Things like, should your mouse be click to focus, or focus follows mouse. I like FFM, but it tends breaks multi window apps. It doesnt fit well in the Mac world, so I bid it a sad goodbye and moved on.
Really. I tried using FFM with CodeTek Virtual Desktop and quickly abandoned it.
Certain click-through inconsistencies can be annoying and that's the sort of UI refinement I'd prefer Apple focus (no pun) more attention on.
Greetings from the other side of the O'ahu, mmmpie, if indeed you're in Honolulu.
Actually, the topics here are generating some interesting discussion.
I must disagree. Hobbes began this thread by asking the question "What can Apple do to advance the state of the art in UI?" A few of us posted with honest attempts to answer this question while adhering to the spirit of his original post...
Unfortunately, the thread rapidly declined into yet another "I think the close button should work this way" debate. Sorry, "debate" is too generous; "argument" seems more appropriate. I don't believe it's an exaggeration to claim that this topic has been thoroughly exhausted elsewhere in these forums.
Props to BuonRutto for a great post on Apple's UI approach - one bright spot in this otherwise dilapidated thread.
Okay, maybe it was more the potential for discussion of interesting topics, like what you wrote about metadata both here and in this thead. I was surprised how much your posts about it were like an eloquent summary of several I've written over the past few months. It's a topic I'd enjoy discussing more and it's still getting attention in that other thread, minus the llama fluff.
Redundancy is inevitable and unavoidable with any larger online communication group.
It's hard to roll out big ideas? about the UI when we're just coming in for little work "breaks".
So true.
Quote:
Originally posted by BuonRotto
I'm trying to wrack my brain for things I'd like to do more easily from the OS. If we keep track of this sort of stuff, we can probably get some idea of what Apple will improve. Problem is, we take how we work for granted, at least I do. It's harder to think about what we're missing when we learn to work with the limitations of the OS so well.
Smart observation. People figure out a system that works for them, and train themselves to work around shortcomings. When they can't, it completely exasperates them -- thus, I imagine, the devolution into endless "close button means quit / close button never means quit" discussions.
It's also a cute little paradox, I think, that the people are generally most interested in UI can generally figure out a system that works for them in a fairly brief time, even more easily blinding them to potential improvements.
Personal example: since OS 7.5 or so, I've been very comfortable managing the window clutter problem by aggressive use of hiding apps. (Toggling windows was a complete roll of the dice, though, until the command-~ was at last made more universal in OS X.) I've pretty much been a fan of the Dock since day one, but was even more delighted to discover the powerful hiding tools built into it (opt. and command-opt clicking). It worked great for me.
At the same, I was fuzzily aware in the back of my head that this was still a power-user solution, that general users weren't going to remember or enjoy pressing modifier keys while clicking on Dock icons (or desktop, or other windows) to manage their clutter. So when Exposé was introduced, it came as that odd, happy combination... as both complete surprise and utterly obvious. So, I'm between systems now. It's difficult for me to break eight years of instinctive hiding, and currently Exposé (system-wide) decides to ignore hidden apps. I've adopted a hybrid use: Exposé all the time to show the desktop, often for application use (esp. Photoshop, what a wonder!) and only very occasionally to switch between all apps.
Anyway. This babbling is all just to note that a problem can still exist even when one has worked out a pretty good system to work around it. You have to start at the the problems, even when they're submerged.
So, what are the problems? What are the (major) aspects of the UE that need to be addressed, or improved? I have my own tally, but I'd like to hear others. Not niggling little interface issues, but core issues: think about the way you use your computer, and what you use it for.
How many steps does it take you to accomplish a task? What do you *not* use your computer, because it's too much of a pain? What lurks in the back of mind, murmuring, "this works, but it could work better"?
The present interface works well - but that's not to say it couldn't be improved. I think we should be looking at improvements rather then something radically different.
I would count Expose' as an improvement in one specific task area
- rather then something 'radically different' altogether.
To be useful, people must feel comfortable with the modification.
Now - what modifications / enhancements could we make ?
