Early peak into Gnome 3’s potential
February 26, 2010
I am stoked to be designing a Gnome that kicks ass for hackers too. McCann is right: the ‘choice‘ between design for office monkeys and hacking monkeys is a false dichotomy. We can make both their lives better: they’re running (almost :=) the same cognitive architecture underneath. Writing for ourselves can be fucking sweet. If we’re eager and happy, we’ll make it better for everyone.
Lets get some new hackers in on this shit. Kids? Time to shine.
Hackfest In Summary
The Gnome 2010 UX Hackfest has been such fun! Today is the last day, reckon we’ll be riotous. It was a wonderful, familiar memory. I know not everyone reciprocates (hi hatemail, bye hatemail), but I missed you Gnome.
It was rad to spool out into airy discussions with everyone from Mark to Mo’, but still land some useful designs at the end. That doesn’t happen on every trip to lala land! Vis a vis the hackfest, this doc is only the tip of the iceberg – and a unfinished tip at that. Many amazing designs for other pieces of gnome have emerged too: from system preferences to empathy. As always, props to my homies @ Canonical & Google for splashin’ the green on us.
I give you: Unfinished Writeup
PS: there’s a couple cool animated mockups at the bottom of this post. they’re good for getting a feel.
I’ll make more. Hopefully loads more static mockups coming too. And once I get my scanner, the mockups will be more readable (sorry).
Anyway, here’s some thoughts in case you don’t want to read the PDF…
We want Gnome 3 to become…
- a transparent tool…
- a tool that accommodates complex tasks without breaking your focus.
- the WORK should occupy your mind, a familiar TOOL should disappear as you work through it, not at it.
- a tool that increasingly disappears in-hand as you become accustomed to it present-at-hand
- a toolbox that’s extremely fast to rummage through
- a toolbox where old/obscure tools can be readily found
- a toolbox where familiar files/tools/apps find your hand right when you need them… and recede into the periphery when you’re done with them
- provide ways to flow past the interruptions of an online existence
- turn task-management into a 500ms operation
- make ordinary laptops more physical, more direct, and much much faster to use for common operations (try to get the advantages sought by kb shortcuts & gestures for a much larger number of operations)
- using an ordinary laptop in ways that naught but touchscreens are now (and touchscreens have their own special problems that detract from their physicality advantages). a toolbox that encourages files, apps and tools to find your hands as you need them, and recede into the periphery when you don’t.
This doc isn’t much past the grey and orange concept-babble that richly fills my inner life. But it’ll give you a taste. And we’ll do better. It’ll get more concrete. Some is known that is not yet described, and some has been described that will turn out badly and need to be ditched. But play around. See how it feels. At this fetal stage, Explore > Reject.
We still need a lot more writing on how these ideas are concretely useful, and how important problems they raise are addressed/ameliorated. No design is perfect. Every design is a tradeoff. You get the center right, and then trim the pain from the edges. You learn lessons from the edges, and use the lessons to revise the core. Rinse, repeat.
I give you: The Task Pooper Animation
I give you: Absolut Trackpad Animation.
sorry the shots are from OS/X, nobody had globalmenu installed with a patched gtk+, and I needed global menus to demonstrate this feature, and was too lazy to do each of the frame mockups by hand.
I give you: Search by Iterative Date Comparison
see doc for meager details
I give you: Browse the pdf, be kind, and play nice.
Lots of stuff here won’t work like expected. That’s ok. We’ll just kill those features. Play around. Don’t just try to shoot it down (that shit is easy, I’ve thrown up a barn-sized target), try to build off it, improve it, riff off it.
And now, let the wild rumpus begin.
February 26, 2010 at 7:20 am
you appear to have a sense of humor.. what on earth are you doing on Planet Gnome?
February 26, 2010 at 7:35 am
So… are all the Absolut Trackpad/Physical Trackpad features triggered on edge touches? Without that, it seems like it would mess up regular trackpad usage- but that limits you to two to four gestures (most non-Mac trackpads have little arrows telling the user to use one or two edges as scrolling sliders).
