Man... I thought we covered this with the last story related to the proposed Khronos 3d api. This is nothing like VRML. It is a javascript api to use graphics hardware.
Nothing to do with markup of any kind (aside from the xml in Collada, which is not necessarily part of the standard). Ugh.
This history of javascript and the DOM is pretty complex and interesting. For example, the practice of wrapping native DOM objects has its roots in a cross-browser hack. IE's implementation of the DOM used COM, which meant that methods couldn't be added at runtime (maybe things got better with.net... I wouldn't know).
It turned out to be really useful, even aside from the cross-browser aspect. It allows jQuery to work with other libraries, and even multiple instances of itself. It gets around the problem of browsers deciding to implement new methods in the DOM, when your library had implemented the same feature earlier. Even Prototype has seen the light on this one... the assurance that your script will play nice with other unknown code is a killer feature.
With the iteration features, I'm guessing you're referring to stuff like each(). However, it's important to note that this is not a method added to a native DOM element; it's a method of a jQuery wrapper. And the practice of wrapping elements is now part of the architectural philosophy of things like jQuery (it's no longer just an IE workaround, it's useful on its own).
On the last paragraph in your comment... Yes. The death of ES4 was a good thing for this reason. We don't need classical inheritance and namespaces.
Also related to iteration, my #1 wish for implementations is optimization for tail recursion. The browsers currently only support 1000-5000 (I think) levels of recursion. After getting hooked on functional languages, using clunky procedural iteration hurts. There's tons of great functional code out there that would get ported to javascript overnight if browsers could manage the stack for recursion. Maybe Microsoft could get some of their Haskell guys to help out.
Well... JSON wasn't really "invented". It's really just the syntax for an object literal in javascript. JSON parsing is not much more than eval() and a regular expression.
It works beautifully.
Not going to touch the rest of your comment. JSON is more useful in this context than xml.
Or they should be. Unfortunately, there are a lot of "how-to"s on the web that use a null key passphrase to make scripting easier. That openssh allows this is a problem, but lots of people have written scripts that depend on it.
I've been using a dual G5 with Tiger for four years as my main development machine. It has never crashed. Not even once. I will be seriously depressed when Apple stops building its OS for PPC.
Then again, I avoid Adobe products. Probably has something to do with it.
Btw, I was an Amiga fan, and I think OS X is beautiful.
Eclipse's plugin architecture has some issues. I'm using NetBeans at the moment because I got really sick of manually adding "natures" to projects. Too many plugins' workflow requires creation of a project in the plugin's nature (e.g., why do all VCS plugins assume that I'm doing a Java project and not a CDT project?).
Eclipse is an example of something that is too extensible. Every once in a while, the "ecosystem" is in a good state and Eclipse is nice. But the rest of the time, the individual plugins are prescribing conflicting workflows.
I hope both projects survive... Interesting things are possible with Eclipse, but NetBeans is a lot more functional out of the box.
D2 and the beta index are working great for me, aside from the hiccup with the comments pane a couple of days ago.
I would like to see meta-moderation revisited, as this is the only way to mitigate coordinated group mods. Since the old mm system was dropped, I've seen an uptick in bizarre moderation.
I know my karma is going to take a hit for this, but I had to say it. Taco, you're doing a fine job. KDawson, I don't hate you. We understand why you have to do Idle. All in all, Slashdot is pretty great.
FWIW, I'm a bit of a pinko in this respect, and I support universal education.
Texas has already implemented what you're advocating. There are essentially no state requirements regarding home schooling in Texas, and TEA has no authority over private schools.
I'm left feeling that I want my state back, but if you're of a Libertarian bent, Texas is Utopia.
Actually, quantum gravity has been a fruitful line of inquiry.
Recent observations of super-distant, super-energetic gamma-ray bursts (specifically, GRB 080916C) have led to some validation of the quantum view.
Totally OT, but there you go.