- More control over the finder interface
-- it's already pretty good - though "jumping windows" annoy me
(icon view - window partly off screen - drag item to window
- can cause it to jump)
More Meta-Data support could help significantly
- Glad the Coloured labels are back - pity only 8 colours to choose from
It would be useful to be able to add "user defined lable" to files & Folders
- and then to be able to search on that lable
ie user to be able to define own categories of items
I have problems keeping track of where I put files
I end up with a complex web of files & folders & topics & categories
- I sometimes put files into the 'wrong place' simply because I can't quickly find the right or best place for it.
I try hard to organise my files - but it's never quite good enough
- basically because I am not able to store enough metadata
So I end up searching on general location, file type and approximate date, with a guessed keyword.
Being able to use my own set of keywords - (without generating great long file names) would help
- Kind of like the coloured labels - it's an orthogonal indexing system.
- think multidimensional - with user defined categories.
Another thought - on the GUI side
- it would be nice to be able to reduce wasted screen space in some circumstances.
eg: the XCode help browser wastes lots of screen space the left hand 'groups' space is not collapsable - why not ?
-- it's already pretty good - though "jumping windows" annoy me
(icon view - window partly off screen - drag item to window
This is a feature of Spring-loaded folders and windows. You can disable it by visiting your Finder preferences and un-checking under the General tab. You can still activate it manually at any time with the spacebar whether checked or not.
In 10.3 it's annoying having to mouse-click buttons to confirm/cancel replacing a preexisting file when copying in Finder. I want <return> (Replace) and command-. or <escape> (Stop) keyboard shortcuts back.
Maybe something along the lines of the app "Spring" but with a better way of associating things.
So when you pull up a document in a particular associated group, a little floating palette or something pops up with pointers to relevant email, Word files, spreadsheets, whatever. So you have this little ecosystem of different projects.
Word 2004 comes close, but it's suite-specific. I'd like to see it system-wide. Maybe someone could come up with some real uses for "Services".
UIs will only change when the way computers interact with each other is changed. look at the technology that exists right now, the current method of user interaction that fails to take into account the networking revolution. just wait and see... my friend and i are writing a new type of system (no joke!) that will change everything. if we can't do it, then the idea will live on through others. what people need to do is realize that the current method of file access is too limited and that the free flow of all information is the key to the future of computing.
i am purposefully being vaugue because obviously, one must protect proprietary ideas...
At this point we're talking evolution rather than revolution, because the combination of decades of legacy compatibility and "training" in specific apps and user interfaces have calcified the PC experience. In addition to hewing pretty closely to the Mac legacy, OS X brought the UNIX legacy on board, and that's even older than Apple is. So OS X has to move gradually as well. Whatever brand-spanking-new thing they decide to implement can't break cp, mv, cat, and the whole phalanx of POSIX commands.
Better filesystems are nothing new. Newton had a brilliant, database-like FS (the "soup"), and there are many more penned by computer scientists over the last 30 years that were ignored as business rushed to standardize on a shockingly primitive system (I remember someone lamenting that DOS/Windows' FAT "set filesystem design back 20 years"). The truly sobering thing to realize is that a lot of the "innovation" people are talking about now is stuff that was hashed out in the '60s, '70s and '80s but never implemented outside of some university. The point is that all too frequently we have what we have by default, not because there haven't been any better ideas in the intervening (or preceding!) decades.
Some things to look out for:
More presentation of context-dependent functions. Context menus came out when 800x600 was pretty good, and when screen compositing was a primitive, fire-and-forget affair. Their problem has always been that most people don't even know they exist. But the role they fill is important, especially in applications whose global menu bars are stuffed to the gills. With sophisticated on-screen compositing and much higher screen resolutions becoming prevalent, expect more "dashboard" like graphical widgets that appear when they're needed and disappear when they aren't. Cocoa user interfaces are incredibly self-aware and dynamic, so they lend themselves to this sort of arrangement anyway.
The sublimation of the operating system. If there are any next-generation filesystems or the like, and in particular any spiffy metadata implementations, they will be used toward this end. Essentially, the idea here is that Finder is sort of like OS 9's memory manager: It enables you to do a lot of low-level housekeeping that's not being done for you. Now, since PCs are wildly general-purpose, Finder will always be there, but it will increasingly become a fallback. Who moves MP3 files around after installing iTunes? Who browses MP3 files in Finder after installing iTunes? Increasingly, the OS will simply be a provider of ways to implement focused, special-purpose solutions like iTunes and the user will deal almost exclusively with applications. This is nothing new or radical, it's just that the mainstream is slowly being pulled this way.