February 26, 2010 at 7:38 am
This document contains a lot of great ideas!
I’m just wondering one thing.
There is a big focus on touchpad awesomeness of use.
I’ve to admit this is great. I did never use my touchpad because it is actually the worst ever interaction mechanism (as it is used now).
I’m using my mouse, never my touchpad. At work my computer has no touchpad at all, etc..
My point is: Is there so many people that are using the touchpad? When you are in office or at home, I see not the benefits to use the touchpad.
February 26, 2010 at 7:53 am
I just love the pooper idea! It seems to be quite a wonderful way of working with projects.
February 26, 2010 at 8:08 am
Lots of cool stuff in there. The pooper seems great, if I understood how it’s supposed to work. Very exciting work in general, but I do have two concerns.
The first is making trackpad use a priority. I don’t even own a computer that has a trackpad, so any compromises made to general usability in service of making it really awesome with a trackpad are going to be strict losses for me (and I’m sure I’m not alone here).
The other is probably going to be a huge flame war in any case, so I might as well get my licks in first. That is, it seems there’s a unified menu bar thing going on in the mock-ups, which would be a real problem for me. I love the idea of collapsing the menu bar / title bar / shell top bar into a single entity in the case of a maximized window on the primary monitor, but I often work with many non-maximized windows, and often two discrete maximized windows (since I work with a two monitor set-up). As a result, accessing the menu of a window that does not have focus is something I do maybe dozens of times per day, and is one (of several) use cases where I find a unified menu bar is a complete failure.
Anyway, I know these are just early mockups and I’m getting into some serious bike-shedding, but the unified menubar is a real pet-peeve of mine, and I just couldn’t resists.
But, to end on a positive note, great work, and I look forward to seeing these ideas prototyped and tested.
February 26, 2010 at 8:11 am
I love your work!
What’s the tool used for create the task popper animated mockup?
February 26, 2010 at 8:14 am
Love your take at in-desktop task management; it would absolutely spare me using GTG for getting-things-done. Keep ’em coming!
February 26, 2010 at 8:14 am
A very interesting read! A priority should be to implement these ideas as soon as possible (even if everything isn’t thought out yet) so it can be tested, improved and to faster become stable.
I wonder (like Daiji above) about the focus on laptops with all the touchpad-oriented designs. I mostly use my laptop (a MacBook Pro) and feel crippled when at my desktop (which is supposed to be my primary working computer).
Is there perhaps some time available among the Gnome 3 designers to also discuss the desktop computer? Optimizing the workflow on a desktop should be an important goal as well.
February 26, 2010 at 8:33 am
@Jon: I don’t know whether this has been discussed as possible behaviour for a unified (global) menu-bar; but why couldn’t we have a cross between traditional menus and the global one?
Make only menus of maximised windows global, i.e. merge with the top bar. Save space and un-divide attention in maximised mode, spare the confusion when working with several windows.
February 26, 2010 at 8:35 am
No buts, I just like your ideas. Keep innovating! Don’t get discouraged or anything. Listen to valid concerns, but stick to the bigger ideas. We still live in a interface stone age.
February 26, 2010 at 8:45 am
Well. That looks interesting. I do always start worrying with these things that my usual model of interaction will be impossible, but then of course I realise that might be a good thing, if the new way is better. Some of this I like a LOT.
Lots of focus on touchpad I see – my laptop has a trackpoint, and my desktop has a mouse. It seems those will be very usable despite the touchpad focus though, okay so we don’t have absolutely mapped corner hotpoint things, but we can get our pointers there faster than a conventional touchpad anyway.
As long as ‘maximised by default’ doesn’t mean ‘very difficult to float a window in the background and keep half an eye on it’, I want this stuff on my computer.
