First impressions: really awful. I find it different, but really interesting. Much better that any other GUI used only with keyboard. *Those* are awful.
# Forward and back arrows do what you expect. Up and down scroll the screen. Page up and down do nothing. I wrote to Brad Lauster just two days ago on this very point, and he was quite open to commentary. Main movement is supposed to be done through text search, but maybe page up/down will be implemented. Remember that this is in very alpha stage.
# The mouse, of course, does nothing at all. The mouse will be used for graphic manipulation and navigation in the zooming interface. For text processing, it uses the keyboard. This is the main point that makes habit forming possible.
# Keys you expect to repeat don't. That triple-tap thing holds firm for everything. Even backspace. Even the arrow keys. All keys do repeat, thank you very much. You have to tap it three times, not press and wait until it repeats. This avoids the error of having autorepeat when you don't want it, and is faster when you do.
# Tildes and backticks are impossible to type, they've become control characters. This program is a prototype. The final product is suposed to run under dedicated hardware (a special keyboard). Afaik- there is a USB "LeapBar" extension for current keyboards.
# The cursor blinks frantically and distractingly in not one, but two colors. Maybe you could file a usability bug report? This project is being user-tested. If it is annoying, it will be removed.
# To access help, you have to hold down capslock while you type. The LeapBar will have a dedicated Command key. With the zooming interface you could simply "zoom out" until you see the Help section. And yes, the textual interface is designed for good typists - that's not strange, since it's supposed for advanced text manipulation.
I stopped there. Guess it needs a little more time in the oven, but so far it's flying in the face of usability. No, its only strange because you have to learn it from scratch. Is not more difficult than a current WIMP "point-and-click" interface is for novice users. Actually it's easier - more consistent, more simple, you don't have to think what you want to do in advance.
It's the same as the Unix command line. A single interface, where all available commands interact at user's whish through a unified syntax. But in Archy the syntax is more powerful than in Bash.
Ok, i should have said "it only forbids *mixing code* between them for distribution purposes".
The GPL doesn't disallow code mixing for your own use, and that's why I don't think this is similar to Microsoft's "you may only use other Microsoft-approved products with this software". Furthermore, MS doesn't allow "code mixing for distribution purposes" at all with their code.
you can't cause a version already released under the GPL to become retroactively proprietary
Maybe the poster meant that, but it doesn't make it true. If you still own the copyright of the code, you can do whatever you want with it. Release it, relicense it, put it in a box and throw away the key.
It doesn't matter what you did in the past with the code, if it's yours then you can put it under a different license or several licenses at once, even if it's a previous version.
Current AI doesn't show creativity at all. In theorem proving, 90% of the required creativity is already present in the stated goal (the theorem), the other 10% is the heuristics programmed to guide the automatic search.
Overview SPIRAL addresses this problem for the performance-critical class of DSP transform algorithms, like the discrete Fourier transform (DFT), the discrete cosine transforms (DCTs), and many others, by automatically generating code that is tuned to the given computing platform.
SPIRAL uses mathematics to describe, generate, and manipulate fast algorithms, translates them into programs, and searches for the best one for a given platform.
The architecture of SPIRAL is displayed in the figure to the left. The user specifies the transform she wants to implement and its size, e.g., a DFT (discrete Fourier transform) of size 1024. First, the Algorithm Generation block generates one, or several, out of many possible fast algorithms for the transform, represented as a formula in the SPIRAL proprietary language SPL (signal processing language). Next the formula is optimized at the mathematical level using a set of formal mathematical manipulation rules. The optimized SPL formula is compiled by the Implementation block into a program in a common language such as C or Fortran and then optimized using common and not so common compiler techniques. Directives at the Implementation level control implementation choices such as the degree of unrolling or scheduling. The final program is compiled and its runtime measured. Based on this runtime, the Search/Learning block triggers the generation of additional algorithms and their implementations using possibly different directives. Iteration of this process leads to a C or Fortran implementation that is adapted to the given computing platform. Optionally, the generated code is verified for correctness.