The sublimation of basic functionality. We're already seeing this: Who needs a word processor? Right now, in Panther, any app that wants a text field with most of the capabilities of a WP can simply ask for one. This will become more true, not less true, and if the OS has, say, Word file format compatibility available as a built-in then instead of a "word processor" you can have any number of apps that include and enhance that capability as necessary. After all, I think that most people's needs could be met by AppleWorks - the AppleWorks that ran on the Apple ][, that is. There are phones that can do this now.
We can also mourn for what has been lost, and maybe push for its return: Cocoa's predecessor on NeXTStep was designed so that the messages passed around within the application were exposed and available to other applications. As a result, there was in effect an OS-standard plugin architecture. Third parties used this to turn the anemic Terminal.app that NeXT shipped into a featureful powerhouse. This capability exists in Mac OS X right now. So why don't we see it? Because commercial vendors want to ship closed, hermetically sealed systems. They can use Carbon, or they can use Cocoa and use a utility to "strip" the application of the hooks that allow other apps to plug in to it. If you like the idea of an OpenDoc that really works, a fertile software landscape that allows a great deal of choice and flexibility, start lobbying for developers to stop stripping their apps. You'll probably hear something about "trade secrets" or "copyright," but somehow when NeXTStep (and later OpenStep) were running huge, mission-critical software designed this way that was never a problem... This is mostly about $$$, partly about branding, and partly about support. The potential gains if every application is a platform (in effect) are immense.
Also, I think the Save command will change. Applications like iCal and iTunes are acclimating people to the idea that you just make a change and it's made. The idea that a change is unsaved until you explicitly say otherwise dates to the era when PCs didn't have any sort of persistent storage built in, so the user had to explicitly save after inserting a floppy. Now? That's silly. Save is still useful for generating documents in particular formats in particular places, so it's not going away, but I think storage is at the point where the application can save state and drop you right back where you were on launch. OmniWeb's Workspaces feature is a big step in this direction.
Finally... the transparent availability of resources. Rendezvous is basically the key to this: What's theoretically available to you is actually available to you at any given time, and the details of how and when to make it available are left up to the system and/or application. This is plug and play for the networked world.
you're spot on in much of what you mentioned amorph. the problem here, is that for some reason, companies that are developing OS's right now have something blocking their ability to see these things.
most importantly, is the part about using the finder (or whatever you call it) as the interface in which to manipulate information. data is data, it's all just different information mediums. text, audio, video, whatever you need to manipulate, it will be done in the finder.
in terms of file access, start thinking outside the box. literally. has anyone used sharescan here? think about that combined with wifi and netboot. i've said too much already.
The sublimation of the operating system. If there are any next-generation filesystems or the like, and in particular any spiffy metadata implementations, they will be used toward this end. Essentially, the idea here is that Finder is sort of like OS 9's memory manager: It enables you to do a lot of low-level housekeeping that's not being done for you. Now, since PCs are wildly general-purpose, Finder will always be there, but it will increasingly become a fallback. Who moves MP3 files around after installing iTunes? Who browses MP3 files in Finder after installing iTunes? Increasingly, the OS will simply be a provider of ways to implement focused, special-purpose solutions like iTunes and the user will deal almost exclusively with applications. This is nothing new or radical, it's just that the mainstream is slowly being pulled this way.
I agree, but to put it more specifically: I think system support for metadata will begin to erase the boundaries between the Finder and the more database-like representation of data in the filesystem, like iTunes. Apple will most likely provide the Finder as a generic file/metadata viewer/manipulator along with an API which programs like iTunes can use to embed the same underlying code in custom interfaces. Much of this has already been demonstrated in BeOS and OS/2 Workplace Shell (for the embedding). For those that remember Lotus Improv on the NeXT, I would really like to see that sort of intuitive and flexible metadata representation available through the Finder itself.
Aesthetics are next to irrelevant to me. Sure, I like it to look pretty, but one pretty is much like another.
Now take a look at the active window. Its shadows are wider so it's easy to distinguish it from other windows. You can identify the active window without thinking about it because this is one way your brain identifies if one thing is closer to you than another.