Btw, I'm also a Texan. It's my understanding that only a few states in the U.S. order textbooks at the state level. You are missing something: Letting conservative think-tanks prescribe curriculum is an unambiguously bad thing; this isn't a law, it's action by the SBOE to amend the curriculum to include language written by the Discovery Institute.
Since Texas buys textbooks for the whole state, the TEKS curriculum has influence over what publishers put in their textbooks (nation-wide).
Your last sentence is odd. The statehouse is absolutely dominated by business-worshipping Republicans, and the judiciary is the same (unfortunately, imo, we elect our judges). We have no state income tax for individuals. If you wanted less government, you'd have to move to Somalia.
With respect, I disagree. People have a lot of definitions for REST, but I don't think yours is totally accurate. REST uses all of the HTTP methods. The workarounds for PUT and DELETE are because of specific servers not supporting them.
Session cookies do not break the REST model. Even when people use the url to pass state, it doesn't contain the state itself, just an id (plenty of room for that). And REST has the bonus of gracefully accommodating people like me (and probably you) who hack the url when the application's interface fails us. Similarly, REST apps make content easier to index.
What you and the originator of this thread are really proposing is making the web stateful and ditching the document metaphor. Lots of people will agree with me that this is not a good idea, and it's not because we're "blinded by 20 years of antipatterns". HTTP is not broken.
Also, I think you should examine what you're using GET for. I've never encountered the problem that you're using as a reason for scrapping HTTP.
Summing up, HTTP is just fine. If you like, you can create a protocol that looks like the one you describe. In fact, many have, but nothing has come close to the usefulness and resiliency of HTTP. So good luck, but hands off HTTP please.
COLLADA is pretty great, though it's pretty tough to generate. I always end up using Blender or SketchUp to get a rough idea of what the tree is going to look like before I write the code (seems to be more efficient than wading through the docs).
Also, COLLADA does a lot more than represent 3D models...
You probably already know this, but for those who don't: COLLADA is also the format used for models in Google Earth.
No offense, but if you're not using the POST method for actions with side-effects, you're doing it wrong.
Note that HTTP has all the CRUD methods and then some (though most servers are configured to make do with POST and GET). In any case, doing everything with GET is dangerous and wrong.
Imo, RESTful systems are the way to go, and the web should remain stateless. In other words, HTTP is great.
Wow. This is the most uninformed discussion I've seen in years.
This has nothing to do with markup (VRML, x3D, etc). It is about the browser exposing an opengl API to javascript.
According to TFA, it's Mozilla that is proposing building the functionality into a browser (but it could be any other ecmascript container). Since Mozilla is contributing to the standard (and started the initiative), we should expect this to absorb Canvas:3D (i.e., this is not a redundancy).
This standardization is a Good Thing. Everyone's concerns about advertisers abusing this should be directed towards advertisers, not this totally useful idea. Also, noscript rocks.
To all who think javascript just sucks, you don't know how to use it. Stop using it like it's C.
And if it replaces that leaky, closed, insecure, inaccessible, non-semantic and patent-encumbered plugin that almost killed the Web, awesome. With any luck, it'll kill DirectX too.
GP probably meant Social Darwinism, which many/. Libertarians (and LIRLs) apparently embrace (whether or not they accept the term).
Every time you see someone post "Capitalism is Evolution", this is what they're really talking about.
It's toxic stuff, and I wish more Libertarians knew that Saint Mises himself rejected it.
Spencer apologetics don't move me, btw. Whether he was a Lamarckian or Darwinist is totally irrelevant, Hofstadter's analysis wasn't a statist conspiracy to discredit Libertarians, and all of the above have little to do with the facts about the beliefs of most Libertarians.
I'd like to point out that the boom/bust cycles only really started after the Federal Reserve was created.
Demonstrably false. Central banks are absolutely essential in modern states. We have a euphemism for ours, but it exists and we need it.
Some fascinating history: J. P. Morgan was once known as the "Savior of the U.S." He used his own capital to rescue the U.S. Treasury twice, during the Panic of 1893 and the Panic of 1907. Now we have an accountable institution acting as lender of last resort, as opposed to depending on one wealthy (but possibly flawed) human being.
