If memory serves, double clicking the ribbon title where the menu items used to be collapses the ribbon.
I like some of the ribbon (context sensitive tool bar part), but don't like the difficulty to find infrequently used things (eg, setting print area in excel).
I wish they just did a context sensitive tool bar and a traditional menu.
The big problem with the ribbon is that it's not quite as context-sensitive as it claims to be. The section of the ribbon that's open is often simply the last one you used, unless it's no longer applicable to the context you're in. If it's a global option (home, page layout or similar) then it's available in all contexts. There were times I'd been working away in Word or Excel for over quarter of an hour without need of the ribbon (using keyboard shortcuts and the context-menu key for all my edits), and then when I needed to go to the ribbon for something, my flow would be broken by the fact that I had to re-orient my mental context to the currently open ribbon. I couldn't work on auto-pilot like I used to. Context-sensitivity should mean that my first available action is defined by the context, but with ribbons, that ain't so.
Things like progress bars actually give you useful information.
To power users, maybe. To average users, a progress bar is something that keeps stopping, making them think that the computer has crashed. Horses for courses -- if you have a task that you can use a progress bar for (ie the task has predictable time) then use it. If it doesn't, don't.
Command line image editing is kind of a pain, but the best for server admin and other tasks for sure.
Not if you want to make global changes to the image. If you're correcting contrast etc, a command line is often much quicker than faffing about with buttons and sliders.
This shows a total failure to understand the point of flat design -- one of the arguments against skeuomorphism is that it often results in low contrast, slowing down mental recognition of icons. Flat design is supposed to mean high contrast. If you have low contrast in your flat UI, you're doing it wrong.
Oh - this - so much this! The backlash against skeuomorphics has produced some of the most boring and bland and ucking fugly interfaces seen by humans.
It's like someone said "Good design is a square of a primary color, with a letter in it." To me, the problem is across the board - UI's are getting ugly, like we are returning to Commodore 64 days when there just wasn't enough resolution to make nice looking stuff.
There has been an overreaction, certainly, but that's due to people not listening to the complaints about skeuomorphism.
The problem with skeuomorphic design is that it ignores the basic principles of what an "icon" is, and how the human brain works. The point is that the gradings and shadings and pseudo-3D projections on Windows XP icons were slower to process cognitively than their Windows 95 equivalents. The trick to a good icon is to find the simplest unambiguous "form" that prototypes the concept to the human brain. A good example of flat design is the icon for iBooks on iOS -- it's a minimal unambiguous representation of a book.
A contrasting bad example would be the icon for Facetime. It's not instantly recognisable as a video camera, and even if it was, "video camera" is not a synonym of "video call".
I specifically used two examples from iOS because this isn't about my views on Apple (the iPad was given to me, my laptop runs Windows and Linux), but just about the principles. Flat design is not a panacea -- it is just one principle of many that make up UI/UX design. Flat design on its own is useless -- flat design is supposed to make it easier to apply the other principles.
The whole force powers in computer-game-world (and in the prequels) really piss me off. If people were running at Mach 3 in living memory, why is that idiot in episode IV mocking Darth Vader for his Jedi skills? The force in the original trilogy is something far more subtle and far more peaceful. It was about awareness and mindfulness, with a little bit of influence and telekinesis thrown in. The only "powerful" stuff was the Emperor's lightning attack, and the good guys didn't even have that. It was almost like saying that good guys only have tools -- weapons are for the bad guys.
I suspect a lot of Star Wars fan behaviour is reactionaryism -- they see the Trekkies wandering around with Spock-ears and want to show that Star Wars is better. Twenty years ago, the proof that Star Wars was better was that the fans didn't get involved in that sort of nonsense, and just watched the films....
Yea, I'm a fan of the original trilogy and it was sad to see George Lucas turn in to his own characters (young Lucas = Anakin, old Lucas = Vader).
Anakin was a whiny wee brat who grew into a petulant unpleasant youth with serious attachment issues who hated anyone who didn't massage his ego. Vader was a galactic-scale badass with an iron strength of will who had the wherewithal to question his own mistakes and set his own course in life.