Changing the colour of the title bar is not the same. You always have to think about "red is active - blue is inactive". It's not an automatic process.
options are *never* bad especially since it doesn't have to affect how anyone else works
This is the premise on which many interface suggestions seem to be based.
Yet, it is patently false. Far more interfaces are ruined by too many options than by too few options.
Options are the easy way out for interface designers and software developers. Rather than spending the time to design and implement a quality interface, they tack on additional features with complete disregard to how the combination of features affects typical use.
Development is always constrained by funding and time. By adding options, time and money is taken away from making seamless, the core-feature set and workflow.
As a general rule, Apple seems to take a different course than much of the industry. Rather than competing for the longest bulleted list of options and features, they concentrate on just making the interface work. When done properly, interfaces satisfy the vast majority of users.
I'm quite glad that Apple chooses to design quality interfaces for the majority of users rather than spending the same development resources on options too numerous to be done well.
Some utilities seek to extend (functionality), or sometimes simply to offer more options to various interface elements.
- Those that offer more options, and even some types of functionality, most often end up as tacky giltz. - which might quite suite some people, but overall is to be avoided.
Very few utilities offer genuinely useful functionality that might be seen as an extension of the interface. (as opposed to separate application functionality)
I was more concerned with somehow making 'a better interface' - or to be more precise a 'better and enhanced' interface.
- In other words can we do any better ?
One such example has been the introduction of tabs in some windows, to collect what would otherwise be multiple windows. - Usually those applications supporting this also offer the option to split them into multiple windows if that's what's wanted - but the tabbed version is usually best.
I was wondering about "the tab not yet invented" - for want of a better description - (obviously not a tab at all) - since that's already been invented.
The acid test is - does it improve things - or is it just glitz ?
One of the most recient such inventions has been the "roll over" control
- used on numerious web pages.
One question that keeps coming up is can anything actually useful be done with 3D representations - or is that just glitz ?
I agree 100%, and as you said, Cocoa is surprisingly able to handle these dynamic user controls and options that are context-based. SketchUp what is calls, "inference" snaps in 3D space. Auto-snaps are nothing new, but these and the parametrics for these objects are far more intelligent than they used to be. I think we'll see these concepts of inference and parameter-driven UIs as we slowly progress out of the desktop metahpor. Though they don't run counter to a desktop idea per se, the way the desktop systems have been implemented is a stalling point. The real trick though is choosing what is dynamic and what is stable and fixed. I think something like a menubar will always remain, just as something like a Finder will always be around too.
Quote:
If there are any next-generation filesystems or the like, and in particular any spiffy metadata implementations, they will be used toward this end. ...Now, since PCs are wildly general-purpose, Finder will always be there, but it will increasingly become a fallback. ...Increasingly, the OS will simply be a provider of ways to implement focused, special-purpose solutions like iTunes and the user will deal almost exclusively with applications.
If Hobbes is still around, this is what I was trying to say in another thread about why the Fidner would be, in my words, "de-centralized," that it wouldn't remain the primary aspect or starting point of the Mac experience.
Quote:
We can also mourn for what has been lost, and maybe push for its return: Cocoa's predecessor on NeXTStep was designed so that the messages passed around within the application were exposed and available to other applications. As a result, there was in effect an OS-standard plugin architecture. ...They can use Carbon, or they can use Cocoa and use a utility to "strip" the application of the hooks that allow other apps to plug in to it.
Some Cocoa vendors still keep this though. They're the ones who make useful Services for other apps to tap into also.
Quote:
Also, I think the Save command will change.
It think it depends on how it's implemented. In some cases, you don't want to screw a user if they decide they want to go back. A few apps already carry undo info with the file, so at any point they can revert to previous states from some time ago. This is mostly true for higher-end apps where it's just too time consuming and risky to not be able to go way back if needed. Right now when I build a model, I save a bunch of files as it progresses so I can always go back to old material after I've altered it. Ideally, there is a more compact but also clean solution so that I'm not handling gigantic files of unused geomoetry until I ask for it.
Quote:
Finally... the transparent availability of resources. Rendezvous is basically the key to this...
A highly flexible, multimedia and networked pasteboard maybe too.
This talk of a more dynamic and transparent view of data across networks (network? what network?) again bring me back to Jeff Raskin's ideas for computer UIs to some extent.