I think Friedman unwittingly started this meme, when he theorized that the Federal Reserve Act led to the specific crisis of the Great Depression (though most economists think he was wrong about even that). There is no evidence (and Friedman did not claim, afaik) that the "boom and bust" cycle (people used to call busts "panics") was smoother before the FRA.
Panic of 1819: Bubble following the War of 1812 collapses, exacerbated by a chaotic and corrupt banking system that emerged in the wake of the non-renewal of the charter for the First Bank in 1811 (keep in mind that during this interval, until the formation of the super-corrupt Second Bank in 1816, all the private banks in this ecosystem printed their own promisary notes, causing massive inflation as the government used this system to finance the war.)
The Long Depression: The original Great Depression. Especially interesting for apparently having been caused by tight monetary policies forced by the idealogical drive to return to the gold standard following the Treasury floating paper during the Civil War.
A case could be made that the Constitutional problems with a central bank simply represent a flaw in the Constitution. Trying to adhere to the Constitution in this regard has caused all kinds of problems. We seemed to have done okay with the nasty hack (the "Reserve" euphemism), though.
There is really no debate outside of the Mises Institute about the need for central banks.
I think the complaint is about the compilers, not the operating systems.
You could go further up the stack. Approaching it from the OS level is an attempt to "parallelize" code compiled to run in a serial manner, even if this code was written to take advantage of OS api locks/mutexes; approaching it from the compiler means finding opportunities to parallelize code written in a serial paradigm; approaching it at the application level means using a writing code in a different paradigm than we've all been using for half a century, but it fixes the problem.
Aside: When talking about locks, we should probably separate data and process concerns.
Speaking of... functional programming is finally getting some respect. Many said it was impossible to write a kernel in C; it would be neat if it were possible implement at least some of it in something like Concurrent Haskell. This is just idle speculation, but it seems like the performance hit of using a functional language near the bottom of the stack would be worth it in the long run, if it makes it possible to use many-core systems that much more efficiently. (Btw, there are purely functional research OS's out there, and there used to be a bunch of LISP machine OS's).
Huh. I wouldn't know, but according to the Reg, there is a maximum of 3 sessions. Funny if true.
The article is worth a read for other reasons. Hachamovitch is hilarious. It's terrifying to watch Microsoft's development chiefs warp technical concerns for the sake of marketing. Check out his response to the Reg's questions about SVG, or his revisionism of IE's purposefully standards-breaking past.
I don't know. Even if he just did nothing to stop Microsoft's resistance it would be bad.
If guys from CERT called me and said, "Hey, could you make The Autorun and NoDriveTypeAutorun registry values actually do something? We worried about this 10 million strong botnet," I'd probably comply. The reality was even worse; Microsoft wrote instructions for users to mitigate the problem which they knew were not effective.
The last thing I would do would be to start a PR war, which they did only to save face about something that has been criticized for over a decade. It's amazing... some slight marketing concern overrode what they were told was a matter of national security.
[...] just because this guy worked for Microsoft doesn't mean he lacks intelligence.
No, but it does mean that he was part of the team fighting US-CERT for months over autorun, at least. He likely helped resist an effort by a division of the department he is to head to fix a security problem that was so bad, they felt it endangered national security.
I wonder if we will be seeing US-CERT standing up to Microsoft the way they did with this (a vector for conficker) with him in charge.
I have a sick feeling about this. This guy was surely part of the Microsoft effort to call this a feature. And what was this "political infighting" that the article alludes to? I hope it wasn't over whether to go after Microsoft for aiding in the creation of the largest botnet to date.
Slashdot Mirror
ffs!
Javascript. API. OpenGL.
Man... I thought we covered this with the last story related to the proposed Khronos 3d api. This is nothing like VRML. It is a javascript api to use graphics hardware.