February 26, 2010 at 9:03 am
One cool thing about this design, and this doesn’t always happen (just luck)… every feature described supports the others nicely, but no features require eachother. The touchpad stuff is cool, but this thing should be awesome w/o it (and it turns out: designing for this touchpad is exactly the same constraint as designing ‘for’ fitt’s law… so basically…. the mouse gets easier too. trackpoints less so, but screen edge has always been their one weakness, and it won’t make them any worse).
Maximized by default is merely maximized by default. We hope to get smart ways of doing passive awareness going (like ‘half keep an eye on it’), but they are quite abstract/fuzzy/vapor now.
February 26, 2010 at 9:08 am
To complete my thoughts..
I think the whole trackpad thing is really good.
But it concerns limited use cases and I’m concerned by this fact.
I can see those use cases:
* When you commute, or travel.
* When you cannot wait to hack some stuffs no matter where you are (laptop on your knees, or anywhere else).
* When you have no fix computer, no desk or no docking station nor usb keyboard or mouse.
Which people are your audience target ?
The two last use cases sound to me like geek cases.
You can rightly say that most of the devices will be mobile devices in the future. This is maybe already the case. But, which are those mobile devices ?
* Phones/Smartphones
* Netbooks
* Tablets
* Notebooks
I see no trackpad on phones/smartphones nor on tablets.
As there is many much better *mobile* devices than notebooks, you may consider this case as insignificant (in a near future).
Stays the netbooks, that are definitely horrible to use with their trackpad. The tiny problem is that tablets are focusing the exact same market. Which one will stay ? I do not know, but there is a chance that trackpads will definitely go away.
As you want to build “the future” of user interface. It may be useful to really think about what will be the future of the devices?
The part on task organisation is the best idea I’ve ever heard on that topic. Keep going with that, it is really awesome!
February 26, 2010 at 9:11 am
Further Comments on Touch Pad Feature
The touchpad is 100% not required to use any of the other good stuff. The touchpad stuff might not even be a good idea. It might feel awful. Too many people may brush it accidentally. Some computers have no touchpad, and some computers have a very small one (thinkpads, notably) where stealing the edges is infeasible.
Further explanation of how the touchpad thing works… they aren’t gestures, per se. I think they’ll end up feeling like gestures though: especially the ones that you most commonly use; that become subconscious.
I didn’t explain absolute edge very well, a couple more animations showing different situations would help.
Think of it this way: imagine that every part of your screen edge corresponded to a part of the trackpad edge. Actions must start from the edge, but once you’ve launched into a menu, you can move out into the center. Make sense?
INSIDE MENUS: the touchpad will be absolute mapped proportional to the MENU at that point, NOT the screen. The last item on the menu is always at the very opposite side of the touchpad. So for short menus, it will be very easy indeed to navigate.
Hardware Limitations: Some menus with some hardware (not all touchpads are equal…) will just be too tweaky to use this. Hopefully, the future is toward bigger/better touchpads (the biggest/best capacitive touchpad is called ‘touchscreen’… though touchscreens have some arm fatigue issues 😉
So it goes like this:
1) Tap and hold at edge, find the menu/item/action you want if you didn’t land perfectly (you’ll get much better, I reckon… we’ll see! this could 100% suck)
2) Swipe to the item you want. For commonly used items, esp in small menus, you should get very very good at this. Your hand will know the way.
3) Release and the item is activated.
So for example, lets grab something from the Task Pooper:
drop finger in the lower left corner, swipe up to the item we want, release. the item is popped from the pooper, and pops onto the screen.
Now lets add it back into the pooper.
drop finger in the lower left corner, swipe up the tiniest bit to the first item (which is “add selected item to this pooper”, obviously these labels are placeholder), and release.
Poof. In 1s you can pop an item from the task list and pop it back on. If this works well, its going to be fucking sweet. Using computers that don’t support this will feel like working w/o your hand.
February 26, 2010 at 9:15 am
@Daiji:
Mobile devices: Touchscreen=trackpad
So this would map really well on MDs.
