Ubuntu Mobile Looks At Qt As GNOME Alternative

Derwent sends along a Computerworld piece which begins: "The Ubuntu Mobile operating system is undergoing its most radical change with a port to the ARM processor for Internet devices and netbooks, and may use Nokia's LGPL Qt development environment as an alternative to GNOME. During a presentation at this year's linux.conf.au conference, Canonical's David Mandala said Ubuntu Mobile has changed a lot over the past year... 'I worked on ARM devices for many years so a full Linux distribution on ARM is exciting,' Mandala said, adding one of the biggest challenges is reminding developers to write applications for 800 by 600 screen resolutions found in smaller devices. 'The standard [resolution] for GNOME [apps] is 800 by 600, but not all apps are. For this reason Ubuntu Mobile uses the GNOME Mobile (Hildon framework) instead of a full GNOME desktop, but since Nokia open sourced Qt under the LGPL it may consider this as an alternative.'"

Comments

Full 'nix for arm?

[serviscope_minor]

There's already a full 'nix for ARM complete with working packaging and so on, in the form of OpenBSD, just in case anyone has forgotten it. Also, the developers need to be reminded that screens are 640x480 on small devices, not 800x600. It would start if they got out of the habit of using excessively lavish button bars with enourmous, heavily padded buttons.

Anyway, it would be nice to see a proper "full" linux distribution. I'm not much of a fan of the special PDA ones since they're cut down. Then again, I'm not much of a fan of ubuntu either, but I appreciate that (say) Arch isn't to everyone's taste.

Re:Full 'nix for arm?

[Mprx]

Padding makes clicking on the buttons faster, as explained by Fitts's law. I don't want my usability compromised because some people are using impractically small screens.

Re:Full 'nix for arm?

[chromatic]

Keyboarding may be faster occasionally, but you'll be surprised how often mousing wins.

I use Vim all day, almost every day. Using a mouse and a word processor is very much not faster for me.

Re:Full 'nix for arm?

[Mprx]

Keyboard shortcuts are not amenable to muscle memory, as the muscle movement differs depending on the previous shortcut. Returning to the home position between each keypresses allows muscle memory, but I'd be very surprised if it were enough to compensate for the movement inefficiency. Consecutive strings of keyboard shortcuts can be memorized by muscle memory (as with typing whole words), but if you use a string of shortcuts frequently enough to memorize in this way it would be better consolidated to a single shortcut.

On the other hand, mouse gestures or pie menus require the exact same movement each time, so are highly amenable to muscle memory.

Re:Full 'nix for arm?

[Arker]

Keyboard shortcuts are not amenable to muscle memory, as the muscle movement differs depending on the previous shortcut. Returning to the home position between each keypresses allows muscle memory, but I'd be very surprised if it were enough to compensate for the movement inefficiency.

You're obviously not a touch typist. That's absurd. A practiced touch typist on a decent keyboard can select a paragraph for manipulation in about the time it takes you to get your hand from the keyboard to the mouse, let alone actually using the mouse to select the paragraph and THEN move the hand back into home position.

Re:Full 'nix for arm?

[Just Some Guy]

Unfortunately for that comparison, mouse input is highly srialized while keyboarding is very parallel. I only have one arm to manipulate a mouse but 10 fingers to manipulate a keyboard.

Re:Full 'nix for arm?

[bhirsch]

Let's see...
Open a new browser tab (winner: keybd)
Close a browser tab (winner: keybd)
Go to a history/bookmarked URL (winner: keybd)
Navigate forward/back in the history (winner: keybd)
Click a link (Tie. I generally hit / and start typing the link text, hit escape, then enter to visit the link. Sometimes moving the mouse over the link is faster.)
Run an application (winner: keybd)
Cut/copy/paste/save/print/quit/etc (winner: keybd)
Scrolling via arrow, pgup/dn, home/end vs. wheel (winner: keybd)
Switching between applications/windows (winner: keybd)
Change a font (I will give this one to mouse)

Sorry, but between tab to move between elements (a must have for any application I use) and standard keyboard shortcuts (control/alt/command/shift modifiers), I have zero desire to use a mouse and rarely do. I don't get how taking tme to move one's hand over to the mouse, moving the cursor over a button and clicking could possibly take less time than two simple key presses (my fingers remember where ctrl and w are, but my hand does not remember how to navigate over the little "x" on the tab).

While mouse gestures are certainly nice, they are, in my experience, far more prone to inaccuracy compared to a key presses. Mistakenly closing a browser tab happens far more often with gestures than the keyboard.

The only time I find myself consistently using a mouse is in OS X Finder, which just doesn't play well with the keyboard. In such situations where the mouse is generally faster, I will usually opt for a keyboard solution (eg, using mv, cp, mkdir, etc from a shell).