Nothing to do with markup of any kind (aside from the xml in Collada, which is not necessarily part of the standard). Ugh.
This history of javascript and the DOM is pretty complex and interesting. For example, the practice of wrapping native DOM objects has its roots in a cross-browser hack. IE's implementation of the DOM used COM, which meant that methods couldn't be added at runtime (maybe things got better with .net... I wouldn't know).
It turned out to be really useful, even aside from the cross-browser aspect. It allows jQuery to work with other libraries, and even multiple instances of itself. It gets around the problem of browsers deciding to implement new methods in the DOM, when your library had implemented the same feature earlier. Even Prototype has seen the light on this one... the assurance that your script will play nice with other unknown code is a killer feature.
With the iteration features, I'm guessing you're referring to stuff like each(). However, it's important to note that this is not a method added to a native DOM element; it's a method of a jQuery wrapper. And the practice of wrapping elements is now part of the architectural philosophy of things like jQuery (it's no longer just an IE workaround, it's useful on its own).
On the last paragraph in your comment... Yes. The death of ES4 was a good thing for this reason. We don't need classical inheritance and namespaces.
Also related to iteration, my #1 wish for implementations is optimization for tail recursion. The browsers currently only support 1000-5000 (I think) levels of recursion. After getting hooked on functional languages, using clunky procedural iteration hurts. There's tons of great functional code out there that would get ported to javascript overnight if browsers could manage the stack for recursion. Maybe Microsoft could get some of their Haskell guys to help out.
Well... JSON wasn't really "invented". It's really just the syntax for an object literal in javascript. JSON parsing is not much more than eval() and a regular expression.
It works beautifully.
Not going to touch the rest of your comment. JSON is more useful in this context than xml.
Btw, Crockford's presentations are really entertaining: http://www.youtube.com/watch?v=hQVTIJBZook
I think you're confusing the DOM api and javascript. jQuery makes the DOM api what it should have been all along.
Or they should be. Unfortunately, there are a lot of "how-to"s on the web that use a null key passphrase to make scripting easier. That openssh allows this is a problem, but lots of people have written scripts that depend on it.
Anecdote, data, blah blah...
I've been using a dual G5 with Tiger for four years as my main development machine. It has never crashed. Not even once. I will be seriously depressed when Apple stops building its OS for PPC.
Then again, I avoid Adobe products. Probably has something to do with it.
Btw, I was an Amiga fan, and I think OS X is beautiful.
Eclipse's plugin architecture has some issues. I'm using NetBeans at the moment because I got really sick of manually adding "natures" to projects. Too many plugins' workflow requires creation of a project in the plugin's nature (e.g., why do all VCS plugins assume that I'm doing a Java project and not a CDT project?).
Eclipse is an example of something that is too extensible. Every once in a while, the "ecosystem" is in a good state and Eclipse is nice. But the rest of the time, the individual plugins are prescribing conflicting workflows.
I hope both projects survive... Interesting things are possible with Eclipse, but NetBeans is a lot more functional out of the box.
D2 and the beta index are working great for me, aside from the hiccup with the comments pane a couple of days ago.
I would like to see meta-moderation revisited, as this is the only way to mitigate coordinated group mods. Since the old mm system was dropped, I've seen an uptick in bizarre moderation.
I know my karma is going to take a hit for this, but I had to say it. Taco, you're doing a fine job. KDawson, I don't hate you. We understand why you have to do Idle. All in all, Slashdot is pretty great.
FWIW, I'm a bit of a pinko in this respect, and I support universal education.
Texas has already implemented what you're advocating. There are essentially no state requirements regarding home schooling in Texas, and TEA has no authority over private schools.
I'm left feeling that I want my state back, but if you're of a Libertarian bent, Texas is Utopia.
Actually, quantum gravity has been a fruitful line of inquiry.
Recent observations of super-distant, super-energetic gamma-ray bursts (specifically, GRB 080916C) have led to some validation of the quantum view.
Totally OT, but there you go.