The biggest trouble seems to be the UI metaphor, the fear that something more abstract is going to be a huge impediment to user comprehension. But if the concept is clear enough, it doesn't have to be literal like the desktop metaphor or the file cabinet or the airport concept.
Comments
Originally posted by Moogs
Someone needs to read some basic UI Design theory / primer.
Hey, shut up! Stop talking about me.
Originally posted by mmmpie
Things like, should your mouse be click to focus, or focus follows mouse. I like FFM, but it tends breaks multi window apps. It doesnt fit well in the Mac world, so I bid it a sad goodbye and moved on.
Really. I tried using FFM with CodeTek Virtual Desktop and quickly abandoned it.
Certain click-through inconsistencies can be annoying and that's the sort of UI refinement I'd prefer Apple focus (no pun) more attention on.
Greetings from the other side of the O'ahu, mmmpie, if indeed you're in Honolulu.
Originally posted by sjk
Actually, the topics here are generating some interesting discussion.
I must disagree. Hobbes began this thread by asking the question "What can Apple do to advance the state of the art in UI?" A few of us posted with honest attempts to answer this question while adhering to the spirit of his original post...
Unfortunately, the thread rapidly declined into yet another "I think the close button should work this way" debate. Sorry, "debate" is too generous; "argument" seems more appropriate. I don't believe it's an exaggeration to claim that this topic has been thoroughly exhausted elsewhere in these forums.
Props to BuonRutto for a great post on Apple's UI approach - one bright spot in this otherwise dilapidated thread.
Originally posted by dglow
I must disagree.
Okay, maybe it was more the potential for discussion of interesting topics, like what you wrote about metadata both here and in this thead. I was surprised how much your posts about it were like an eloquent summary of several I've written over the past few months. It's a topic I'd enjoy discussing more and it's still getting attention in that other thread, minus the llama fluff.
Redundancy is inevitable and unavoidable with any larger online communication group.
Originally posted by BuonRotto
It's hard to roll out big ideas? about the UI when we're just coming in for little work "breaks".
So true.
Originally posted by BuonRotto
I'm trying to wrack my brain for things I'd like to do more easily from the OS. If we keep track of this sort of stuff, we can probably get some idea of what Apple will improve. Problem is, we take how we work for granted, at least I do. It's harder to think about what we're missing when we learn to work with the limitations of the OS so well.
Smart observation. People figure out a system that works for them, and train themselves to work around shortcomings. When they can't, it completely exasperates them -- thus, I imagine, the devolution into endless "close button means quit / close button never means quit" discussions.
It's also a cute little paradox, I think, that the people are generally most interested in UI can generally figure out a system that works for them in a fairly brief time, even more easily blinding them to potential improvements.
Personal example: since OS 7.5 or so, I've been very comfortable managing the window clutter problem by aggressive use of hiding apps. (Toggling windows was a complete roll of the dice, though, until the command-~ was at last made more universal in OS X.) I've pretty much been a fan of the Dock since day one, but was even more delighted to discover the powerful hiding tools built into it (opt. and command-opt clicking). It worked great for me.
At the same, I was fuzzily aware in the back of my head that this was still a power-user solution, that general users weren't going to remember or enjoy pressing modifier keys while clicking on Dock icons (or desktop, or other windows) to manage their clutter. So when Exposé was introduced, it came as that odd, happy combination... as both complete surprise and utterly obvious. So, I'm between systems now. It's difficult for me to break eight years of instinctive hiding, and currently Exposé (system-wide) decides to ignore hidden apps. I've adopted a hybrid use: Exposé all the time to show the desktop, often for application use (esp. Photoshop, what a wonder!) and only very occasionally to switch between all apps.
Anyway. This babbling is all just to note that a problem can still exist even when one has worked out a pretty good system to work around it. You have to start at the the problems, even when they're submerged.
So, what are the problems? What are the (major) aspects of the UE that need to be addressed, or improved? I have my own tally, but I'd like to hear others. Not niggling little interface issues, but core issues: think about the way you use your computer, and what you use it for.
How many steps does it take you to accomplish a task? What do you *not* use your computer, because it's too much of a pain? What lurks in the back of mind, murmuring, "this works, but it could work better"?
Originally posted by Kickaha
Been done.