In summary, SPIRAL automatically searches in the space of structurally different algorithms and the space of their implementations for the best match to a given computing platform. Further, many optimizations are performed at a high level of program representation, thus overcoming compiler limitations.
That's one of the most pervasive design errors in today computers. Really, a good computer design should trear user input as sacred - because everything else can be recomputed, but user data is unique and precious.
Come on guys. We have transactional databases, we have huge space in hard disks, we have no reason to lose a single keypress from the user. Do we enjoy having jokes on how people react when all their work of five hours is lost forever? Is "press the Save Button often" the best solution we can engineer?
You got it wrong. Capitalism is based on private property. The problem is that when you try to turn information into a capitalist product, you have a paradox: information should be free to have a free market (one with perfect competition), but then how can you make a profit by selling it?
But property requires government protection. In information market, private property (be it over closed source software, music or books) does distort the market. So free content != capitalism.
Autopackage is still useful por people who *know* how to use software that is not packaged in the distro. Didnt you read the bit in the FAQ about the writer making their own packages? I would love to have a program doing all the dependency checks for me when I want to install Beagle, or X.org, or any other beta software that is not yet packaged.
Slashdot Mirror
So, uh, you recreated Emacs?
Yes, but modeless.
Yes, just like if it were a command line...
It's the same concept of Emacs. Must be good, since Emacs is considered by programmers as one of the best development environments.
The main benefit of Archy over Emacs is that it has been engineered with ease of use in mind, not just ease to extend.
First impressions: really awful.
I find it different, but really interesting. Much better that any other GUI used only with keyboard. *Those* are awful.
# Forward and back arrows do what you expect. Up and down scroll the screen. Page up and down do nothing.
I wrote to Brad Lauster just two days ago on this very point, and he was quite open to commentary. Main movement is supposed to be done through text search, but maybe page up/down will be implemented. Remember that this is in very alpha stage.
# The mouse, of course, does nothing at all.
The mouse will be used for graphic manipulation and navigation in the zooming interface. For text processing, it uses the keyboard. This is the main point that makes habit forming possible.
# Keys you expect to repeat don't. That triple-tap thing holds firm for everything. Even backspace. Even the arrow keys.
All keys do repeat, thank you very much. You have to tap it three times, not press and wait until it repeats. This avoids the error of having autorepeat when you don't want it, and is faster when you do.
# Tildes and backticks are impossible to type, they've become control characters.
This program is a prototype. The final product is suposed to run under dedicated hardware (a special keyboard). Afaik- there is a USB "LeapBar" extension for current keyboards.
# The cursor blinks frantically and distractingly in not one, but two colors.
Maybe you could file a usability bug report? This project is being user-tested. If it is annoying, it will be removed.
# To access help, you have to hold down capslock while you type.
The LeapBar will have a dedicated Command key.
With the zooming interface you could simply "zoom out" until you see the Help section.
And yes, the textual interface is designed for good typists - that's not strange, since it's supposed for advanced text manipulation.
I stopped there. Guess it needs a little more time in the oven, but so far it's flying in the face of usability.
No, its only strange because you have to learn it from scratch. Is not more difficult than a current WIMP "point-and-click" interface is for novice users. Actually it's easier - more consistent, more simple, you don't have to think what you want to do in advance.
Not a problem at all, since Archy doesn't have "files".
It's the same as the Unix command line. A single interface, where all available commands interact at user's whish through a unified syntax. But in Archy the syntax is more powerful than in Bash.
There are humans that show creativity, but there aren't computers that show creativity. Not yet.
So instances of "computer creativity" like theorem proving are really hidden instances of a creative human.
Ok, i should have said "it only forbids *mixing code* between them for distribution purposes".
The GPL doesn't disallow code mixing for your own use, and that's why I don't think this is similar to Microsoft's "you may only use other Microsoft-approved products with this software". Furthermore, MS doesn't allow "code mixing for distribution purposes" at all with their code.