February 26, 2010 at 9:17 am
it was done frame by frame by hand. the ‘video’ on the right is just sequential camera shots. the screen video on the left is just sequential screen grabs. it takes a long time. i hate doing it, but its effective. if anyone can suggest a better way of doing this, i’d be grateful.
February 26, 2010 at 9:19 am
Now that I look closer: Aren’t that tabs in a global menu-bar? I quite like that idea from Google Chrome OS.
February 26, 2010 at 9:19 am
most mobile devices w/ a touchscreen don’t even need this. it would be sweet on netbooks as they start sporting bigger touchpads though. that will be nice.
a sunlight readable, physical feeling, no HD so less fragile netbook would be a very different sort of computer. it would be “in the world” in a way these tweaky monsters we’re tethered to today could only dream of.
I once read that the only design goal the PalmPilot failed on was they wanted it to feel/be droppable, but they just couldn’t make it work. Ah well.
February 26, 2010 at 9:42 am
@Tom
Ok then 🙂
It was not so clear to me that there is no difference (from a functional point of view). That is why I was worried 😀
Anyway, the problem with computers that have no track/touchpad/screen stays. But clearly it is good to go in that direction. Thanks you for the enlightenment and keep going.
February 26, 2010 at 10:26 am
what about mobile devices? will gnome 3.0 target them by providing a ‘shell’ for mobile or your development tools (as it is now) for manufactures to create their UI vision?
What about desktop widgets?
February 26, 2010 at 11:46 am
Hey man, lose the menu bar, they are here for decades, maybe is to let it go. I recently saw a mockup of a video controls running on top of Docky, it looked nice!
Be careful not to make gnome-shell too alike to OSX.
February 26, 2010 at 12:20 pm
This looks excellent. What will happen with the incoming events bar that is currently at the bottom of gnome-shell? How does this deal with incoming events, communication, etc, at all?
February 26, 2010 at 12:32 pm
Wow, this totally rocks! Workspaces as tasks, archiving workspaces, whacking incoming tasks for later, searching by comparison… You got my vote!
February 26, 2010 at 1:03 pm
Love the ideas, like your views and I think there’s really a great potential in this preview, but could you please slow down on the “shit man f**kin’ shit rocks kickass” vocabulary ?
GNOME kicks ass, and GNOME 3 will even more, but I feel like GNOME’s always been the clean, clever, sober desktop, so let’s stay modest, this will emphasize even more on our awesomeness.
Again, no offense, thanks for your work !
February 26, 2010 at 1:27 pm
Gnome-Shell user for months now. I love how it and with stuff like this Gnome is really taking the lead on Desktop GUI interfaces and not just following. Keep up the good thinking.
I think the idea of a predictive mouse could go far. I see that feature being copied by others in the future. We should go ahead and implement it into Gnome-Shell with the activities hot corner. It would likely be even easier to predict when someone goes there being a corner and not an entire side.
February 26, 2010 at 1:43 pm
This is all really interesting. But why should tasks be constrained to the desktop you’re on? Imagine sweeping them, in the style used in Avatar, from one machine to an adjacent machine and getting all the windows and sessions associated with it dropped on the new machine. Or how about porting it across a network to another user through, say, UbuntuOne. Why should one be limited to only one task list? With reference to the email being used as a task manager, any improvement on local tasks could immediately be presented as improvements to doing tasks within your group/company/office/whatever. In addition to 1hr/1day/1week on the bottom you can have Phil/Marketing/HomeMachine.
I think you’ve hit the nail on the head on how to deal with tasks in the more general sense. We don’t deal with documents and we don’t run programs – we do tasks. Tasks should not be constrained by time (as you identify) but they also shouldn’t be constrained by space, machine, person etc.
February 26, 2010 at 2:12 pm
This is very clever. I have a few nitpicks, though:
When the Desktop concept was introduced at PARC, the idea was that you moved a file you were working with from the network onto the Desktop folder, opened it from there, and moved the file back to its rightful place when you were done. It was a very rigid metaphor based on the office. What you’re proposing with the Task Pooper fulfils that idea in a much better way. As such, having a separate Desktop folder now makes little sense. Shouldn’t that folder finally be deprecated?