Now I understand.
I would count Expose' as an improvement in one specific task area
- rather then something 'radically different' altogether.
To be useful, people must feel comfortable with the modification.
Now - what modifications / enhancements could we make ?
- More control over the finder interface
-- it's already pretty good - though "jumping windows" annoy me
(icon view - window partly off screen - drag item to window
- can cause it to jump)
More Meta-Data support could help significantly
- Glad the Coloured labels are back - pity only 8 colours to choose from
It would be useful to be able to add "user defined lable" to files & Folders
- and then to be able to search on that lable
ie user to be able to define own categories of items
I have problems keeping track of where I put files
I end up with a complex web of files & folders & topics & categories
- I sometimes put files into the 'wrong place' simply because I can't quickly find the right or best place for it.
I try hard to organise my files - but it's never quite good enough
- basically because I am not able to store enough metadata
So I end up searching on general location, file type and approximate date, with a guessed keyword.
Being able to use my own set of keywords - (without generating great long file names) would help
- Kind of like the coloured labels - it's an orthogonal indexing system.
- think multidimensional - with user defined categories.
Another thought - on the GUI side
- it would be nice to be able to reduce wasted screen space in some circumstances.
eg: the XCode help browser wastes lots of screen space the left hand 'groups' space is not collapsable - why not ?
Originally posted by Zoz
-- it's already pretty good - though "jumping windows" annoy me
(icon view - window partly off screen - drag item to window
This is a feature of Spring-loaded folders and windows. You can disable it by visiting your Finder preferences and un-checking under the General tab. You can still activate it manually at any time with the spacebar whether checked or not.
So when you pull up a document in a particular associated group, a little floating palette or something pops up with pointers to relevant email, Word files, spreadsheets, whatever. So you have this little ecosystem of different projects.
Word 2004 comes close, but it's suite-specific. I'd like to see it system-wide. Maybe someone could come up with some real uses for "Services".
UIs will only change when the way computers interact with each other is changed. look at the technology that exists right now, the current method of user interaction that fails to take into account the networking revolution. just wait and see... my friend and i are writing a new type of system (no joke!) that will change everything. if we can't do it, then the idea will live on through others. what people need to do is realize that the current method of file access is too limited and that the free flow of all information is the key to the future of computing.
i am purposefully being vaugue because obviously, one must protect proprietary ideas...
Originally posted by yogisfunhouse
...one must protect proprietary ideas...
Yes, patent it, hell M$ patented double click so file access should be a breeze.
Better filesystems are nothing new. Newton had a brilliant, database-like FS (the "soup"), and there are many more penned by computer scientists over the last 30 years that were ignored as business rushed to standardize on a shockingly primitive system (I remember someone lamenting that DOS/Windows' FAT "set filesystem design back 20 years"). The truly sobering thing to realize is that a lot of the "innovation" people are talking about now is stuff that was hashed out in the '60s, '70s and '80s but never implemented outside of some university. The point is that all too frequently we have what we have by default, not because there haven't been any better ideas in the intervening (or preceding!) decades.
Some things to look out for:
More presentation of context-dependent functions. Context menus came out when 800x600 was pretty good, and when screen compositing was a primitive, fire-and-forget affair. Their problem has always been that most people don't even know they exist. But the role they fill is important, especially in applications whose global menu bars are stuffed to the gills. With sophisticated on-screen compositing and much higher screen resolutions becoming prevalent, expect more "dashboard" like graphical widgets that appear when they're needed and disappear when they aren't. Cocoa user interfaces are incredibly self-aware and dynamic, so they lend themselves to this sort of arrangement anyway.
The sublimation of the operating system. If there are any next-generation filesystems or the like, and in particular any spiffy metadata implementations, they will be used toward this end. Essentially, the idea here is that Finder is sort of like OS 9's memory manager: It enables you to do a lot of low-level housekeeping that's not being done for you. Now, since PCs are wildly general-purpose, Finder will always be there, but it will increasingly become a fallback. Who moves MP3 files around after installing iTunes? Who browses MP3 files in Finder after installing iTunes? Increasingly, the OS will simply be a provider of ways to implement focused, special-purpose solutions like iTunes and the user will deal almost exclusively with applications. This is nothing new or radical, it's just that the mainstream is slowly being pulled this way.
