put method code directly on your instance objects, then you have to go through hideous convolutions of method calls on the class to simulate this by passing it as data. But with prototype OO, you can just put the code right where it needs to be, tweak the instance exactly how it needs to be tweaked, and it just works.
To really make this work you also need to have a language that allows the code you put on instance methods to also be special-cased in the same way; in other words, it needs to support the concept of *closures*. IE, the sweet spot for GUI development is a language that combines functional and prototype OO features. Javascript does this, and that's why it wins in the UI game.
"So your objections against nuclear power are concentration of wealth and power? Are you an environmentalist or a closet anarchist?"
Why should anarchism be in the closet, especially here on Slashdot of all places? It's the political expression of the end-to-end principle that makes the Internet work, and the free/open-source software philosophy. Everyone's free to participate. Don't put big heavy social machinery into the core, put it on the edges - empower every person and group to be autonomous as much as possible. Much the same concept behind federalism, or at least confederation.
Whatever you might have thought you learned in high school, political anarchism is not about throwing bombs.
Though I'm possibly not precisely an anarchist as the political group is defined, I'm pretty much with E F Schumacher, one of the fathers of modern environmentalism. You may have heard of the terms 'intermediate technology' and 'small is beautiful'? Same principle, again. Small stuff empowers the individual, is environmentally friendly, allows rapid evolution of solutions. Big stuff - heavy, dangerous, capital-intensive plant - has political implications - centralisation of power rather than individual empowerment. And politics is just pragmatism that looks at the future. It's about the shape of social and economic interactions that investment in certain technologies implies.
Don't believe the nuclear hype. Look at the companies behind it. Look at their track record. How much do you trust them not only not to screw up, but not to screw you over?
If nuclear power were the Internet, it would be like Ma Bell, IBM and (for you youngies who don't remember Ma Bell) the RIAA building it based on everything being time-sharing servers and mandatory DRM. Deep centralisation, no user-servicible parts on the end - backed by *serious* hazard to life and health if you try to modify the system.
You're a Slashdot geek, and yet you're cool with that for your energy future?
Affect/effect are one of those amusingly nasty little hand grenades in English. Handy crib sheet:
Affect, n: emotional response. "The Minister for Granola appeared to be displaying flattened affect during his speech, leading to suspicions that he was abusing his own product."
Effect, n: causal result. "The effect of the proposed granola reform would be catastrophic."
Affect, v: alter. "The proposed reforms will affect the granola industry greatly."
Effect, v: put into immediate action. "If elected, I will effect sweeping reforms of the granola trade."
"He's probably more interested in business management than flash and, with a special ops background, may belive that if you're going to build a secret ninja hacker cyberforce, you might want to do it with a somewhat lower public profile."
It surprises me how few people are putting forward this, the obvious explanation.
"Cancel" the program, heh, yes of course, Sir. Very good thinking Sir. Wink wink. Say no more.
"Quite clearly we are losing the media war in Europe (battle over their hearts and minds, if you like). The two stories I read on the Brittish and German news are extremely pro-Russia (and strongly anti-US). I think this is yet another sign of our declining power in the world, and it makes me sad."
Why do you think that power over the 'hearts and minds' of allegedly free people is inherently a *good* thing? Seriously, on what planet does that ideology even *begin* to make sense?
The world's hearts and minds are their own. Quit thinking you own them and have the right to control them.
"IT departments who just spent 6 years getting XP stable enough"
I second this. On a yearly rollout cycle, it literally does take *years* to bed in a new version of Windows and work through all the subtle little bugs that need tweaking.
XP's finally working great for us. We don't want to break it.
I think the biggest single usability problem is that our current coding tools and 'best practices' force interface design and low-level code to be merged together in a big monolithic box called 'the application'. This is a technological dead end.
Increasingly, what I do on a computer is NOT 'run applications', but 'browse data and construct new data sets'. But the tools used for 'application design' are locked into a developer-user separation that gives us this negative feedback cycle: coders can't design, interface designers can't code, and users whose data and machine it actually is and know what works for them have no way to modify the design of either.
Find a way to SAFELY separate 'interface' from 'internals' at every possible level so that the USER can tweak the interface to 'scratch their own itch', and you'll see an explosion of open-source interface design similar to what happened when the Web appeared.
A key I think is switching from imperative, even OO programming, to a dataflow mindset. Ted Nelson is on the right track with Xanadu and 'applitudes', but he can't coherently describe his vision and nobody is listening. And he's patented his ideas. Pay attention to him, filter out the crap, then try to work around his patents.