Similarly, I don’t understand why you’re still pushing folders when tags are just absolutely better. Even those top-level folders can be very limiting: Do music videos go in Music or Video? Do audio books go in Documents or Music? It can get pretty frustrating, and tags can easily fix all that. Considering that you seem to be pushing against (forced) hierarchies, getting folders replaced with tags in this design would go a long way towards making it all make more sense.
February 26, 2010 at 2:19 pm
@Chris
I think that would be totally sweet, as long as I have a separate menubar for each of my maximized windows (on dualhead).
February 26, 2010 at 2:46 pm
It seems a usability review of gnome-shell was not on the list for the hackfest? I was waiting for it the whole time since a few UI experts say it is far from good…
On the other hand we now get to see something like the globalmenu implementation as potential GNOME 3 idea; something the gnome-shell guys said “no” to.
Feels like nobody really can define GNOME 3’s vision regarding UI in a clear sense.
I’d prefer to see a clear vision and some targets as the result of this hackfest, not some loose ideas.
Anyways, keep it up.
February 26, 2010 at 4:02 pm
First of all, I absolutely love _all_ your ideas, as brainstormy as they are, that’s the whole point! Second of all are some buts 🙂
– in the timeframes we are looking at getting all these realized to their full potential the desktop (as a platform) will be overdue. think global / cloud / distributed. The interaction you are designing seems implicitly constrained to a single system and seems to ignore anything beyond. see following points.
– your design is very single-screen oriented. think multi-sceen, multi-device, multi location. i may have an external monitor and i may want to have different activities on the two displays, or just one. i may have a second computer vnc-ing in. i may have my android phone tapping in. why can’t i drag my playlist window to my phone in the “virtual desktop manager” and continue enjoying music from the couch? or grab a browser window I had open at home — from work? or have a shared window-space with my colleague sitting next to me?
– touchpad is but one input medium. the mouse is still here, touchscreens are still coming. any key ui component should have efficient interaction models for many if not all. just an idea, for mice a pie menu might work, upon a special click or a special keypress, a circular menu could pop up, with the 4 screen corners as if were shifted to touch right at the pointer: so towards the top left would be your bottom right corner. The interaction would be exactly the same as if you tapped on the bottom right corner of your touchpad and started moving your pointer north-east.
– i’d expect screen sizes to differ more and more. the main issue i have with a global menu on a huge screen is the distance to travel from a tiny window in the bottom right corner to get to the file menu is a real pain. whatever you put top-left will have the same issue unless there is a means to “teleport”… similarly a cute dock with big bouncy icons will waste a lot of real estate on a small screen. certainly, there will be no one-size-fits-all (pun intended) solution for all screens, therefore any high-level interaction concept should not be rooted in such minute implementation details as screen corners or bouncy docks. An abstraction is needed that would say something like: on medium sized screens one of the corners is the place to render the “Tasks” UI controls. On a massive screen use pie menus for increased locality. On small devices render as a zoom UI to drill down. The point is not the actual solution, it is the abstraction that allows *essentially the same thing* to no not only scream on a macbook pro.
In summary, I think your design is too much constrained by targeting the Desktop, and that is limiting full potential way too early in the brainstorming process. Even so, your ideas are very refreshing and looking forward to more of them!
Greg
ps: the top-right corner could provide access to presence, as is now very common. “IM” status, lock screen… Maybe entry point to communicate with others.
February 26, 2010 at 4:17 pm
Please, please, please, get them to implement all the stuff you guys thought up. It sounds absolutely brilliant! The task pooper even made it to Ars Technica 🙂
February 26, 2010 at 4:42 pm
One quick thought: touchscreens are not like big touchpads. Scaling a menu so that the middle of the touchscreen = the middle of the menu makes no sense, whereas I think the idea of the menu-relative touchpad is great.