The sublimation of basic functionality. We're already seeing this: Who needs a word processor? Right now, in Panther, any app that wants a text field with most of the capabilities of a WP can simply ask for one. This will become more true, not less true, and if the OS has, say, Word file format compatibility available as a built-in
We can also mourn for what has been lost, and maybe push for its return: Cocoa's predecessor on NeXTStep was designed so that the messages passed around within the application were exposed and available to other applications. As a result, there was in effect an OS-standard plugin architecture. Third parties used this to turn the anemic Terminal.app that NeXT shipped into a featureful powerhouse. This capability exists in Mac OS X right now. So why don't we see it? Because commercial vendors want to ship closed, hermetically sealed systems. They can use Carbon, or they can use Cocoa and use a utility to "strip" the application of the hooks that allow other apps to plug in to it. If you like the idea of an OpenDoc that really works, a fertile software landscape that allows a great deal of choice and flexibility, start lobbying for developers to stop stripping their apps. You'll probably hear something about "trade secrets" or "copyright," but somehow when NeXTStep (and later OpenStep) were running huge, mission-critical software designed this way that was never a problem... This is mostly about $$$, partly about branding, and partly about support. The potential gains if every application is a platform (in effect) are immense.
Also, I think the Save command will change. Applications like iCal and iTunes are acclimating people to the idea that you just make a change and it's made. The idea that a change is unsaved until you explicitly say otherwise dates to the era when PCs didn't have any sort of persistent storage built in, so the user had to explicitly save after inserting a floppy. Now? That's silly. Save is still useful for generating documents in particular formats in particular places, so it's not going away, but I think storage is at the point where the application can save state and drop you right back where you were on launch. OmniWeb's Workspaces feature is a big step in this direction.
Finally... the transparent availability of resources. Rendezvous is basically the key to this: What's theoretically available to you is actually available to you at any given time, and the details of how and when to make it available are left up to the system and/or application. This is plug and play for the networked world.
There's probably more, but I've gotta go.
most importantly, is the part about using the finder (or whatever you call it) as the interface in which to manipulate information. data is data, it's all just different information mediums. text, audio, video, whatever you need to manipulate, it will be done in the finder.
in terms of file access, start thinking outside the box. literally. has anyone used sharescan here? think about that combined with wifi and netboot. i've said too much already.
Originally posted by Amorph
The sublimation of the operating system. If there are any next-generation filesystems or the like, and in particular any spiffy metadata implementations, they will be used toward this end. Essentially, the idea here is that Finder is sort of like OS 9's memory manager: It enables you to do a lot of low-level housekeeping that's not being done for you. Now, since PCs are wildly general-purpose, Finder will always be there, but it will increasingly become a fallback. Who moves MP3 files around after installing iTunes? Who browses MP3 files in Finder after installing iTunes? Increasingly, the OS will simply be a provider of ways to implement focused, special-purpose solutions like iTunes and the user will deal almost exclusively with applications. This is nothing new or radical, it's just that the mainstream is slowly being pulled this way.
I agree, but to put it more specifically: I think system support for metadata will begin to erase the boundaries between the Finder and the more database-like representation of data in the filesystem, like iTunes. Apple will most likely provide the Finder as a generic file/metadata viewer/manipulator along with an API which programs like iTunes can use to embed the same underlying code in custom interfaces. Much of this has already been demonstrated in BeOS and OS/2 Workplace Shell (for the embedding). For those that remember Lotus Improv on the NeXT, I would really like to see that sort of intuitive and flexible metadata representation available through the Finder itself.
Originally posted by mmmpie
Aesthetics are next to irrelevant to me. Sure, I like it to look pretty, but one pretty is much like another.
Now take a look at the active window. Its shadows are wider so it's easy to distinguish it from other windows. You can identify the active window without thinking about it because this is one way your brain identifies if one thing is closer to you than another.
Changing the colour of the title bar is not the same. You always have to think about "red is active - blue is inactive". It's not an automatic process.
Things like this divide intelligent from pretty.
Originally posted by Concord
options are *never* bad especially since it doesn't have to affect how anyone else works
This is the premise on which many interface suggestions seem to be based.
Yet, it is patently false. Far more interfaces are ruined by too many options than by too few options.