Look at Cells, FrTime, Flapjax, and spreadsheets, then generalise that model. That's the missing layer we need to have instead of things like COM. It might be possible to write something like that as a layer over D-BUS, expose *all* data in a system from the filesystem to mouse clicks as *fully* scriptable dataflow events, then add a *thin* GUI layer that lets anyone at runtime construct and run an 'interface' from a small text file that aggregates such events into widgets.
If you want, then add a thin GUI builder which lets you construct such a text file by pointing and clicking, but it's not necessary (and it's necessary that ALL functionality exposed by a GUI layer ALSO be available to a non-imperative, composable scripting language).
Then you'll have something which could seriously take over the world: a Web-like dataflow fabric which can construct runtime pipelines of events and filters and make tiny, user-specific 'tasklets' much like accounting types build spreadsheets.
Even the simplest of 'users' would then be able to be 'interface designers' in the same way a file manager allows anyone to be a librarian. Let people publish such tasklets freely to the Net (like RSS aggregators or Google mashups), and you've solved the interface problem.
The key is to make all your data-processing widgets fully composable, which means to make them pure-functional, which is what existing OO frameworks aren't.
"Why is that? Well, most likely because the average user never needs automation at any large scale. The more advanced users who require automation because they actually do regularly and repeatedly perform tasks on large amounts of material are willing to take the time to learn to use existing tools which, as you say, work just fine. "
Actually, I don't believe that's true. I think the average user does need automation.
The problem is that the way our current GUIs are designed, it is very very difficult for an end-user to automate anything without leaving the shell.
I have this problem working in IT. Most of what I do requires integrating data from multiple sources and programs. It is exceptionally difficult to do this with the tools we have as given. I can write scripts and applications, but scripts and apps are only part of the problem. They only solve *tasks*. There is still no easy way in our current OS designs of tying multiple simultaneous tasks together into a user-configurable workflow.
This is a design problem in the IT world which we will need to overcome to get to the next level of user-friendliness.
Want to see more user-friendly software? *Make it so you don't have to be a 'programmer' to write a GUI*.
By which I don't mean 'write another silly point-and-click shell', but write a programming language that is actually suited to the job of designing GUIs and doesn't force the user to deal with irrelevant crud like our current C++-inspired methodology does. The C family is probably the worst possible language class for trying to develop dynamically runtime-configurable, parallel, pipes of composable widgets that consume and emit event streams. We only use these primitive tools because of historical inertia. The concept of 'compilation', like manual 'file saving', can and should be automated away.
We need a language and operating system that lets us describe 'programs' as primarily data-driven relationships and abstracts out any concept of 'control flow', because it's not relevant to the task of structuring a user's data workspace. Look into homoiconic languages like Lisp and Scheme, dataflow engines like FrTime and Flapjax, and the Xanadu concept of 'applitudes'. Or into the good old spreadsheet. There's a reason Excel is used extensively as a computation fabric by end-users. Generalise the cell-recomputation model, link it to a thin, completely runtime-user-reconfigurable GUI layer, and you have the user interface of the future.
Currently, it is hard for end-users to write GUIs *because GUI design tools are fundamentally broken*. They are broken because they are unsafe: they require the coder to manually write far too much C++-style event-driven 'glue code' and 'design patterns', rather than just writing a terse description of 'what connects to what' and let the infrastructure figure it out and optimise the low-level code. We need to be using data-driven, declarative/functional languages, rather than brittle imperative ones. Because the GUI 'application design' space is not in fact imperative in nature. It's not about implementing 'applications' that solve 'tasks' - it should be about *storing and processing data*. 'Tasks' are things a user needs to organise for themselves in the same way they decide what folder to store files in.
Fix the languages, and the GUI design problem will fix itself. I imagine a world where 'application designers' will just publish the basic data-transformation fundamentals of software, and 'interfaces' are things like HTML themes or Firefox skins or RSS aggregators : rapidly-changing things which the end-user free-software ecosystem can safely reconfigure until they actually do what users want.
"But what does taste have to do with it? Shouldn't all that matters be the aptness of the comparison?"
Yes, exactly.
I'm sick and tired of political discourse being filled with bizarre weasel words like 'tasteless' and 'tired' and 'offensive' and 'disloyal' to describe claims the speaker doesn't like which are, objectively, either *true* or *false*.
I'm sick of politics being about 'opinions'.
Politics *isn't* about opinions. It's about reality. People have strong opinions *about* it, yes, just as they do about science, but those opinions do not *determine* the truth or falsity of political claims: results do. It's not a matter of fashion. Truth doesn't get 'tired' or 'stale' from multiple repetition and lies do not get proven consensus acceptance no matter how many years have passed.