Oh! You mean that you can't take away the GPL! Ok, that's true. But you can make a proprietary auto-fork.
you can't cause a version already released under the GPL to become retroactively proprietary
Maybe the poster meant that, but it doesn't make it true. If you still own the copyright of the code, you can do whatever you want with it. Release it, relicense it, put it in a box and throw away the key.
It doesn't matter what you did in the past with the code, if it's yours then you can put it under a different license or several licenses at once, even if it's a previous version.
Are you mad? The GPL doesn't disallow to *use* programs that are incompatible with the license, it only forbids *mixing code* between them.
Write all your open code under BSD license, sir, and in 10 years there will not be any commercial open source code at all.
This is what happened to Unix, and the reason why Stallman designed the GPL.
Current AI doesn't show creativity at all. In theorem proving, 90% of the required creativity is already present in the stated goal (the theorem), the other 10% is the heuristics programmed to guide the automatic search.
Overview
SPIRAL addresses this problem for the performance-critical class of DSP transform algorithms, like the discrete Fourier transform (DFT), the discrete cosine transforms (DCTs), and many others, by automatically generating code that is tuned to the given computing platform.
SPIRAL uses mathematics to describe, generate, and manipulate fast algorithms, translates them into programs, and searches for the best one for a given platform.
The architecture of SPIRAL is displayed in the figure to the left. The user specifies the transform she wants to implement and its size, e.g., a DFT (discrete Fourier transform) of size 1024. First, the Algorithm Generation block generates one, or several, out of many possible fast algorithms for the transform, represented as a formula in the SPIRAL proprietary language SPL (signal processing language). Next the formula is optimized at the mathematical level using a set of formal mathematical manipulation rules. The optimized SPL formula is compiled by the Implementation block into a program in a common language such as C or Fortran and then optimized using common and not so common compiler techniques. Directives at the Implementation level control implementation choices such as the degree of unrolling or scheduling. The final program is compiled and its runtime measured. Based on this runtime, the Search/Learning block triggers the generation of additional algorithms and their implementations using possibly different directives. Iteration of this process leads to a C or Fortran implementation that is adapted to the given computing platform. Optionally, the generated code is verified for correctness.
In summary, SPIRAL automatically searches in the space of structurally different algorithms and the space of their implementations for the best match to a given computing platform. Further, many optimizations are performed at a high level of program representation, thus overcoming compiler limitations.
It is not even funny. Almost. Well, a little bit.
It's a well known fact that new technologies don't catch up until they can be used for pr0n.
Why do you need special hardware to achieve this? It should be done by the OS.
That's one of the most pervasive design errors in today computers. Really, a good computer design should trear user input as sacred - because everything else can be recomputed, but user data is unique and precious.
Come on guys. We have transactional databases, we have huge space in hard disks, we have no reason to lose a single keypress from the user. Do we enjoy having jokes on how people react when all their work of five hours is lost forever? Is "press the Save Button often" the best solution we can engineer?
I, for one, welcome our new teenage high-school robot-making MIT-defeating overlords.
You got it wrong. Capitalism is based on private property. The problem is that when you try to turn information into a capitalist product, you have a paradox: information should be free to have a free market (one with perfect competition), but then how can you make a profit by selling it?
But property requires government protection. In information market, private property (be it over closed source software, music or books) does distort the market. So free content != capitalism.
Quite ironic that your link to kindergarten art states: In Kindergarten art class, we explore various two and three diminsional media .
Picasso worked with the 4th dimension, like Einstein did. Tell that to your average 5 year old kindergarden children.
Does ./configure also download and install all the recursively unmet dependencies?
I thought so.
Autopackage is still useful por people who *know* how to use software that is not packaged in the distro. Didnt you read the bit in the FAQ about the writer making their own packages? I would love to have a program doing all the dependency checks for me when I want to install Beagle, or X.org, or any other beta software that is not yet packaged.
And people stopped taking physics seriously after crap like Einstein, too...
nuff said.