Timothy Zahn is a cheap hack who writes dimestore pulp. His Star Wars novels wavered between slavish consistency with the films (meaning no character development) and utter disregard for the films (creatures that exist outwith the energy that binds the entire universe together...?) leaving a sense of sheer mediocrity.
Back in the day, when Lucas was just a filmmaker, Star Wars was conceived literally as "a cowboy western in space".
It was SUPPOSED to be action-packed and a little cheesy, with hammy 2d archetypes for characters...
The way that this subsequently has ended up hallowed in some peoples' minds (including Lucas, who never has apparently missed a step on his own hagiography) does a disservice to what it was really intended to be.
Yes and no.
The original drew heavily on Kurosawa's Hidden Fortress, and so bolstered up the lightweight 2d archetypes with "twilight of the samurai", creating a very strong effect: "oh, weren't things simpler in the old days, and we can go back to that". I don't really credit Lucas for constructing that -- I imagine a lot of it was more accident (and plagiarism) than design.
However, the real enduring power of the trilogy came out of what Lawrence Kasdan did -- he managed to maintain the shallow good-and-evil fairy-story, combine it with Lucas's slightly pretentious ideas about Skywalker's internal conflict and build in a scale that more than made up for a lack of depth. It was a simple story that was well paced and well told, and there is really nothing wrong with that.
On the other hand, keeping individual device speeds done helps prevent users killing the network. I know, traffic management can be done digitally, but having 10/100 instead of gigabit on the device is basically fool-proof, and if you're running an office of 200 non-technical users on a single network connection, there will be a significant minority of fools in there that think it's OK to download stupidly large files in the middle of the working day.
I hope they've verified that nothing bad happens when you stack the speakers module directly above or below the HDD module and then play music at top volume for a few hours:)
And vice versa: I wonder if the noise-reducing microphone is specifically tuned to reject the vibration noise of the rapidly spinning plates (which frequently spin up/down, don't forget) that it will be physically coupled to.
This is why the broader sphere of AI (and not just the classical AI) has to encompass multiple disciplines. If you were to create an academic sphere called "intelligence", it would take in neuroscience, neuropsychology, psychology and arguably even sociology. Every part of AI tries to model a particular level of abstraction of intelligent behaviour, and a single level of abstraction will never be enough
Why does "AI" have to be dominated by "classical AI"? Classical AI didn't even attempt to model intelligence -- it was pure behaviourism at heart -- so I don't really know if it can justify being called AI at all.
In that sense it's analogous to sigma-style mahematic operations, and that sort of implied iteration is perfectly acceptable in the declarative paradigm. As such, we don't need recursion for all of our list processing functions
In fact, this sigma-style iteration where there are no interdependancies between iterations is crucial to parallelisation, so traditional imperative iterative loops are a major barrier to progress.
One of my favourite tings at the moment is Python's list comprehensions, which I feel sit on a weird boundary between declarative and imperative. It's declarative because you don't specify how the iteration works but it's arguably imperative because everything except the iteration is explicit. But I think this is enough to make it declarative. I'm not talking it through the loop, and then iteration could be evaluated 0 upwards, -1 downwards or in any random order you want -- it's just not relevant at the programmer's level.
In that sense it's analogous to sigma-style mahematic operations, and that sort of implied iteration is perfectly acceptable in the declarative paradigm. As such, we don't need recursion for all of our list processing functions
Isn't that the problem we have in teaching programming? our refusal as programmers to accept declarative programming as "real" programming? The "programming" (i.e. scheduling and process planning),of past centuries was declarative, not imperative, and that's much easier to get our heads around as human beings.
"But... optimisation!" I hear you cry. How many of us really need to develop high-performance systems? I had a task that I spent months trying to code up imperatively, and when I said "stuff that" and fired up Prolog, I got my results in weeks instead of months. I'm happyrunning it as a batch job for now and optimising the code (i.e. rewriting in another language) later if I need to. For now, it means I can focus on other parts of the system and get the whole thing to full proof-of-concept rather than getting stuck on one part of it.