February 26, 2010 at 5:01 pm
So, it’s like the bastard child of Remember The Milk and Tracker, with a few shreds of Docky thrown in. No doubt it’ll be written in JavaScript.
February 26, 2010 at 5:16 pm
All of this is truly exciting! I definitely love the task pooper thingie 🙂
It would be awesome if tasks in the pooper could be shared on all my computers, along with the data they involve, so that I could potentially forget questions like “which computer was I working project x on last time?”
Hope to see some prototype soon 🙂
February 26, 2010 at 5:32 pm
I appreciate the idea of the task popper/(dock?).
However, it’s more a schedule poper as I see it. Tasks in my mind are entities of their own, that live not only in a calendar but also in the dimension of what is needed on the desktop.
Here would be my vision of this:
* Have separate desktops for each task (open documents, web pages, chat sessions, e-mails, contacts, bookmarks, browser/search history, …) and that persists across login sessions (under a name) and ideally across machines (web sync like Mozilla Weave)
* When I start a task I might want to start with the desktop of a nother task (snapshot) or a blank slate
* Have a system to do fast task switching (like the multi screen desktops, but not limited to the number of those) and also a system that notifies me of activities that belong to other tasks (like chat responses, mail replies, etc.) Ideally filtered by a priority/importance level of the task intruding (or filter by the planned work – stuff relevant to tasks that I have planned ot work on today or tomorrow –> scheduling)
* Make task contexts such as personal/work/hobby/moonlighting (or even sub contexts like projects) and allow me to swtich between those and filter notifications with different filters for disruptions in those. For example notify me of the important private reply at work, but not of the important reply on hobby. But while in hobby notify me of anything that is important.
* Allow to set availability levels for IM with switching contexts (tasks with different priorities and task contexts) automatically. Sometimes even seperate profiles.
* Allow the kind of scheduling that is indicated with the Task popper and within the time buckets a simple order/ranking to make a plan
* Give me desktop search with clear priority for task local results over global results.
* All tasks are archived, the waste basked is only task specific. May be have a Shredder to really delete things across tasks?
In simple terms give me a desktop per task (where I can have the relevant documents, contacts, e-mails, etc. just persistent around, even if they may be shared across tasks) that I can switch between easily and a virtual room that I can control what communication comes through the walls to intrude with the tasks that I’m at.
And I don’t start talking about sharing resources/documents accross users for colaboration. That would be the ultimate wirtual room/desktop, although who loves someone else rearranging one’s desktop, so I probably wnated my private view 😉
February 26, 2010 at 5:38 pm
I love the task pooper idea. That’s something I’ve been wanting for AGES!
I do worry about the network manager icon that’s there, though. It seems to me that long term system status indicators (or really any status indicators) have a completely different scope. Instead, I think gnome-shell’s current message tray design is going in the right direction for the rest of the task pooper’s functionality. (That being the automated portion).
I’m also with others here that this completely obsoletes the desktop folder metaphor and it should die a quick death.
The Intellihide feature in some docks springs to mind as a worthwhile thing. When no windows are above it, the task pooper could appear (maybe some window management conventions could be imagined to strengthen how we get from there to it appearing when the mouse is at the edge).
On a fresh workspace, one could get something like the Moblin home zone, which is a beautiful piece of software. (Shows inserted devices, upcoming appointments, tasks, recent documents in a fashion which is actually organized and makes good use of space…).
That’s enough babbling from me. Thanks for your amazingly uplifting posts, Seth!
February 26, 2010 at 5:47 pm
“Wow, this totally rocks! Workspaces as tasks, archiving workspaces, whacking incoming tasks for later, searching by comparison… You got my vote!”
Very much agreed.
