Seriously, if you can't recognize the hundreds of improvements since 2.2, you probably don't need them. It's all made of subtle changes to preferences panels, admin wizards and widgets placement. According to all HCI gurus, those are the kind of changes that make an interface feel 'better' for non programmers.
For all those folks comparing it to Firefox "Search web for selection", this is a very different beast. What Y!Q does is adding metadata to your query, so that only pages are listed which both
1) contain the terms you selected and 2) are related to the metadata of the original website.
Try typing while holding down your thumb(s)! It's a major pain and impacts my speed. Plus, what the hell are you supposed to do for a command with a space in it? You can still write the command name beforehand, whithout holding the special key, then select the text and invoke the command. Typing while holding the key is only required for search. Besides that, you usually have two thumbs so one remains free to press the spacebar if needed.
A single button click, followed by typing the command, possibly editing it and possibly cancelling it, followed by another click (or more realistically, the enter key) makes a lot more sense.
The single main idea behind Raskin's proposed interface is to automatize the user actions by habituating the user and turning cursor positioning into a reflex. You can buy this premise or not, but the LEAPing technique achieves this, and clicking with the mouse and then moving hands to keyboard does not.
It's called a quasimode. It has psychological benefits for interaction that have been checked with empirical experiments - old good science, you know. Having a dedicated key for entering the quasimode would be required for doing it a good interface, though.
Or maybe it's just a dinky (8MB?) demo that doesn't do their paradigm justice.
It is. I've read the book, and the demo doesn't do justice - all the interesting data-manipulation tools have not been implemented. The only remaining bit is the navigational idea of zoom-it-to-the-infinite.
It's not, except that they're visual instead of textual (as in command line symlinks) - so it's better for the end user even when it's the same abstract concept.
But then the grandparent had said that this interface didn't have indirect references, when it does.
Is this thing suppose to manage all your documents and applications? Does that mean everything is being displayed and active at the same time? The CPU and memory requirements of this must be off the chart.
How is this different to a windowing interface? There are programming techniques to overcome this requirements, and they are very similar in both interfaces.
Also you really should read about the thing before you critizise it (I must be new here...). The system is not mainly navigated by zooming in/out: it has a really simple and powerful find tool accessed entirely with keyboard - not that different to Emacs, but more user friendly.
This interface is supposed to be used with a dedicated keyboard. This keyboard would have a COMMAND key near the spacebar, so the "hold" in this "hold+type" sequence should be done with the thumb.
Then I found a page describing some arcane "command language" thing called THE which reminded me of using Emacs or vi, not exactly user friendly.
Exactly. But why were Emacs and vi not user friendly? It was because they were heavily modal interfaces, which made learning to use them a real pain. Raskin claims to have identified the design errors of those tools and constructed a better interface based on similar principles.
Emacs and vi are regarded as the most efficient programming environments for people in the know. The new interface would provide an equivalent interface without the steep learning curve and with increased usability. Read The Humane Interface (coralized link) for further explanation of the main concepts.
The interface does not get rid of symlinks, they're actually replaced by something better: portals.
An object *can* be in several places, but in all them you see the real object - updated to the second and fully functional, not just a proxy of the object with its properties crippled. Also the information is browsed visually only when doing visual tasks, otherwise you browse it with incremental text search (like the one found in Firefox).
This guy Raskin is incredibly insightful on what makes computers a pain to use. The proposed system combines the power of Emacs without the hard learning curve. Read a description of the navigation tools of this system in The Humane Environment.
Actually the very idea of this new interface is using the Find command for navigating the whole operating system - an incremental find, like the one in Firefox. So this webpage is an act of coherence.
If you don't like find-as-you-type, you'd be better not using the interface described in the article.
Bad guess. The system allowed to interpret arbitrary text as names of system commands, not as executable code. In that sense it was not different to a CLI except for that it was in-place execution rather than a one-dimensional text stream.
T.H.E. was supposed to be released as Open Source, but this really didn't happen as far as I know. That's not important though, as this interface implementation was never finished - the important bit was the system specification.
IMO, something as polished as an operating system kernel can only be created in a corporate setting. There are too many egos wanting different things, it'd be impossible to get a team of 100 coders to all agree to work towards the same set of goals. One guy wants X, another wants Y.
Slashdot Mirror
So you really enjoy using a circular control (the knob) that is controlled with a linear device (the mouse)??
Windows, Mac, KDE, GNOME and every other sane interfaces all actively avoid circular knobs in the computer screen because they're a bad idea.
You must be a KDE user...
Seriously, if you can't recognize the hundreds of improvements since 2.2, you probably don't need them. It's all made of subtle changes to preferences panels, admin wizards and widgets placement. According to all HCI gurus, those are the kind of changes that make an interface feel 'better' for non programmers.
actually they have a very good tool in Longhorn that does precisely that, it's called SiS (Stuff i've Seen).