That's why history is an active research subject - we're too polite to admit it, but it's because HISTORY MAKERS LIE, and history is largely the science of sorting through the lies after the fact and determining from documentary evidence just how we were deceived and then guessing at why. If we could take public figures' words at face value at the time they say them, we wouldn't *need* either history or political science.
Political claims, like every other claim, need to be verified against objective, trustworthy evidence, one political system is NOT identical to another, and torture and oppression remain torture and oppression regardless of whether it's the 'good guys' or the 'bad guys' doing it.
And while Israel's policies toward Palestine are not identical to Nazi Germany's toward the Jews, there sure are depressing similarities.
And it just goes to show that suffering atrocity and swearing 'never again shall this happen TO US' is not quite the same as swearing 'never again shall we allow this happen, full stop'.
The other interesting thing is that America's 'torture lite' techniques are, from what I've read, not only nearly identical to the Cold War KGB's (sleep deprivation, sound, heat/cold, stress positions) -- but they actually migrated to the US military lexicon FROM the Soviets via the Cold War.
What you hate, you really DO stand a strong chance of becoming. Literally and not figuratively. Rivals copy each other and adopt the methodologies which seem to work.
"Turns out, the government really did have a Stargate Project [wikipedia.org] -- it was just about psychics, not aliens. And they didn't find any. Of either."
Actually they did. You might want to read Mind-Reach, the 1977 original book about SCANATE, the SRI project that later became GRILL FLAME then was closed (at least officially) by the CIA under the name STAR GATE. Some of their 'hits' detailed in this book are pretty darn impressive.
The tricky thing about remote viewing is not that it doesn't work, but that it's hard to separate the 'signal' from the 'noise'.
And of course, the results are incompatible with assumptions made by some of our fundamental physical theories. Whatever information channel ESP uses, it does not appear to obey the inverse square law or respect light cones, so it's not EM-based. This makes it difficult to figure out how to maximise the effect, since we don't have a good mathematical model for how it works. Some scientists (or science-believing people, as opposed to active researchers) are uncomfortable with admitting this kind of uncertainty into their personal models of the universe. It's a lot easier to believe that we really do understand how the universe works than to realise that actually, we only understand parts of it and our working assumptions may need to be rethought.
But when you get significant results that contradict theory, it's the theory that should change, if you're doing science.
"You mean like Jimmy Carter? Obama is dangerously close to him both in policies and experience level. And you remember the damage Carter did to us in just four years."
Damage?
Oh, you mean by correctly predicting the energy crisis and proposing serious workable solutions, anticipating Peak Oil by three decades?
And then getting run out of town for his honesty and far-sightedness?
That kind of damage?
You could use a lot more Jimmy Carters, and if Obama is a tenth of the US President he was, you folks will be counting yourself lucky.
"Hydrogen for example is light, small in volume..."
No, actually, that's exactly wrong. Being light by definition means it is NOT "small in volume". It takes a huge volume of H2 in room temperature gas form to store a similar amount of energy to a heavier molecule. The volume problem is why it's a pain to store unless you go to cryogenics, hydrides, or other complicated systems.
"Terry Pratchett is an economical retard. In that book he states that an eternal and free energy/labor supply equates to poverty for everyone."
Well, it does, in our current market- and employment-based economy. Marshall Brain makes essentially the same point in his 'Robot Freedom' essays: http://marshallbrain.com/robotic-nation.htm
That doesn't mean massive automation is a bad thing. What it does mean is that to cope with the existence of such automation, we will need to transition our economy from one where 'you don't eat unless you don't work / sell / finance others' labour', to a guaranteed Universal Basic Income system where you get 'credit' for goods and services simply based on being a living human.
And to most Americans, that transition sounds like the worst kind of SOCIALISM.
Even to those who aren't ideologically predisposed to knee-jerk hatred of a communical resource-sharing society not driven by private fear and greed, there are some very curly issues involved based around the fact that we simply haven't yet deployed such a system on a wide scale.
It's sort of like the 'second system effect' in programming: we can see that the current economy is horribly broken and crashes regularly and kills people, but it is a patchwork of both bugs and bug-fixes. Starting from scratch with a whole new concept is attractive, but is likely to crash even harder if we don't work out the bugs first in small systems.
It doesn't help that the most widely publicised Economy 2.0, Marxist-Lenist Communism, was a collossal social and environmental disaster. It gives the false impression that 'there is no alternative' to the present system.
I'm hopeful the Free / Open Source Software movement is the first step toward demonstrating that an 'economy of abundance' is at least *conceivable*, but it's a long way from showing how the same principles could organise, for example, a supermarket.
Making the transition *without* invoking bloody revolution, as the Communists tried, and without destabilising the entire house of cards that is our present what-passes-for-an-economy, is going to be the tricky bit.