Btw, I'm also a Texan. It's my understanding that only a few states in the U.S. order textbooks at the state level. You are missing something: Letting conservative think-tanks prescribe curriculum is an unambiguously bad thing; this isn't a law, it's action by the SBOE to amend the curriculum to include language written by the Discovery Institute.
SBOE TEKS (Texas Essential Knowledge and Skills)
Since Texas buys textbooks for the whole state, the TEKS curriculum has influence over what publishers put in their textbooks (nation-wide).
Your last sentence is odd. The statehouse is absolutely dominated by business-worshipping Republicans, and the judiciary is the same (unfortunately, imo, we elect our judges). We have no state income tax for individuals. If you wanted less government, you'd have to move to Somalia.
I've got karma to burn.
Microsoft:ISO::United States:UN.
With respect, I disagree. People have a lot of definitions for REST, but I don't think yours is totally accurate. REST uses all of the HTTP methods. The workarounds for PUT and DELETE are because of specific servers not supporting them.
Session cookies do not break the REST model. Even when people use the url to pass state, it doesn't contain the state itself, just an id (plenty of room for that). And REST has the bonus of gracefully accommodating people like me (and probably you) who hack the url when the application's interface fails us. Similarly, REST apps make content easier to index.
What you and the originator of this thread are really proposing is making the web stateful and ditching the document metaphor. Lots of people will agree with me that this is not a good idea, and it's not because we're "blinded by 20 years of antipatterns". HTTP is not broken.
Also, I think you should examine what you're using GET for. I've never encountered the problem that you're using as a reason for scrapping HTTP.
Summing up, HTTP is just fine. If you like, you can create a protocol that looks like the one you describe. In fact, many have, but nothing has come close to the usefulness and resiliency of HTTP. So good luck, but hands off HTTP please.
COLLADA is pretty great, though it's pretty tough to generate. I always end up using Blender or SketchUp to get a rough idea of what the tree is going to look like before I write the code (seems to be more efficient than wading through the docs).
Also, COLLADA does a lot more than represent 3D models...
You probably already know this, but for those who don't: COLLADA is also the format used for models in Google Earth.
Read.
No offense, but if you're not using the POST method for actions with side-effects, you're doing it wrong.
Note that HTTP has all the CRUD methods and then some (though most servers are configured to make do with POST and GET). In any case, doing everything with GET is dangerous and wrong.
Imo, RESTful systems are the way to go, and the web should remain stateless. In other words, HTTP is great.
Wow. This is the most uninformed discussion I've seen in years.
This has nothing to do with markup (VRML, x3D, etc). It is about the browser exposing an opengl API to javascript.
According to TFA, it's Mozilla that is proposing building the functionality into a browser (but it could be any other ecmascript container). Since Mozilla is contributing to the standard (and started the initiative), we should expect this to absorb Canvas:3D (i.e., this is not a redundancy).
This standardization is a Good Thing. Everyone's concerns about advertisers abusing this should be directed towards advertisers, not this totally useful idea. Also, noscript rocks.
To all who think javascript just sucks, you don't know how to use it. Stop using it like it's C.
And if it replaces that leaky, closed, insecure, inaccessible, non-semantic and patent-encumbered plugin that almost killed the Web, awesome. With any luck, it'll kill DirectX too.
[...] not least of all because it actually has fairly original gameplay (it's not a clone of some other game)
I like Wesnoth, and have wasted weeks at a time with it, but it was based on Master of Monsters.
Also, check this out: http://www.wesnoth.org/wiki/WesnothPhilosophy
Spencer, Herbert "Social Statics", c. XXV, "Poor-Laws"
GP probably meant Social Darwinism, which many /. Libertarians (and LIRLs) apparently embrace (whether or not they accept the term).
Every time you see someone post "Capitalism is Evolution", this is what they're really talking about.
It's toxic stuff, and I wish more Libertarians knew that Saint Mises himself rejected it.