Besides, most of the heavy lifting in a lot of coding tasks is done in libraries, and from the programmer's point of view, that's effectively declarative programming -- I have no control of how SDL draws a circle on screen. Yes, I can start looking at how different libraries' internal code works and whatnot, but going down that road means hitting the conclusion that libraries are *worse* than declarative code, because declarative coding means that (in theory) one line can be evaluated whichever way is most efficient whereas a library call encodes a specific operational choice.
What we need is what I call "declarative-first programming" -- a language/environment that lets you code up a declarative algorithm to build your first working prototype then progressively optimise it section-by-section using more specific code until all the sections where efficiency matters are procedural, and all the lightweight (not processor/memory intensive) stuff is declarative.
I don't really buy the types-of-teacher/learner thing. I believe their is a near-optimal teaching strategy for every subject, and that you only need to tinker round the edges to incorporate learners' prior knowledge and individual aims. I reckon any apparent differences in learners beyond that are best explained as "tolerances" -- most learners can cope with a limited amount of inefficiency in teaching, and those tolerances may be individual. I aim to be a better teacher with every lesson I teach. If I was to say "learner differences" I would be professing a belief that you can only ever successfully teach a small part of your class at any one time. It would make my job easier, but it sounds a little dog-ate-my-homework to me.
If you equate programming with standard maths, you're missing the importance of mutability. Maths is semantic and imperative, procedural programming is declarative. Mathematical variables are an entirely different beast to computer variables.
I think the next step after this would be an open index where people can just catalogue papers and individual observations and how they support/refute one another. And goodbye selective citation.
As I've commented further up, I think the real value here is counterexamples: if you have a paper that's built on selective evidence, you don't need to prove an alternative theory, but rather just find sufficient counterexamples to demonstrate that the paper is unreliable.
Slashdot Mirror
"The problem is that UX/UI people like to invent new and exciting stuff, while they should be making stuff familiar and boring."
Which is completely unlike non-UI/UX open source programming, right?
Yes. A heck of a lot of open-source software is just functional clones of commercial packages.
If memory serves, double clicking the ribbon title where the menu items used to be collapses the ribbon.
I like some of the ribbon (context sensitive tool bar part), but don't like the difficulty to find infrequently used things (eg, setting print area in excel).
I wish they just did a context sensitive tool bar and a traditional menu.
The big problem with the ribbon is that it's not quite as context-sensitive as it claims to be. The section of the ribbon that's open is often simply the last one you used, unless it's no longer applicable to the context you're in. If it's a global option (home, page layout or similar) then it's available in all contexts. There were times I'd been working away in Word or Excel for over quarter of an hour without need of the ribbon (using keyboard shortcuts and the context-menu key for all my edits), and then when I needed to go to the ribbon for something, my flow would be broken by the fact that I had to re-orient my mental context to the currently open ribbon. I couldn't work on auto-pilot like I used to. Context-sensitivity should mean that my first available action is defined by the context, but with ribbons, that ain't so.
Things like progress bars actually give you useful information.
To power users, maybe. To average users, a progress bar is something that keeps stopping, making them think that the computer has crashed. Horses for courses -- if you have a task that you can use a progress bar for (ie the task has predictable time) then use it. If it doesn't, don't.
Command line image editing is kind of a pain, but the best for server admin and other tasks for sure.
Not if you want to make global changes to the image. If you're correcting contrast etc, a command line is often much quicker than faffing about with buttons and sliders.
Flat widgets, low washed out contrast,
This shows a total failure to understand the point of flat design -- one of the arguments against skeuomorphism is that it often results in low contrast, slowing down mental recognition of icons. Flat design is supposed to mean high contrast. If you have low contrast in your flat UI, you're doing it wrong.
Oh - this - so much this! The backlash against skeuomorphics has produced some of the most boring and bland and ucking fugly interfaces seen by humans.
It's like someone said "Good design is a square of a primary color, with a letter in it." To me, the problem is across the board - UI's are getting ugly, like we are returning to Commodore 64 days when there just wasn't enough resolution to make nice looking stuff.
There has been an overreaction, certainly, but that's due to people not listening to the complaints about skeuomorphism.
