The Fedora 14 Beta was released today, but as a Network World article points out, it “will be the first Red Hat supported distribution to let users choose MeeGo as their desktop.” This new release will also include the Sugar interface, intended for netbooks, and “will also be the first version to fully incorporate Red Hat’s VDI (virtual desktop infrastructure), called SPICE, or Simple Protocol for Independent Computing Environments. SPICE will allow Fedora to host virtual desktops that can be accessed over a network.”We’ve all been waiting a long time for the desktop Linux revolution that never came. But then while we were all playing with our iPhones and not paying attention, the Linux revolution snuck through the back door and took us by surprise. Microsoft initially outmaneuvered the early Linux netbook gains, but then Android came on the scene and made that defeat look small in comparison to Linux’s seeming triumph in the handheld space.
But I think the lesson that we’ve all learned is that a computing platform revolution isn’t really about the operating system anymore, but about the ecosystem. Both iOS and Android have succeeded because they’re part of a movement to create a vibrant community of users, vendors, app developers, and web service providers that provide a useable, complete package. If Linux is going to make a meaningful splash in this new tablet space, whether it’s Android, MeeGo, or something else, a solid OS foundation is only going to get you so far. I’m not sure whether Red Hat can bring anything to the table to help with the ecosystem-building, and I’m not sure that Intel (a MeeGo sponsor) can either (or for that matter, Nokia, the other MeeGo backer). They just don’t have the kind of juice that Apple and Google have.
Fedora could be the platform upon which a Linux tablet empire is built, or it could just as easily amount to naught. Who will pick up the banner and advance the ecosystem?
Most tablets present multitouch as a big feature for interaction. And tablets need many specific features to be implemented and prepared.
I don’t see why Fedora is selected as a TabletOS.
At least ubuntu has the new Unity interface designed for a touch interaction with big buttons on the left.
The uTouch stack is also developed for ubuntu and released as opensource.
The problem is that making the UI of the OS touch-friendly, whether with Unity or MeeGo, is only a very small part of the job… You have to remake the interface of all applications. We’d have to push everybody to remake their interfaces: LibreOffice, FireFox, the Gimp, Inkscape, Ardour, Blender, MPlayer, insert-a-popular-open-source-app-here, etc. Not impossible, but open-source UI sucked for years [/troll], why would it change? (maybe yes, it will, because Qt and Gnome matured, and well-coded apps have a good separation between the UI and the working bits, but that’s also a question of mentality, and ressources).
That’s true.
UI of most opensource tools sucks.
Just look at Gimp, they have accepted to change the UI only later. (And everybody is still waiting for the next version with 1 window)
If you’re trying to find elements of photoshop’s workflow in an UI that’s not photoshop’s, that UI surely sucks
What I mean is that this judgement of GIMP’s UI is highly subjective. When you’re experimented with it and get the basic workflow, you understand how powerful it is. Flexibility, discoverability, readability, quick access to a lot of functionality, pen tablet friendly widgets, much better curve editing… These are all good sides of the GIMP UI.
I think that like most people who unilaterally hate gimp’s UI, you do it for the same reason as I hate Photoshop’s : not being used to it.
Edited 2010-09-29 06:17 UTC
Correct. People whining about Gimp not being MDI, etc., really should whine about their own inability to learn. It’s a much more serious problem.
The question is not to imitate photoshop or not. Quit that line of thinking. Design & ergonomy are NOT completely subjective: these are fields that have been under heavy research for decades. Clutter is bad, wasting space is bad, etc. Have a look at this blog post:
http://humanized.com/weblog/2007/10/05/make_oss_humane/
I know that very well, as I’ve been studying them for some times. BUT it takes some time to fully grasp the good and bad sides of an UI. Spontaneously, when you see an alien UI that does not respect the conventions you’re used to, you have, as a human being, a natural tendency to notice what is absent rather than what is present.
E.g. when I see photoshop’s UI, my first thought is “uuuugh why do I have to open silly menus when I want to access my tools ? And why are those buttons so tiny ? I can’t click them with a pen tablet ! And who had that crappy idea of putting everything behind non-discoverable modifier keys and context menus ?”. I suppose that if I took the time to learn it, I’d end up discovering some advantages in most of their UI choices.
Not all UIs are equal. But contrary to popular belief, you cannot measure the quality of an UI at first sight without in-depth knowledge of usability rules and the targeted user, or in-depth use… In books about ergonomy, you discover why who your user is matters more than generic rules supposed to apply to all human beings.