Spencer apologetics don't move me, btw. Whether he was a Lamarckian or Darwinist is totally irrelevant, Hofstadter's analysis wasn't a statist conspiracy to discredit Libertarians, and all of the above have little to do with the facts about the beliefs of most Libertarians.
I'd like to point out that the boom/bust cycles only really started after the Federal Reserve was created.
Demonstrably false. Central banks are absolutely essential in modern states. We have a euphemism for ours, but it exists and we need it.
Some fascinating history: J. P. Morgan was once known as the "Savior of the U.S." He used his own capital to rescue the U.S. Treasury twice, during the Panic of 1893 and the Panic of 1907. Now we have an accountable institution acting as lender of last resort, as opposed to depending on one wealthy (but possibly flawed) human being.
I think Friedman unwittingly started this meme, when he theorized that the Federal Reserve Act led to the specific crisis of the Great Depression (though most economists think he was wrong about even that). There is no evidence (and Friedman did not claim, afaik) that the "boom and bust" cycle (people used to call busts "panics") was smoother before the FRA.
More financial crisis fun:
More...
A case could be made that the Constitutional problems with a central bank simply represent a flaw in the Constitution. Trying to adhere to the Constitution in this regard has caused all kinds of problems. We seemed to have done okay with the nasty hack (the "Reserve" euphemism), though.
There is really no debate outside of the Mises Institute about the need for central banks.
I think the complaint is about the compilers, not the operating systems.
You could go further up the stack. Approaching it from the OS level is an attempt to "parallelize" code compiled to run in a serial manner, even if this code was written to take advantage of OS api locks/mutexes; approaching it from the compiler means finding opportunities to parallelize code written in a serial paradigm; approaching it at the application level means using a writing code in a different paradigm than we've all been using for half a century, but it fixes the problem.
Aside: When talking about locks, we should probably separate data and process concerns.
Speaking of... functional programming is finally getting some respect. Many said it was impossible to write a kernel in C; it would be neat if it were possible implement at least some of it in something like Concurrent Haskell. This is just idle speculation, but it seems like the performance hit of using a functional language near the bottom of the stack would be worth it in the long run, if it makes it possible to use many-core systems that much more efficiently. (Btw, there are purely functional research OS's out there, and there used to be a bunch of LISP machine OS's).
Huh. I wouldn't know, but according to the Reg, there is a maximum of 3 sessions. Funny if true.
The article is worth a read for other reasons. Hachamovitch is hilarious. It's terrifying to watch Microsoft's development chiefs warp technical concerns for the sake of marketing. Check out his response to the Reg's questions about SVG, or his revisionism of IE's purposefully standards-breaking past.
I don't know. Even if he just did nothing to stop Microsoft's resistance it would be bad.
If guys from CERT called me and said, "Hey, could you make The Autorun and NoDriveTypeAutorun registry values actually do something? We worried about this 10 million strong botnet," I'd probably comply. The reality was even worse; Microsoft wrote instructions for users to mitigate the problem which they knew were not effective.
The last thing I would do would be to start a PR war, which they did only to save face about something that has been criticized for over a decade. It's amazing... some slight marketing concern overrode what they were told was a matter of national security.
Funny... the wikipedia page on autorun was just stealth edited to remove all mention of the problem.
RTFA. I'm talking about about CERT and conficker.
FWIW, I don't like the very idea of DHS. Also, fuck you.
[...] just because this guy worked for Microsoft doesn't mean he lacks intelligence.
No, but it does mean that he was part of the team fighting US-CERT for months over autorun, at least. He likely helped resist an effort by a division of the department he is to head to fix a security problem that was so bad, they felt it endangered national security.
I wonder if we will be seeing US-CERT standing up to Microsoft the way they did with this (a vector for conficker) with him in charge.
I have a sick feeling about this. This guy was surely part of the Microsoft effort to call this a feature. And what was this "political infighting" that the article alludes to? I hope it wasn't over whether to go after Microsoft for aiding in the creation of the largest botnet to date.