The problem with skeuomorphic design is that it ignores the basic principles of what an "icon" is, and how the human brain works. The point is that the gradings and shadings and pseudo-3D projections on Windows XP icons were slower to process cognitively than their Windows 95 equivalents. The trick to a good icon is to find the simplest unambiguous "form" that prototypes the concept to the human brain. A good example of flat design is the icon for iBooks on iOS -- it's a minimal unambiguous representation of a book.
A contrasting bad example would be the icon for Facetime. It's not instantly recognisable as a video camera, and even if it was, "video camera" is not a synonym of "video call".
I specifically used two examples from iOS because this isn't about my views on Apple (the iPad was given to me, my laptop runs Windows and Linux), but just about the principles. Flat design is not a panacea -- it is just one principle of many that make up UI/UX design. Flat design on its own is useless -- flat design is supposed to make it easier to apply the other principles.
The whole force powers in computer-game-world (and in the prequels) really piss me off. If people were running at Mach 3 in living memory, why is that idiot in episode IV mocking Darth Vader for his Jedi skills? The force in the original trilogy is something far more subtle and far more peaceful. It was about awareness and mindfulness, with a little bit of influence and telekinesis thrown in. The only "powerful" stuff was the Emperor's lightning attack, and the good guys didn't even have that. It was almost like saying that good guys only have tools -- weapons are for the bad guys.
I suspect a lot of Star Wars fan behaviour is reactionaryism -- they see the Trekkies wandering around with Spock-ears and want to show that Star Wars is better. Twenty years ago, the proof that Star Wars was better was that the fans didn't get involved in that sort of nonsense, and just watched the films....
Yea, I'm a fan of the original trilogy and it was sad to see George Lucas turn in to his own characters (young Lucas = Anakin, old Lucas = Vader).
Anakin was a whiny wee brat who grew into a petulant unpleasant youth with serious attachment issues who hated anyone who didn't massage his ego. Vader was a galactic-scale badass with an iron strength of will who had the wherewithal to question his own mistakes and set his own course in life.
Timothy Zahn is a cheap hack who writes dimestore pulp. His Star Wars novels wavered between slavish consistency with the films (meaning no character development) and utter disregard for the films (creatures that exist outwith the energy that binds the entire universe together...?) leaving a sense of sheer mediocrity.
Back in the day, when Lucas was just a filmmaker, Star Wars was conceived literally as "a cowboy western in space".
It was SUPPOSED to be action-packed and a little cheesy, with hammy 2d archetypes for characters...
The way that this subsequently has ended up hallowed in some peoples' minds (including Lucas, who never has apparently missed a step on his own hagiography) does a disservice to what it was really intended to be.
Yes and no.
The original drew heavily on Kurosawa's Hidden Fortress, and so bolstered up the lightweight 2d archetypes with "twilight of the samurai", creating a very strong effect: "oh, weren't things simpler in the old days, and we can go back to that". I don't really credit Lucas for constructing that -- I imagine a lot of it was more accident (and plagiarism) than design.
However, the real enduring power of the trilogy came out of what Lawrence Kasdan did -- he managed to maintain the shallow good-and-evil fairy-story, combine it with Lucas's slightly pretentious ideas about Skywalker's internal conflict and build in a scale that more than made up for a lack of depth. It was a simple story that was well paced and well told, and there is really nothing wrong with that.
Then you have the wrong equipment and your buyers have sold you short.
On the other hand, keeping individual device speeds done helps prevent users killing the network. I know, traffic management can be done digitally, but having 10/100 instead of gigabit on the device is basically fool-proof, and if you're running an office of 200 non-technical users on a single network connection, there will be a significant minority of fools in there that think it's OK to download stupidly large files in the middle of the working day.
I hope they've verified that nothing bad happens when you stack the speakers module directly above or below the HDD module and then play music at top volume for a few hours :)
And vice versa: I wonder if the noise-reducing microphone is specifically tuned to reject the vibration noise of the rapidly spinning plates (which frequently spin up/down, don't forget) that it will be physically coupled to.