Your original point was that GIMP’s UI is universally bad (“sucks” in your own words). My point is that this is wrong. GIMP has its issues, but it’s quite good when you’re used to working with it
Clutter and wasting space are opposite extremes
PS tries to put lots of tiny widgets in few room while Gimp takes more space.
But consider that most computers have wide screens nowadays. As much as they’re crap for actual work (I can develop this point), they have a single advantage : more width for tools. The GIMP tools are good for tablet use, at the expense of that extra width of widescreens that you don’t need anyway. It’s not about wasting space, it’s about using it properly !
It’s large (and old, too). Extract the part of it which you want to use in your argumentation.
Edited 2010-09-29 12:01 UTC
I was especially thinking at “MISTAKE 2” and “MISTAKE 5”.
Mistake 2 : Well, I’m not sure that I agree with him on this one. He seems to disagree with use of the layer/selection/path paradigm in GIMP because it’s too complicated and he thinks that it was only introduced for photoshop compatibility.
But there’s a reason why photoshop is using this paradigm : though difficult to learn, it’s powerful in the end. AFAIK, a UI paradigm for these tasks that’s at the same time simpler and just as powerful has not been found yet.
Maybe the author means that GIMP should switch to a simpler paradigm (like that of MSpaint) at the expense of image editing power. But then I disagree : casual editing is not the main usage pattern envisioned by gimp designers, so they should remain consistent with that, even if it alienates casual users. Alas, software design is about choice.
Mistake 5 : I don’t think that GIMP devs don’t listen to their users. I use it since 2.4, and there have been great improvements in usability in successive releases… Let me quote some things.
-2.6 : Tool windows (puts an end to the window focus hell), modifiable selections (VERY handy), menus moved on top of the windows in order to make them simpler (the older ones were quite obfuscated), better lasso (polygonal selection on 2.4 was awful)
-2.8 : Easier alpha to selection (it’s often done, so it should be optimized), layer grouping, much better brush dynamics, single-mode windows for those who like it (and can be disabled for people who like it the other way like me), better brushes included by default…
Edited 2010-09-29 13:56 UTC
As a person who uses Gimp to do my basic image manipulations, I get messed many times with losing the focus of the tool window.
If it is a single window, I never got this problem.
Do you have any scientific studies approving that gimp UI (or similar) is better than photoshop one ?
(Adobe has many research centers, so I don’t think they like providing their users crappy UIs)
Look at Inkscape, it has a very very usable UI. You can’t lose focus on what you are doing anytime.
Since gimp 2.6, tool windows are used. They don’t loose focus, don’t appear in the taskbar, and you can hide them with TAB. Still think that the problem is using multiple windows ?
And you ? Do you have scientific studies proving that photoshop’s UI is better than that of gimp ?
My point is not arguing that GIMP is better. It’s that although different, it is very capable, too
Adobe can sell whatever they want, people who need the software for work will buy it anyway. It’s called a monopoly. Photoshop CS2-CS3 is a good example of that : no major evolution except that of hardware resource usage.
Funny, I hate inkscape’s UI too
First becaus I think it’s a waste of space, second because I heavily dislike having to move my mouse on kilometric distances before I can finally click my target. They should have learned about the second part of the Fitts law (that about distance). And don’t get me started about those horrible incomprehensible toolbars on the top.
This shows best the importance of being used to an interface, IMO.
Edited 2010-09-29 12:14 UTC
Bingo. Like other poster stated, both Gimp and Photoshop are good on their own rights. As a freelance designer, I mainly use Gimp for budget reason, lower footprint (Adobe CS product are power hungry). While Photoshop is likely complete due to CMYK support for printing, Gimp lack of support can be completed with Krita (thanks to the ability to import xcf and convert into CMYK).
For me, an excellent design can use any software image tool regardless their quirk.
I think that like most people who unilaterally hate gimp’s UI, you do it for the same reason as I hate Photoshop’s : not being used to it.
Just to stick a note here, I am one of those people who are used to GIMP’s UI, not having used Photoshop ever, yet I still absolutely hate how GIMP looks and behaves. It just is a horrible mess.
That’s why I said “most” ^^
Well, that’s kind of the point of MeeGo, isn’t it? Not to take existing apps off the shelf and plop them on a tablet, which worked so well for Windows over the past 5 years [/sarcasm], but to code up touch interfaces in QT4 for apps and release them as MeeGo apps.
It’s the only sane approach to touch applications.