Much of this may have to do with my late switch to Windows from DOS and work generally keeping me in a terminal window, but I still maintain I can use a keyboard faster for most common tasks than anyone with a mouse.

Re:Full 'nix for arm?

[Arker]

How many people browse the web with only a keyboard? How many people edit pictures/music/movies with only a keyboard?

How many people that *think* they've been educated on computer usage never learned to type to begin with? How many keyboards these days are so shoddily made they are effectively useless for those of us that do know how to use them, clearly designed for use by hunt-and-peckers only? How many computer programs just assume the current answers to those facts and dont bother to even consider exceptions to the rule in their design and implementation?

How many people use a dozen programs in a different day, each with different keyboard shortcuts that would need to be muscle-memoried?

How many programmers cant be bothered to use the standard shortcut keys of the system(s) they target with their programs? Particularly in X, unfortunately, many even deliberately remove the ability of the user to *set* those keys to the standard they are used to? (Yes, I'm looking at you Gnome.)

Much like "touch typing is just faster" falls apart once special characters and numbers are introduced in large quantities

Actually, it doesnt. Numbers are no problem at all, as long as you do them often enough to keep your fingers in practice. Special characters can be problematic, yes, but they dont have to be. Windows makes it bloody impossible, but with only a little practice I found I can type most special characters without ever breaking stride on a mac with a standard English keymap. (Assuming, of course, someone's already done the sensible thing and replaced the pretty but unusable little toy Apple pretends is a keyboard in the trash and gone to the trouble of fitting a decent one on the machine.) X keymappings are absolute horror to setup in my experience, but at least they're editable.

or when one is not transcribing but rather writing original text (where regardless of typing technique, most people can press keys considerably faster than they can think of keys to press)

I dont see how you think pausing to think has any bearing on the subject. Sure, you pause to think, then you type, then you pause to think... pausing to grab the mouse and run it around the screen still adds time and more importantly it breaks the concentration on the subject matter, so this doesnt help the case for the 'do everything with the mouse' argument at all.

"keyboard shortcuts are faster" falls apart when you start using apps that are unfamiliar, unpredictable or simply make more sense with a mouse

Absolutely true. However this, again, does very little for the argument. Just how often do you need to learn a new program? For most people it's relatively rare, you spend far more time *using* programs you know than learning new ones.

Furthermore it would be far more rare to need to learn new programs if we didnt keep fixing things that arent broken - i.e. replacing perfectly good, functional programs we already know how to use with new, buggy, bloated junk that gets pushed on us for marketing reasons. For most people that use a computer today, the thing is primarily a bloody typewriter. They spend most of their time in 'Word.' Yet every couple of years they have to get a new version of word, with a new learning curve and new and higher hardware requirements just to keep doing what they were doing all along - using the computer as a typewriter.

Finally, on a decently designed system, the commands that are 'common' to many programs should have system-wide standard keystrokes anyway. Even windows gets this concept - ctrl-s is always supposed to be save, ctrl-c copy, etc.

Of course, the best option is just to give people both and let them use whatever they want. We're probably only talking milliseconds here anyway, so perception is all that really counts.

I

Re:Full 'nix for arm?

[gknoy]

My reflexive control-X control-S (when I'm /not/ in Emacs) would beg to differ. It's gotten so ingrained that I'll use those shortcuts in other things as well; similarly, shift-delete will delete a whole line in my Other Editor, which annoys the heck out of me, as i'm expecting to cut what I have highlighted. ;) So, I'm pretty sure muscle memory works quite well with keystrokes ... whereas when mousing, I am always looking at where it's going. Perhaps with a tablet it'd be easier.... but even then I had trouble and needed to look.

Re:Full 'nix for arm?

[Lord Bitman]

I use vim exclusively, but counting out how many words I want to move, followed by typing [ESC(which is the "capslock" key)] 17w, or even just hitting "w" or "W" repeatedly while tracking with my eyes wherever the cursor has ended up /this/ time based on whatever is considered a "word boundary"...

has _ALWAYS_ been slower than moving my hand to the mouse and clicking. and USUALLY been slower than just holding down an arrow key, especially if using an editor which sanely handles the use of arrow keys to move between lines on the screen.

Yeah, I realize that if I think for 20 seconds, I can come up with a sequence of commands which will do the specific task I require, and will require I know which "physical" line of the file I'm on.. but I really don't give a shit if I could otherwise do what I want in under two seconds.

I'd be an emacs user if it didn't blow so much. (still waiting for it to act as advertised...)

Re:Full 'nix for arm?

[Anonymous Coward]

Keyboard shortcuts are not amenable to muscle memory.

O rly? :wq

Re:Full 'nix for arm?

[Curien]

Open a new browser tab (winner: keybd)
Close a browser tab (winner: keybd)
Navigate forward/back in the history (winner: keybd)

Those are all arguably faster with gestures.