...street illumination, water service and waste disposal used to be luxuries. What if access to information counted as a required service in future?
For all those folks comparing it to Firefox "Search web for selection", this is a very different beast. What Y!Q does is adding metadata to your query, so that only pages are listed which both
1) contain the terms you selected and
2) are related to the metadata of the original website.
I wonder what the results could be if this technology were merged with masive metadata with distributed generation.
Try typing while holding down your thumb(s)! It's a major pain and impacts my speed. Plus, what the hell are you supposed to do for a command with a space in it?
You can still write the command name beforehand, whithout holding the special key, then select the text and invoke the command. Typing while holding the key is only required for search. Besides that, you usually have two thumbs so one remains free to press the spacebar if needed.
A single button click, followed by typing the command, possibly editing it and possibly cancelling it, followed by another click (or more realistically, the enter key) makes a lot more sense.
The single main idea behind Raskin's proposed interface is to automatize the user actions by habituating the user and turning cursor positioning into a reflex. You can buy this premise or not, but the LEAPing technique achieves this, and clicking with the mouse and then moving hands to keyboard does not.
It's called a quasimode. It has psychological benefits for interaction that have been checked with empirical experiments - old good science, you know. Having a dedicated key for entering the quasimode would be required for doing it a good interface, though.
Yes but Emacs is modal, while THE is modeless.
Ok, thanks for the info. It has been a long time since I last checked the code release of THE.
Or maybe it's just a dinky (8MB?) demo that doesn't do their paradigm justice.
It is. I've read the book, and the demo doesn't do justice - all the interesting data-manipulation tools have not been implemented. The only remaining bit is the navigational idea of zoom-it-to-the-infinite.
Yes, it's a window manager that resembles a shell. Actually, it resembles Emacs without the Meta-Caps-Shift-Esc-touch-your-nose madness.
Yes, 'cause people love computers as they're right now, true?
It's not, except that they're visual instead of textual (as in command line symlinks) - so it's better for the end user even when it's the same abstract concept.
But then the grandparent had said that this interface didn't have indirect references, when it does.
Not so. You just have to search for the text in the link (by typing it's first letter, for example), then clicking it (with ENTER).
Mod parent +1 Insightful. He actually has RTFA.
Is this thing suppose to manage all your documents and applications? Does that mean everything is being displayed and active at the same time? The CPU and memory requirements of this must be off the chart.
How is this different to a windowing interface? There are programming techniques to overcome this requirements, and they are very similar in both interfaces.
Also you really should read about the thing before you critizise it (I must be new here...). The system is not mainly navigated by zooming in/out: it has a really simple and powerful find tool accessed entirely with keyboard - not that different to Emacs, but more user friendly.
This interface is supposed to be used with a dedicated keyboard. This keyboard would have a COMMAND key near the spacebar, so the "hold" in this "hold+type" sequence should be done with the thumb.
Then I found a page describing some arcane "command language" thing called THE which reminded me of using Emacs or vi, not exactly user friendly.
Exactly. But why were Emacs and vi not user friendly? It was because they were heavily modal interfaces, which made learning to use them a real pain. Raskin claims to have identified the design errors of those tools and constructed a better interface based on similar principles.
Emacs and vi are regarded as the most efficient programming environments for people in the know. The new interface would provide an equivalent interface without the steep learning curve and with increased usability. Read The Humane Interface (coralized link) for further explanation of the main concepts.
The interface does not get rid of symlinks, they're actually replaced by something better: portals.
An object *can* be in several places, but in all them you see the real object - updated to the second and fully functional, not just a proxy of the object with its properties crippled. Also the information is browsed visually only when doing visual tasks, otherwise you browse it with incremental text search (like the one found in Firefox).
This guy Raskin is incredibly insightful on what makes computers a pain to use. The proposed system combines the power of Emacs without the hard learning curve. Read a description of the navigation tools of this system in The Humane Environment.
Actually the very idea of this new interface is using the Find command for navigating the whole operating system - an incremental find, like the one in Firefox. So this webpage is an act of coherence.
If you don't like find-as-you-type, you'd be better not using the interface described in the article.
Bad guess. The system allowed to interpret arbitrary text as names of system commands, not as executable code. In that sense it was not different to a CLI except for that it was in-place execution rather than a one-dimensional text stream.
There are articles describing Jef Raskin and The Humane Interface in the Wikipedia.
T.H.E. was supposed to be released as Open Source, but this really didn't happen as far as I know. That's not important though, as this interface implementation was never finished - the important bit was the system specification.
Maybe, but also many of them will prefer to file a lawsuit than to shoot someone who bothers them.
So you think you'd be better served by the /. editors also doing the comment moderations by hand?
IMO, something as polished as an operating system kernel can only be created in a corporate setting. There are too many egos wanting different things, it'd be impossible to get a team of 100 coders to all agree to work towards the same set of goals. One guy wants X, another wants Y.
Oh wait...