Yes. I see Iron Man, in both comic and film versions, as attempting to straddle a delicate line between left-wing and right-wing positions on war, to avoid offending anyone.
The result to me was a sort of standard wishy-washy 'yay, hoo-ra for war, kill our enemies, death to tyrants, hooray for defence contractors, but booooo those naaasty Arms Dealers on our side who SOMEHOW manage to sell OUR WEAPONS to the WRONG PEOPLE! Who kill AMERICAN BOYS IN UNIFORM (instead of Afghanistani kids having lunch)' position that we've seen for years.
It's not actually an honest representation of what happens in wars and arms manufacture, and how the American defence industry is funded, and the shifting geopolitical alliances that see people like Al Quaeda, the Taliban and the mujihadeen be CIA assets and 'freedom fighters' against the Soviets in, say, Rambo III but Evil Incarnate in . But it's par for the course for American popular fiction.
And the shadowy 'rogue arms dealer' is up there as one of the all-time favourite movie villains. Especially favourite on the left wing because you can have a villain who's NOT a foreigner, so avoid racism, and you can be all moral and righteous about nonproliferation, how weapons should be used with responsibility, etc. While still toeing the right-wing line about death to tyrants, compulsory freedom at the point of an American gun to all nations, take my nukes from my cold dead hand, etc.
But does anyone ever ask where those evil 'arms dealers' GET their super-top-secret weapons?
They don't fall off the back of a truck, and they don't get sold to *anyone* without top-level US Government approval, that's for sure.
"Like it or not, you need an army to remain a secure/solvent country. And sometimes you need to go to war. But that doesn't mean war should be painted in a positive or glorious light."
I'm sorry, but that simply is not logical.
If you think war is NECESSARY, you by definition DO think that it's honourable. That it's an occupation that good people need to be involved in, and should have their choice to do so supported rather than criticised.
That means that given you believe that 'you need an army to remain a secure/solvent country', then yes, you DO think that fighting a war in defence of one's country IS 'positive' and 'glorious'. Glory doesn't always mean 'fun' and 'pleasant'. It means 'it was a horrible thing but someone needed to do it, and I'm glad they did'.
But if war really is a BAD thing - not just an unpleasant necessity like cleaning toilets, digging the garden or nursing the sick - NOBODY needs to do it. In fact, everyone needs to NOT do it.
I'm anti-war, and I have the courage to admit that that means that I think that people who have chosen to fight wars even for the best of reasons have actually made a terrible mistake.
Why is this so hard to accept? Why this constant double-talk of 'oh, I'm anti-war, but pro-military'?
You simply can't be both. Not unless you transform the armed forces into a force whose mission does NOT involve killing people.
Slashdot Mirror
"i agree with you except i think you mean "write something new about something old""
Innovate kind of like you did last year, none of this new stuff.
I don't know, fly casual.
"For something to be OO, it has to support inheritance. Prototyping is more like making copies of something and then adding some stuff on."
Sorry, wrong. That's one definition of "prototyping" but it's not the type used by Javascript (or the nicer prototype OO language Lua).
You *do* have inheritance in Javascript. The 'Prototype' property is looked up dynamically and acts just like a class in a class-based language.
Followup note: the reason why this is a big deal is that there is actually an unexamined problem in the whole definition of OO, in that *functions need to be objects in their own right* and they currently aren't - even prototype OO still has two fundamental entities, 'objects' and 'functions'. So many OO languages just don't grasp the concept that you could assign functions as values to slots at runtime. Prototype OO with functional features does, but it only takes that concept so far. There's still a lot of manual 'glue code' needed to build a GUI; ideally, you should be able to describe a set of widgets and bound events purely as a data structure and let the system figure out what if any 'code' needs to be generated. Dataflow engines like Flapjax are taking the next step in that direction; there's still a few more needed to completely close the loop and get to a really humane programming environment.
But that's a story for another day.
"I will also add, that I don't like prototyping as a concept and prefer OO over it. "
Huh? That makes no sense. Prototyping *is* OO. What it is not is *class-based* OO. It's in fact a much more flexible and natural way of doing OO and allows for better modelling of real-world situations.
The point of prototype-based OO is that you don't have two separate entities for 'class' and 'object': you only have one, and then you have the 'prototype' relation which is runtime-assignable to implement inheritance. Logically, this means you have much cleaner semantics (you don't need, ie, a meta-object protocol to define what kind of object a 'class' is); pragmatically, if you do it right, it means you can *change the class of an object at runtime*, which is in fact something that happens in The Real World.
(A Person starts out as a Baby, becomes a Child, then an Adult, all over the lifecycle of one object. Try doing *that* in Java with fixed class hierarchies. That's a bug in the Platonic idea of immutable 'classes', not a bug in the Real World.)