Go to a history/bookmarked URL (winner: keybd)

With most browsers, I'd say this is true. With Opera's Speed Dial, it becomes a matter of whether you've memorized the name of the site (or bookmark). Just open a new tab, and you see a (fairly large) picture of each of your bookmarked sites. Click on the one you want to visit (or Ctrl+Number). This method makes both the keyboard and the mouse faster (and near-equivalent, IMO) by removing the dependence on the user's memory for speed.

Run an application (winner: keybd)

Only if you've memorized the name of the command. What you're really saying here is that searching for a program is slower than already having it memorized. I guarantee you that clicking on a desktop shortcut (and you have your dekstop set for single-click mode, right?) is faster than executing the key combo, then typing a word, then hitting enter.

Cut/copy/paste/save/print/quit/etc (winner: keybd)

Cut,copy,paste is a process, not just a key combo. I've found that the process usually works best when you use both the keyboard and the mouse. The mouse is better for selecting large blocks of text and getting the cursor to the general area where you want it. The keyboard is better for precision movements.

Scrolling via arrow, pgup/dn, home/end vs. wheel (winner: keybd)

Let's have a race. We each have an identical several-hundred-page document. You use the page keys, I'll click on the scroll bar. First to the middle wins. Also, your mention of the scroll wheel belies your inexpertise with the mouse. The correct tool is the "middle-click drag" auto-scroll feature.

Switching between applications/windows (winner: keybd)

The advantage of the keyboard over the mouse is its parallel nature. Alt-tab is an inherently serial process, so it eliminates the advantage completely. (If you happen to know that the window you want is the previously-active window, then sure, alt-tab is inherently faster. But that's an incredibly special-case scenario.)

Click a link (Tie. I generally hit / and start typing the link text, hit escape, then enter to visit the link. Sometimes moving the mouse over the link is faster.)

That method is incredibly limited. You can't click on buttons or images. You lose context as you type since the screen's jumping around thanks to search-as-you-type. It fails miserably is the link text is repeated multiple times (eg "reply to this" on slashdot). IHBT. IHL. HAND.

While mouse gestures are certainly nice, they are, in my experience, far more prone to inaccuracy compared to a key presses. Mistakenly closing a browser tab happens far more often with gestures than the keyboard.

That is not at all my experience. Perhaps what you mean to say is that you are prone to imprecision with the mouse (especially if the next item is true for you). From this and other assertions of yours, the only conclusion I can make is that you're not very good with the mouse.

Why just netbooks?

[DeHackEd]

Sure the big blocky feel of pretty much every window manager out there sucks on my Eee, but this is one reason I stick with GTK+ 1.x. I don't have a 1280x1024 monitor just so I can see the same material I could see on an 800x600 10 years ago but with cleaner rounded edges.

And I have the bigger Eee. 1024x600 resolution, and some dialogs don't even fit on the screen.

Re:Why just netbooks?

[Just Some Guy]

I'm a huge fan of Netbook Remix and "maximus". The former provides an awesome launcher sort of like the Eee's default interface but way better, and the second provides fullscreen, borderless windows. You might see what you think of it.

Yah for the LGPL

[QuantumG]

For too many years the GPL has been killing adoption of Qt. That's a fact. Maybe it shouldn't have. Maybe people should be willing to be dictated to on what license they can use for their product because they dare to use the Qt framework. Maybe that's your opinion.

Of course, now that so many people are piling on-board to use Qt thanks to the license change, I wonder how many of them have actually bothered to read the LGPL. My favourite part is section 4.

You may convey a Combined Work under terms of your choice that, taken together, effectively do not restrict modification of the portions of the Library contained in the Combined Work and reverse engineering for debugging such modifications,

Yeah, didn't see that did ya? Almost every boiler plate EULA includes a clause prohibiting reverse engineering and I wonder how many have not been updated to comply with the LGPL (thankfully a lot of us can just ignore these restrictions as the government in our part of the world recognizes reverse engineering as a right that cannot be contracted out of).

I'll be looking for violations.. just for shits and giggles.

Re:Yah for the LGPL

[QuantumG]

Note the words "reverse engineering". If you forbid reverse engineering, as typical EULAs do, for any purpose, then that is forbidding reverse engineering for debugging modifications to the library. So they at least need to modify their EULA to permit reverse engineering for this purpose. And it also means they can't put any anti-debugging tricks in the application, because it will interfere with that reverse engineering.

Re:Yah for the LGPL

[kripkenstein]

Yeah, didn't see that did ya? Almost every boiler plate EULA includes a clause prohibiting reverse engineering

It is my understanding that you can reverse-engineer the LGPLed library, but nothing else. And that isn't much of an issue: the LGPLed part should come with the complete source (or equivalent), so being able to debug/disassemble binaries for which you have the source isn't a big deal. In fact the purpose of this clause (again, to my understanding) is simply to be able to ensure that the source of the LGPLed library indeed corresponds to the binary of said LGPLed library.