Yes. But how will that articulate with Fedora and OpenSUSE also implementing MeeGo ? Will the MeeGo repositories be accessible? Will the apps be compatible despite the differences in implementation and library versions? Or will we have completely different ecosystems between the “official” netbook MeeGo (managed by Intel?) and others?
Woah, we seem to have veered from the tablet UX to the netbook UX in your comment. Let me straighten my glasses. Ok, ready.
Intel’s ecosystem for netbooks is called AppUp, which supports both Windows and Meego (and legacy Moblin). It’s reasonable to assume that a different Linux distro using the Meego UX on x86 processors such as Atom would support both AppUp and the native distro repository, similar to how Ubuntu has for years supported both Universe and the (not terribly well-known) click-n-run warehouse.
Back to touch apps – smartphones and tablets generally get their apps from the operating system’s app store, so for Meego that’s probably Nokia’s Ovi store. A tablet running (say) Ubuntu with Meego UX would probably support both Ovi and Universe, just as Maemo on N900 supports both Ovi and the native Maemo repository.
In both cases, though, Meego is far more open than iOS or even Android. You can load any apps you want from any source. App stores and repositories just make it much simpler for “normal” people.
Yeah, sorry for the mix between netbook and tablet (although their share the same problem : everything needs to have big UI elements, for netbooks to be readable, for tablets to be touchable).
But those apps rely on libraries and API; won’t there be a widening gap between MeeGo-pure, and MeeGo-on-top-of-SUSE-or-Fedora? If, next year, PulseAudio is replaced by a new sound API (just an example, could be anything else – X.org replaced by Wayland, or something I can’t think about, any of the hundreds of libs), Fedora will certainly switch to it, bleeding-edge as it is, OpenSuse will follow a little later, but how would the AppUp/Ovi apps deal with it?
Edited 2010-09-30 21:35 UTC
Good question. I’m not enough of an authority to know for certain, although my understanding to date is that QT4 abstracts the underlying systems. So when writing Meego apps, you don’t write to PulseAudio or Wayland or (shudder) X.org – you write to QT4, period.
That’s how QT4 apps run on Windows, Macs, desktop Linux, WinMo 6.5, Meego, Maemo and Symbian with basically a recompile. Abstraction.
I’m not claiming that you won’t want to make *any* code changes, you understand – a desktop QT4 app won’t “feel” right on a Meego tablet. Just saying that your concern about “hundreds of libs” being unique depending on which platform hosts the Meego UX’s, and requiring significant code rework to make them compatible with each Meego instance, is probably off-base. Probably. And given (say) several Atom-based netbooks running Red Hat with Meego, Suse with Meego, and Intel’s Meego, binary Meego apps should run just fine on all three just like Gnome apps or KDE apps run unmodified on Red Hat, Suse and Ubuntu for x86 today.
But we’ll know once we have several distinct Meego products in the market, won’t we?
You may be reading a bit too much into this. As we keep saying, Fedora is a community project. Meego in F14 is a *Fedora* feature, not a Red Hat one. This is actually being driven by the community, in large part by Peter Robinson. Some of us at RH are interested in it too, but it’s no kind of official Red Hat ‘thing’.
It’s also somewhat on the wire for F14; there’ll be a Meego package set of a kind in there (in fact, Maemo was in F13 already) but it’s probably going to be a little raw. Especially using it on a tablet is not likely to be practical without a lot of knowledge. The biggest hurdle here is the kernel; most tablets are ARM devices and it’s not yet possible to build a single united kernel for all ARM devices, each device needs a dedicated kernel. Various people are working on resolving this but it’s likely to be in the F15 or F16 timeframe. That’s about when I’d start looking for really interesting stuff in terms of strapping Fedora onto your Galaxy Tab or whatever. It’ll be a bit of a wait.
dell
Edited 2010-09-29 05:38 UTC
I remain sceptical about tablets in general (netbooks are far more useful at the moment), so the big news for me was that they’ve ditched systemd in F14 beta and won’t include it in F14 final either. We’ll have to wait until F15 next year to see systemd, which is a shame because I’d rather have several seconds taken off my boot time than worry about Meego/Sugar/etc.
Maybe because there where still quiet a few people with problems with Upstart on Ubuntu ? Thus showing parallel startup might need a bit more work before they release it on the general public.
AFAIK, systemd was just not ready so deciding to pull the plug is probably the best thing to do in the circumstances.
As somethnig like systemd is at the core of the whole boot process then it has to be right otherwise the system is as much use to the end user as a Windows Box displaying the BSOD.
I already had planned to test it out:
http://www.spicespace.org/
I think it could be really interresting.