The other case where prototype OO rocks the house is where you have a heavily data-driven model with lots of singleton objects. In prototype OO, there's nothing at all weird about putting special-purpose methods or slots onto instance objects - in class-based OO, methods can only be put on classes, and instance objects can't have special-case slots at all.
Where this is a *huge* win is in GUI code, where you have all these special-case object instances representing widgets *with bound events*.
If you can't put method code directly on your instance objects, then you have to go through hideous convolutions of method calls on the class to simulate this by passing it as data. But with prototype OO, you can just put the code right where it needs to be, tweak the instance exactly how it needs to be tweaked, and it just works.
To really make this work you also need to have a language that allows the code you put on instance methods to also be special-cased in the same way; in other words, it needs to support the concept of *closures*. IE, the sweet spot for GUI development is a language that combines functional and prototype OO features. Javascript does this, and that's why it wins in the UI game.
"So your objections against nuclear power are concentration of wealth and power? Are you an environmentalist or a closet anarchist?"
Why should anarchism be in the closet, especially here on Slashdot of all places? It's the political expression of the end-to-end principle that makes the Internet work, and the free/open-source software philosophy. Everyone's free to participate. Don't put big heavy social machinery into the core, put it on the edges - empower every person and group to be autonomous as much as possible. Much the same concept behind federalism, or at least confederation.
Whatever you might have thought you learned in high school, political anarchism is not about throwing bombs.
Though I'm possibly not precisely an anarchist as the political group is defined, I'm pretty much with E F Schumacher, one of the fathers of modern environmentalism. You may have heard of the terms 'intermediate technology' and 'small is beautiful'? Same principle, again. Small stuff empowers the individual, is environmentally friendly, allows rapid evolution of solutions. Big stuff - heavy, dangerous, capital-intensive plant - has political implications - centralisation of power rather than individual empowerment. And politics is just pragmatism that looks at the future. It's about the shape of social and economic interactions that investment in certain technologies implies.
Don't believe the nuclear hype. Look at the companies behind it. Look at their track record. How much do you trust them not only not to screw up, but not to screw you over?
If nuclear power were the Internet, it would be like Ma Bell, IBM and (for you youngies who don't remember Ma Bell) the RIAA building it based on everything being time-sharing servers and mandatory DRM. Deep centralisation, no user-servicible parts on the end - backed by *serious* hazard to life and health if you try to modify the system.
You're a Slashdot geek, and yet you're cool with that for your energy future?
Don't forget Duke Nukem 2, which was still a side-scroller but was pretty impressive.
Affect/effect are one of those amusingly nasty little hand grenades in English. Handy crib sheet:
Affect, n: emotional response. "The Minister for Granola appeared to be displaying flattened affect during his speech, leading to suspicions that he was abusing his own product."
Effect, n: causal result. "The effect of the proposed granola reform would be catastrophic."
Affect, v: alter. "The proposed reforms will affect the granola industry greatly."
Effect, v: put into immediate action. "If elected, I will effect sweeping reforms of the granola trade."
"He's probably more interested in business management than flash and, with a special ops background, may belive that if you're going to build a secret ninja hacker cyberforce, you might want to do it with a somewhat lower public profile."
It surprises me how few people are putting forward this, the obvious explanation.
"Cancel" the program, heh, yes of course, Sir. Very good thinking Sir. Wink wink. Say no more.
"Quite clearly we are losing the media war in Europe (battle over their hearts and minds, if you like). The two stories I read on the Brittish and German news are extremely pro-Russia (and strongly anti-US). I think this is yet another sign of our declining power in the world, and it makes me sad."
Why do you think that power over the 'hearts and minds' of allegedly free people is inherently a *good* thing? Seriously, on what planet does that ideology even *begin* to make sense?
The world's hearts and minds are their own. Quit thinking you own them and have the right to control them.
"IT departments who just spent 6 years getting XP stable enough"
I second this. On a yearly rollout cycle, it literally does take *years* to bed in a new version of Windows and work through all the subtle little bugs that need tweaking.
XP's finally working great for us. We don't want to break it.
"In my experience physicists are generally rather cool, worldly people who have well developed personal lives."
And in any case, they have lasers.
I think the biggest single usability problem is that our current coding tools and 'best practices' force interface design and low-level code to be merged together in a big monolithic box called 'the application'. This is a technological dead end.
Increasingly, what I do on a computer is NOT 'run applications', but 'browse data and construct new data sets'. But the tools used for 'application design' are locked into a developer-user separation that gives us this negative feedback cycle: coders can't design, interface designers can't code, and users whose data and machine it actually is and know what works for them have no way to modify the design of either.