——-8<——8<——8<——8<——-
Now for some criticism:
The two most problematic ideas, which I believe falls into the "will turn out badly and need to be ditched" category, are:
– Apps starting maximized: Desktop screens are getting larger and larger. *Almost no app* deserves an entire 30in monitor! The main exceptions, of course, are movie players and specialized apps (like audio editors, video editors, 3d modeling, etc)
– Absolute touchpad: Seth himself pointed out the main problems: "Too many people may brush it accidentally. Some computers have no touchpad, and some computers have a very small one". Also, it can be confusing and lead to unexpected results *very* easily.
——-8<——8<——8<——8<——-
BUT: I would still like to commend the Gnome hackers for coming up with some amazing ideas. I also think Gnome Shell deserves some love, for innovation.
Just, please, make sure everything that ships is well-tested (usability wise), and that this type of dialog with the community continues to take place — or Gnome 3 could end up exploding like a KDE 4.
Keep rocking!
February 26, 2010 at 7:25 pm
I usually flame mock-ups that have this level of innovation to hell and back, but frankly I am finding your ideas sound and coherent. I will criticise more when we get to something more specific, I promise. 🙂 But so far – congratulations, I am well impressed.
February 26, 2010 at 8:07 pm
Its worth noting that this is for the Gnome 3 series, not Gnome 3.0. Its very important to develop an idea where you’re trying to go, and the call to this is even stronger when you’re contemplating a rewrite. You need to know that once you’re done with the first bit, there are more interesting bits to land instead of defering them as “future problems”. When you do that you often find you’ve walked your way into the end of a dark tunnel to nowhere. Lets look for the light at the end first!
I see this manifest in articles/talk around ‘cloud computing’… many seem to subconsciously assume there will be interesting things invented to use cloud computing once its arrived. I suspect there won’t be all that many.
That said, we do have one curious twist on ‘cloud computing’. Mo from Red Hat came up with this, and I think its rather brilliant. In our world where window/workspace is working (lets say, gnome 3.9), make each workspace a potential virtual machine. Let all your workspaces (all VMs, movable between machines) be mapped into the Gnome Shell switcher. Let you drag windows between them, etc. That’s a sort of cloud computey answer. Would that be useful for concrete tasks? Probably. Exactly which, though, is going to depend a lot on how the world unfolds.
February 26, 2010 at 8:17 pm
We talked a lot about the Gnome shell. We took a straw man poll in the room: by my site, most people thought it was good, many people thought it was promising, and nobody was willing to be the bad guy and say it was bad. So, by the numbers, I’d say its “getting there”. Jon McCann and Jeremy Perry, its principal designers, are clever fellows who are quite eager to find the right solution, not just advocate their solution. That will be critical to getting this to awesome.
The GNOME UI vision is a little murky. No doubt. This is often true inside companies too, but its ‘hid within the walls’. In open source, you’re more apt to see inside the sausage factory. We are still selling eachother on specifics, and the ever present question of “ok, we agree X is great, but who’s going to actually program X? are there volunteers who are eager for this work? are there companies who have available resources and want to fund the development?” etc.
This means that we find lots of mostly good ideas, and compete them against eachother to find agreement on what’s best. Its a social process, and that’s inevitably a little messy. It later in the GNOME 3 cycle than would be ideal for this; but I think you’re going find Gnome 3’s release date may push back to accommodate reality. Still, the gnome shell is on the ground, and its getting better by the day. Better to release when ready than release too early (up to a point, debian).
February 26, 2010 at 8:26 pm
1) We have made “change the desktop fundamentally / delete it” a non-goal. I agree there is redundancy. It tweaks me a bit in my heart when I touch that part of the design. BUT, we don’t have any clear alternatives, having the desktop won’t be ‘bad’ per se, and its good to leave a safety net in case the pooper simply falls through (or falls through for a significant minority).
We already have nautilus written, and it works well. If the pooper is fucking brilliant in the real world, we can always deprecate the desktop / morph it into something different in a later release (if ever). no point in jumping the gun unless removing it would somehow clear things up. I suspect it’ll mostly stay buried, and the home folder, projects folders and some task poopery will be used more, but I don’t see any harm in this specific case in having it back there, good, solid, reliable. Deleting desktop for the sake of the conceptual rush it might give us would be satisfying in some strange way (novelty? power?), but probably not a prudent decision. Enough wild stuff is already on the table that I think has more fundamental importance than having a few clickable objects between the windows and the background photo/wallpaper.