I for one welcome our new bitcoin mining overlords
FTFY
I hardly see neural nets as a particularly efficient way to do concrete maths...
This is why the broader sphere of AI (and not just the classical AI) has to encompass multiple disciplines. If you were to create an academic sphere called "intelligence", it would take in neuroscience, neuropsychology, psychology and arguably even sociology. Every part of AI tries to model a particular level of abstraction of intelligent behaviour, and a single level of abstraction will never be enough
Why does "AI" have to be dominated by "classical AI"? Classical AI didn't even attempt to model intelligence -- it was pure behaviourism at heart -- so I don't really know if it can justify being called AI at all.
In that sense it's analogous to sigma-style mahematic operations, and that sort of implied iteration is perfectly acceptable in the declarative paradigm. As such, we don't need recursion for all of our list processing functions
In fact, this sigma-style iteration where there are no interdependancies between iterations is crucial to parallelisation, so traditional imperative iterative loops are a major barrier to progress.
One of my favourite tings at the moment is Python's list comprehensions, which I feel sit on a weird boundary between declarative and imperative. It's declarative because you don't specify how the iteration works but it's arguably imperative because everything except the iteration is explicit. But I think this is enough to make it declarative. I'm not talking it through the loop, and then iteration could be evaluated 0 upwards, -1 downwards or in any random order you want -- it's just not relevant at the programmer's level.
In that sense it's analogous to sigma-style mahematic operations, and that sort of implied iteration is perfectly acceptable in the declarative paradigm. As such, we don't need recursion for all of our list processing functions
"in procedural programming languages..."
Isn't that the problem we have in teaching programming? our refusal as programmers to accept declarative programming as "real" programming? The "programming" (i.e. scheduling and process planning),of past centuries was declarative, not imperative, and that's much easier to get our heads around as human beings.
"But... optimisation!" I hear you cry. How many of us really need to develop high-performance systems? I had a task that I spent months trying to code up imperatively, and when I said "stuff that" and fired up Prolog, I got my results in weeks instead of months. I'm happyrunning it as a batch job for now and optimising the code (i.e. rewriting in another language) later if I need to. For now, it means I can focus on other parts of the system and get the whole thing to full proof-of-concept rather than getting stuck on one part of it.
Besides, most of the heavy lifting in a lot of coding tasks is done in libraries, and from the programmer's point of view, that's effectively declarative programming -- I have no control of how SDL draws a circle on screen. Yes, I can start looking at how different libraries' internal code works and whatnot, but going down that road means hitting the conclusion that libraries are *worse* than declarative code, because declarative coding means that (in theory) one line can be evaluated whichever way is most efficient whereas a library call encodes a specific operational choice.
What we need is what I call "declarative-first programming" -- a language/environment that lets you code up a declarative algorithm to build your first working prototype then progressively optimise it section-by-section using more specific code until all the sections where efficiency matters are procedural, and all the lightweight (not processor/memory intensive) stuff is declarative.
I don't really buy the types-of-teacher/learner thing. I believe their is a near-optimal teaching strategy for every subject, and that you only need to tinker round the edges to incorporate learners' prior knowledge and individual aims. I reckon any apparent differences in learners beyond that are best explained as "tolerances" -- most learners can cope with a limited amount of inefficiency in teaching, and those tolerances may be individual. I aim to be a better teacher with every lesson I teach. If I was to say "learner differences" I would be professing a belief that you can only ever successfully teach a small part of your class at any one time. It would make my job easier, but it sounds a little dog-ate-my-homework to me.
If the programming makes the problems seem more relevant, there's no reason you can't learn both at once.
If you equate programming with standard maths, you're missing the importance of mutability. Maths is semantic and imperative, procedural programming is declarative. Mathematical variables are an entirely different beast to computer variables.
I think the next step after this would be an open index where people can just catalogue papers and individual observations and how they support/refute one another. And goodbye selective citation.
As I've commented further up, I think the real value here is counterexamples: if you have a paper that's built on selective evidence, you don't need to prove an alternative theory, but rather just find sufficient counterexamples to demonstrate that the paper is unreliable.