Find a way to SAFELY separate 'interface' from 'internals' at every possible level so that the USER can tweak the interface to 'scratch their own itch', and you'll see an explosion of open-source interface design similar to what happened when the Web appeared.
A key I think is switching from imperative, even OO programming, to a dataflow mindset. Ted Nelson is on the right track with Xanadu and 'applitudes', but he can't coherently describe his vision and nobody is listening. And he's patented his ideas. Pay attention to him, filter out the crap, then try to work around his patents.
Look at Cells, FrTime, Flapjax, and spreadsheets, then generalise that model. That's the missing layer we need to have instead of things like COM. It might be possible to write something like that as a layer over D-BUS, expose *all* data in a system from the filesystem to mouse clicks as *fully* scriptable dataflow events, then add a *thin* GUI layer that lets anyone at runtime construct and run an 'interface' from a small text file that aggregates such events into widgets.
If you want, then add a thin GUI builder which lets you construct such a text file by pointing and clicking, but it's not necessary (and it's necessary that ALL functionality exposed by a GUI layer ALSO be available to a non-imperative, composable scripting language).
Then you'll have something which could seriously take over the world: a Web-like dataflow fabric which can construct runtime pipelines of events and filters and make tiny, user-specific 'tasklets' much like accounting types build spreadsheets.
Even the simplest of 'users' would then be able to be 'interface designers' in the same way a file manager allows anyone to be a librarian. Let people publish such tasklets freely to the Net (like RSS aggregators or Google mashups), and you've solved the interface problem.
The key is to make all your data-processing widgets fully composable, which means to make them pure-functional, which is what existing OO frameworks aren't.
So OSS is less like a restaurant and more like the kitchen in a backpackers hostel?
"Why is that? Well, most likely because the average user never needs automation at any large scale. The more advanced users who require automation because they actually do regularly and repeatedly perform tasks on large amounts of material are willing to take the time to learn to use existing tools which, as you say, work just fine. "
Actually, I don't believe that's true. I think the average user does need automation.
The problem is that the way our current GUIs are designed, it is very very difficult for an end-user to automate anything without leaving the shell.
I have this problem working in IT. Most of what I do requires integrating data from multiple sources and programs. It is exceptionally difficult to do this with the tools we have as given. I can write scripts and applications, but scripts and apps are only part of the problem. They only solve *tasks*. There is still no easy way in our current OS designs of tying multiple simultaneous tasks together into a user-configurable workflow.
This is a design problem in the IT world which we will need to overcome to get to the next level of user-friendliness.
Want to see more user-friendly software? *Make it so you don't have to be a 'programmer' to write a GUI*.
By which I don't mean 'write another silly point-and-click shell', but write a programming language that is actually suited to the job of designing GUIs and doesn't force the user to deal with irrelevant crud like our current C++-inspired methodology does. The C family is probably the worst possible language class for trying to develop dynamically runtime-configurable, parallel, pipes of composable widgets that consume and emit event streams. We only use these primitive tools because of historical inertia. The concept of 'compilation', like manual 'file saving', can and should be automated away.
We need a language and operating system that lets us describe 'programs' as primarily data-driven relationships and abstracts out any concept of 'control flow', because it's not relevant to the task of structuring a user's data workspace. Look into homoiconic languages like Lisp and Scheme, dataflow engines like FrTime and Flapjax, and the Xanadu concept of 'applitudes'. Or into the good old spreadsheet. There's a reason Excel is used extensively as a computation fabric by end-users. Generalise the cell-recomputation model, link it to a thin, completely runtime-user-reconfigurable GUI layer, and you have the user interface of the future.
Currently, it is hard for end-users to write GUIs *because GUI design tools are fundamentally broken*. They are broken because they are unsafe: they require the coder to manually write far too much C++-style event-driven 'glue code' and 'design patterns', rather than just writing a terse description of 'what connects to what' and let the infrastructure figure it out and optimise the low-level code. We need to be using data-driven, declarative/functional languages, rather than brittle imperative ones. Because the GUI 'application design' space is not in fact imperative in nature. It's not about implementing 'applications' that solve 'tasks' - it should be about *storing and processing data*. 'Tasks' are things a user needs to organise for themselves in the same way they decide what folder to store files in.
Fix the languages, and the GUI design problem will fix itself. I imagine a world where 'application designers' will just publish the basic data-transformation fundamentals of software, and 'interfaces' are things like HTML themes or Firefox skins or RSS aggregators : rapidly-changing things which the end-user free-software ecosystem can safely reconfigure until they actually do what users want.
"But what does taste have to do with it? Shouldn't all that matters be the aptness of the comparison?"
Yes, exactly.