2) tags have ups and downs. from a strict proceduralist perspective, perhaps we can argue that tags are a formal superset of foldres. however, the cognitive models are significtantly different. it is no longer possible to say something is ‘in” something else, there is no sense of containment. from a purely practical standpoint, there are many cases in the interface where the ability to know what something’s “container” or “parent” is is important. For example, just take the case of “Show in file manager” after clicking on a song in Rhythmbox. The song has two labels: “music by boris” and “stuff from jenny”. Which do you show? Do we just add multiple items to the right click? What if a file has 50 labels? Are there limits? These are not insoluble, but its is definitely not true that tags are ‘absolutely’ better. There are many axes on which they are worse.
February 26, 2010 at 8:29 pm
hi greg, this design is brilliant on mice too. same constraints, and its not particularly worse or better on touchscreens. a major design goal for gnome shell is to (future) accomodate very different sorts of input methods, and contexts. we want to build idioms that can be transposed to other screen sizes, and types of input, even if some of the code is different (i.e. it would be foolish to be exactly the same on each, no customization to local strengths is probably sub-optimal, but we want most of our idioms to transpose well to most hardware)
February 26, 2010 at 9:48 pm
I have a couple of small criticisms:
– On the touchpad/touchscreen issue. Don’t assume a touchpad is the same as a touch screen. its actually much harder to hit the edge of a touchscreen than the edge of a touchpad.
– Discoverability, one thing that’s never been really clear for me with ‘pops from side of screen’ UIs is how to make them discoverable. Ideas?
On the non criticism side of things: SETH, IT IS AWESOME TO HAVE YOU BACK! I’ve wanted to meet you since I started in GNOME and it was great to finally get the chance at the UI hackfest. Keep rocking our worlds.
Thanks,
Rob
February 26, 2010 at 9:59 pm
There is only one thing that would convince me to use Gnome for desktop development, and that is GTK documentation. Until GTK is as learn-able and usable as Qt, my preference is clear.
February 27, 2010 at 10:26 am
Really cool and imaginative, would love to use a lot of this stuff. And now for the criticism part…
I get the same impression as with the GNOME Shell design document. I don’t know what’s a fixed part of the design and what’s a good default. Take panel applets in gnome 2 as an example. The OS provides a lot of them, some are there by default, but you can add and remove them as you wish. On the other hand, the panel itself can’t be removed (not all of them).
Will the task pooper be like a panel applet (removable) or like the panel itself (non removable)?
February 27, 2010 at 12:27 pm
The Task Pooper concept is great. Are there any plans to put that in an seperate application?
February 27, 2010 at 4:26 pm
[…] Early peak into Gnome 3’s potential The Gnome 2010 UX Hackfest has been such fun! Today is the last day, reckon we’ll be riotous. It was a wonderful, familiar memory. I know not everyone reciprocates (hi hatemail, bye hatemail), but I missed you Gnome. […]
February 27, 2010 at 11:11 pm
I received 1 st loans when I was 20 and that supported my business a lot. However, I need the consolidation loans as well.
February 28, 2010 at 8:44 am
Maximised windows by default… not sure I like that. I assume it will work something like Ubuntu’s netbook remix – something that works great on my netbook’s small (1024×600?) screen.
However, it works terribly on the much larger (1920×1200) screen of my desktop box, and suspect it would be dreadful on the dual-head setup I use at work. I rarely run anything maximised in either of those environments, and would find it most annoying to have to un-maximise windows every time I start an application…
February 28, 2010 at 10:39 pm
[…] was some exciting stuff at the hackfest. Seeing Seth’s ideas emerge was really cool. I also like the look of the GNOME tweaks app and the new control center is […]