So yes, EULAs might need to change - good point. But the change should only be "you can reverse-engineer the LGPLed library X, but nothing else," and that would appear after "you can get the source code to the LGPLed library X from here."

Bottom line, if you are ok with using an LGPLed library - and many otherwise GPL-averse corporations are - then this should not be an issue.

Re:Whiney complaints (send to /dev/null)

[Anonymous Coward]

>@ Canonical
>Debian system

WTF?

Re:Too big

[Arker]

I saw a MIPS based netbook for about US $150 a week or two ago. Trying to remember where.

It strikes me that the best way to improve usability of X apps might be to send these little babies off to as many developers as you can find - and then preferably putting a gun to their heads and forcing them to try and use their apps on them.

The gun to the head part, of course, is tongue in cheek - but wow! seldom is such a bad idea so tempting.

Re:Whiney complaints (send to /dev/null)

[Anonymous Coward]

If you want to build from source you should be competent enough to figure out how to download it; it's really not hard. Otherwise, let your vendor do the packaging for you. Most Linux distributions make it so you don't have to care about building anything; and the BSDs make building easy anyhow.

Granted, if you are building from source, Qt's build method is mildly stupid compared to the (end-user) ease of autotools or CMake. But really, if you're just wanting to run programs, let your vendor take care of it all for you.

Re:Whiney complaints (send to /dev/null)

[tftp]

WTF? What is that all about to someone who just wants to run an application that uses Qt?

If you want only to run a Qt-based app then you do not need to do anything except to install the application. It should install the Qt runtime libraries for you.

Why the hell am I even looking at this when I just want to run an application?

A good question indeed :-)

If you want Qt widely used you need to make it easy to get and install.

If you are a developer then installation of Qt is the least of your worries. If you are an end user then, as I said, you should not install Qt at all.

AT LAST

[NekoXP]

Hoo-fucking-ray!

At last some common sense..

Qt outstrips GTK/GNOME just as a GUI toolkit and a bunch of middleware, even before you start thinking about stuff like KDE.

The only thing stopping it's use - at least in the strange mix of preinstalled Linux distributions on standard hardware - was that weird problem of having to have every one of your developers buy a license just to run their app - on a Dell for example - if their license was even slightly incompatible. That was a real turn-off if you were a hardware company wanting to take advantage of open source and build communities around open source software.

I'm glad that so soon after Nokia announced the LGPL relicensing, people are taking notice of what is quite obviously a far superior middleware solution than the GTK/GNOME nightmare, and considering developing solutions that work because of code quality and wealth of features, and not *just* because it's GPL.

Re:Whiney complaints (send to /dev/null)

[Draek]

I know my comment will be burried for saying this, but this kind of crap is what we all know is wrong with open source software. The front end delivery is done by geeks and bean counters who don't actually use the products as end users.

You may notice the fact that QT was originally developed by a commercial company, Trolltech. You may also notice the fact that since, until lately, they sold commercial licenses for the same software they licensed as GPL, practically all contributions to the 'main' branch of QT were done by Trolltech (and now Nokia) employees. Therefore, if anything, this proves the failings of cathedral-style development, of which closed-source is the biggest exponent.

Ohh and also, being a person unwilling to use pre-compiled packages to be able to use a library you do *not* plan to use as a developer puts you amongst the minority of a minority of a minority of users, therefore do not be surprised if Trolltech/Nokia doesn't care about you at all.

What *is* happening?

[dhasenan]

Ubuntu Mobile is not switching to Qt.

Ubuntu Mobile is not even considering switching to Qt.

At some point in the future, they may consider switching.

How is this news?

Nokia open sourced Qt under the LGPL

[John Hasler]

Qt was already Open Source, of course, under the GPL.

Qt is *NOT* Gnome alternative itself

[hubert.lepicki]

OK, Qt isn't even close to Gnome in terms of being a desktop environment. In fact, it isn't a desktop environment at all - so it can't be alternative to Gnome. It can be alternative to GTK, which is underlying library for Gnome. What I guess is the case - Ubuntu might look for KDE as an alternative to Gnome desktop, or create something new based on QT that'll fit more on small screens.

Only in your dreams

[Anonymous Coward]

"For too many years the GPL has been killing adoption of Qt. That's a fact."

It's only a fact in the fantasy world you inhabit.

Without GPL for Qt, KDE would not have gained the foothold it has, and Qt would still be a niche player.

Nokia doesn't care about GPL, open source, or free software. It is just doing what it feels is necessary to undermine its rivals.