I'm sick and tired of political discourse being filled with bizarre weasel words like 'tasteless' and 'tired' and 'offensive' and 'disloyal' to describe claims the speaker doesn't like which are, objectively, either *true* or *false*.
I'm sick of politics being about 'opinions'.
Politics *isn't* about opinions. It's about reality. People have strong opinions *about* it, yes, just as they do about science, but those opinions do not *determine* the truth or falsity of political claims: results do. It's not a matter of fashion. Truth doesn't get 'tired' or 'stale' from multiple repetition and lies do not get proven consensus acceptance no matter how many years have passed.
That's why history is an active research subject - we're too polite to admit it, but it's because HISTORY MAKERS LIE, and history is largely the science of sorting through the lies after the fact and determining from documentary evidence just how we were deceived and then guessing at why. If we could take public figures' words at face value at the time they say them, we wouldn't *need* either history or political science.
Political claims, like every other claim, need to be verified against objective, trustworthy evidence, one political system is NOT identical to another, and torture and oppression remain torture and oppression regardless of whether it's the 'good guys' or the 'bad guys' doing it.
And while Israel's policies toward Palestine are not identical to Nazi Germany's toward the Jews, there sure are depressing similarities.
And it just goes to show that suffering atrocity and swearing 'never again shall this happen TO US' is not quite the same as swearing 'never again shall we allow this happen, full stop'.
The other interesting thing is that America's 'torture lite' techniques are, from what I've read, not only nearly identical to the Cold War KGB's (sleep deprivation, sound, heat/cold, stress positions) -- but they actually migrated to the US military lexicon FROM the Soviets via the Cold War.
What you hate, you really DO stand a strong chance of becoming. Literally and not figuratively. Rivals copy each other and adopt the methodologies which seem to work.
That second link should be http://www.amazon.com/Irreducible-Mind-hard-find-contemporary/dp/0742547922/ref=pd_bbs_sr_1?ie=UTF8&s=books&qid=1217801555&sr=8-1
The mother of all ESP textbooks.
"Turns out, the government really did have a Stargate Project [wikipedia.org] -- it was just about psychics, not aliens. And they didn't find any. Of either."
Actually they did. You might want to read Mind-Reach, the 1977 original book about SCANATE, the SRI project that later became GRILL FLAME then was closed (at least officially) by the CIA under the name STAR GATE. Some of their 'hits' detailed in this book are pretty darn impressive.
http://www.amazon.com/Mind-Reach-Scientists-Psychic-Abilities-Consciousness/dp/1571744142/ref=pd_bbs_sr_1?ie=UTF8&s=books&qid=1217800041&sr=8-1
The tricky thing about remote viewing is not that it doesn't work, but that it's hard to separate the 'signal' from the 'noise'.
And of course, the results are incompatible with assumptions made by some of our fundamental physical theories. Whatever information channel ESP uses, it does not appear to obey the inverse square law or respect light cones, so it's not EM-based. This makes it difficult to figure out how to maximise the effect, since we don't have a good mathematical model for how it works. Some scientists (or science-believing people, as opposed to active researchers) are uncomfortable with admitting this kind of uncertainty into their personal models of the universe. It's a lot easier to believe that we really do understand how the universe works than to realise that actually, we only understand parts of it and our working assumptions may need to be rethought.
But when you get significant results that contradict theory, it's the theory that should change, if you're doing science.
See also http://www.amazon.com/Extraordinary-Knowing-Science-Skepticism-Inexplicable/dp/0553382233/ref=pd_bbs_3?ie=UTF8&s=books&qid=1217801555&sr=8-3
and http://www.amazon.com/Extraordinary-Knowing-Science-Skepticism-Inexplicable/dp/0553382233/ref=pd_bbs_3?ie=UTF8&s=books&qid=1217801555&sr=8-3
for some very recent books detailing the experimental support for the reality of ESP.
"You mean like Jimmy Carter? Obama is dangerously close to him both in policies and experience level. And you remember the damage Carter did to us in just four years."
Damage?
Oh, you mean by correctly predicting the energy crisis and proposing serious workable solutions, anticipating Peak Oil by three decades?
And then getting run out of town for his honesty and far-sightedness?
That kind of damage?
You could use a lot more Jimmy Carters, and if Obama is a tenth of the US President he was, you folks will be counting yourself lucky.
"Others could simply be flushed into the volcano."
Now that's an intriguing option for the modern bathroom.
"Hydrogen for example is light, small in volume..."
No, actually, that's exactly wrong. Being light by definition means it is NOT "small in volume". It takes a huge volume of H2 in room temperature gas form to store a similar amount of energy to a heavier molecule. The volume problem is why it's a pain to store unless you go to cryogenics, hydrides, or other complicated systems.