Options are the easy way out for interface designers and software developers. Rather than spending the time to design and implement a quality interface, they tack on additional features with complete disregard to how the combination of features affects typical use.
Development is always constrained by funding and time. By adding options, time and money is taken away from making seamless, the core-feature set and workflow.
As a general rule, Apple seems to take a different course than much of the industry. Rather than competing for the longest bulleted list of options and features, they concentrate on just making the interface work. When done properly, interfaces satisfy the vast majority of users.
I'm quite glad that Apple chooses to design quality interfaces for the majority of users rather than spending the same development resources on options too numerous to be done well.
Some utilities seek to extend (functionality), or sometimes simply to offer more options to various interface elements.
- Those that offer more options, and even some types of functionality, most often end up as tacky giltz. - which might quite suite some people, but overall is to be avoided.
Very few utilities offer genuinely useful functionality that might be seen as an extension of the interface. (as opposed to separate application functionality)
I was more concerned with somehow making 'a better interface' - or to be more precise a 'better and enhanced' interface.
- In other words can we do any better ?
One such example has been the introduction of tabs in some windows, to collect what would otherwise be multiple windows. - Usually those applications supporting this also offer the option to split them into multiple windows if that's what's wanted - but the tabbed version is usually best.
I was wondering about "the tab not yet invented" - for want of a better description - (obviously not a tab at all) - since that's already been invented.
The acid test is - does it improve things - or is it just glitz ?
One of the most recient such inventions has been the "roll over" control
- used on numerious web pages.
One question that keeps coming up is can anything actually useful be done with 3D representations - or is that just glitz ?
Originally posted by Amorph
Some things to look out for:
More presentation of context-dependent functions.
I agree 100%, and as you said, Cocoa is surprisingly able to handle these dynamic user controls and options that are context-based. SketchUp what is calls, "inference" snaps in 3D space. Auto-snaps are nothing new, but these and the parametrics for these objects are far more intelligent than they used to be. I think we'll see these concepts of inference and parameter-driven UIs as we slowly progress out of the desktop metahpor. Though they don't run counter to a desktop idea per se, the way the desktop systems have been implemented is a stalling point. The real trick though is choosing what is dynamic and what is stable and fixed. I think something like a menubar will always remain, just as something like a Finder will always be around too.
If there are any next-generation filesystems or the like, and in particular any spiffy metadata implementations, they will be used toward this end. ...Now, since PCs are wildly general-purpose, Finder will always be there, but it will increasingly become a fallback. ...Increasingly, the OS will simply be a provider of ways to implement focused, special-purpose solutions like iTunes and the user will deal almost exclusively with applications.
If Hobbes is still around, this is what I was trying to say in another thread about why the Fidner would be, in my words, "de-centralized," that it wouldn't remain the primary aspect or starting point of the Mac experience.
We can also mourn for what has been lost, and maybe push for its return: Cocoa's predecessor on NeXTStep was designed so that the messages passed around within the application were exposed and available to other applications. As a result, there was in effect an OS-standard plugin architecture. ...They can use Carbon, or they can use Cocoa and use a utility to "strip" the application of the hooks that allow other apps to plug in to it.
Some Cocoa vendors still keep this though. They're the ones who make useful Services for other apps to tap into also.
Also, I think the Save command will change.
It think it depends on how it's implemented. In some cases, you don't want to screw a user if they decide they want to go back. A few apps already carry undo info with the file, so at any point they can revert to previous states from some time ago. This is mostly true for higher-end apps where it's just too time consuming and risky to not be able to go way back if needed. Right now when I build a model, I save a bunch of files as it progresses so I can always go back to old material after I've altered it. Ideally, there is a more compact but also clean solution so that I'm not handling gigantic files of unused geomoetry until I ask for it.
Finally... the transparent availability of resources. Rendezvous is basically the key to this...
A highly flexible, multimedia and networked pasteboard maybe too.
This talk of a more dynamic and transparent view of data across networks (network? what network?) again bring me back to Jeff Raskin's ideas for computer UIs to some extent.
The biggest trouble seems to be the UI metaphor, the fear that something more abstract is going to be a huge impediment to user comprehension. But if the concept is clear enough, it doesn't have to be literal like the desktop metaphor or the file cabinet or the airport concept.