"Terry Pratchett is an economical retard. In that book he states that an eternal and free energy/labor supply equates to poverty for everyone."
Well, it does, in our current market- and employment-based economy. Marshall Brain makes essentially the same point in his 'Robot Freedom' essays: http://marshallbrain.com/robotic-nation.htm
That doesn't mean massive automation is a bad thing. What it does mean is that to cope with the existence of such automation, we will need to transition our economy from one where 'you don't eat unless you don't work / sell / finance others' labour', to a guaranteed Universal Basic Income system where you get 'credit' for goods and services simply based on being a living human.
And to most Americans, that transition sounds like the worst kind of SOCIALISM.
Even to those who aren't ideologically predisposed to knee-jerk hatred of a communical resource-sharing society not driven by private fear and greed, there are some very curly issues involved based around the fact that we simply haven't yet deployed such a system on a wide scale.
It's sort of like the 'second system effect' in programming: we can see that the current economy is horribly broken and crashes regularly and kills people, but it is a patchwork of both bugs and bug-fixes. Starting from scratch with a whole new concept is attractive, but is likely to crash even harder if we don't work out the bugs first in small systems.
It doesn't help that the most widely publicised Economy 2.0, Marxist-Lenist Communism, was a collossal social and environmental disaster. It gives the false impression that 'there is no alternative' to the present system.
I'm hopeful the Free / Open Source Software movement is the first step toward demonstrating that an 'economy of abundance' is at least *conceivable*, but it's a long way from showing how the same principles could organise, for example, a supermarket.
Making the transition *without* invoking bloody revolution, as the Communists tried, and without destabilising the entire house of cards that is our present what-passes-for-an-economy, is going to be the tricky bit.
Is it just me, or does anyone else reading the two rovers' names mentally translate them as Motive and Opportunity?
Just me? Thought so.
Yes. I see Iron Man, in both comic and film versions, as attempting to straddle a delicate line between left-wing and right-wing positions on war, to avoid offending anyone.
The result to me was a sort of standard wishy-washy 'yay, hoo-ra for war, kill our enemies, death to tyrants, hooray for defence contractors, but booooo those naaasty Arms Dealers on our side who SOMEHOW manage to sell OUR WEAPONS to the WRONG PEOPLE! Who kill AMERICAN BOYS IN UNIFORM (instead of Afghanistani kids having lunch)' position that we've seen for years.
It's not actually an honest representation of what happens in wars and arms manufacture, and how the American defence industry is funded, and the shifting geopolitical alliances that see people like Al Quaeda, the Taliban and the mujihadeen be CIA assets and 'freedom fighters' against the Soviets in, say, Rambo III but Evil Incarnate in . But it's par for the course for American popular fiction.
And the shadowy 'rogue arms dealer' is up there as one of the all-time favourite movie villains. Especially favourite on the left wing because you can have a villain who's NOT a foreigner, so avoid racism, and you can be all moral and righteous about nonproliferation, how weapons should be used with responsibility, etc. While still toeing the right-wing line about death to tyrants, compulsory freedom at the point of an American gun to all nations, take my nukes from my cold dead hand, etc.
But does anyone ever ask where those evil 'arms dealers' GET their super-top-secret weapons?
They don't fall off the back of a truck, and they don't get sold to *anyone* without top-level US Government approval, that's for sure.
"Like it or not, you need an army to remain a secure/solvent country. And sometimes you need to go to war. But that doesn't mean war should be painted in a positive or glorious light."
I'm sorry, but that simply is not logical.
If you think war is NECESSARY, you by definition DO think that it's honourable. That it's an occupation that good people need to be involved in, and should have their choice to do so supported rather than criticised.
That means that given you believe that 'you need an army to remain a secure/solvent country', then yes, you DO think that fighting a war in defence of one's country IS 'positive' and 'glorious'. Glory doesn't always mean 'fun' and 'pleasant'. It means 'it was a horrible thing but someone needed to do it, and I'm glad they did'.
But if war really is a BAD thing - not just an unpleasant necessity like cleaning toilets, digging the garden or nursing the sick - NOBODY needs to do it. In fact, everyone needs to NOT do it.
I'm anti-war, and I have the courage to admit that that means that I think that people who have chosen to fight wars even for the best of reasons have actually made a terrible mistake.
Why is this so hard to accept? Why this constant double-talk of 'oh, I'm anti-war, but pro-military'?
You simply can't be both. Not unless you transform the armed forces into a force whose mission does NOT involve killing people.
"Cosmic rays travel through the Universe with enough energies to boil a cup of water (in one single proton)."
Nice cuppa tea and a positron?