For handheld computers, there are basically two UI design patterns. The first one, the keyboard UI, has been almost invented and immediately perfected by Psion in their Series 3xx palmtop computer line. It really would take far too long to explain how the Psion UI works (although, as a longtime Psion afficionado, I’m sorely tempted!) but let’s suffice by saying that the keyboard-based user interface of those computers was simple, intuitive (as far as any computer metaphore can be intuitive),easy to learn and as frugal as possible. That last property means that the interface was designed in such a way that any user action could be performed with as little input as possible (I’d almost dare to claim that Psion managed to find the path with the least possible inputs for everything — I sure like to see proof of the opposite).
The second user interface for handhelds is the tablet UI, for touchscreen devices. Again, I claim that the perfect tablet UI already exists: It is the Newton user interface. For those who want to know more about it, see this page, written by Sean Luke, one of the legends of the Newton user community. Everything he writes in that article should be compulsory reading for anyone even thinking about improving the Itablet (that’s my personal moniker for the “Nokia Internet Tablet”. What can I say? I hate lower case “i”s). What I’d like to do here is focus the attention on just a couple of key features of the Newton (and therefore the ideal tablet) UI.
The first one is the general screen layout. If you look at a couple of the Newton screenshots on Sean Luke’s page, you might notice something rather odd about their layout compared to Hildon’s. The top of the Newton screen is remarkably devoid of menus, boxes or indeed anything else that the user might need to select while using the main program. Instead, Newton’s menus are at the bottom of the screen! This is not because of some strange whim Jobs had (because Jobs never had anything to do with the Newton), but it was done for a very specific and good reason.
On a tablet the user interacts with the system via the touchscreen with a stylus or finger (no fingers on the Newton, but you get the drift). The reasoning was this: With top drop-down menus, everytime the user interacts with the menu, her/his hand will obscure the screen, which will make handling the device needlessly more difficult. The Newton puts most of its menus on the bottom, with options sliding upwards, so that your hand is never in the way of your work. In fact, the only options you’ll find on top are the menus for filing and exporting; stuff you’ll do when the work is done.
I don’t think Hildon will ever be adapted to the bottom/slide up menu style of the Newton, mainly because I don’t believe Nokia can’t be bothered but also because they have committed themselves to the mishmash that is Hildon.
But there may still be hope. The KDE port is rapidly maturing, and KDE is a very adaptable environment. I wouldn’t be surprized at all if it turned out to be possible to shift all the menu bars in KDE apps to the bottom of the window (someone please prove me wrong on this?).
My second point about the tablet UI is about entering text. Hildon has three methods, and they all suck. Virtual keyboards on a tablet suck anyway, and Nokia’s implementations have brought along their own suckiness. The stylus keyboard is just too small to be workable for anything but extremely short entries. And the large, finger keyboard has some annoying quirks that rapidly make one want to gauge their own eyeballs out to throw them at the screen (one example: it doesn’t go away on its own after you’ve typed and entered a command in an xterm. You have to click it away to see if your command was executed).
Ah, but Hildon has handwriting input as well, right? Yes, printed character recognizing in a dedicated sub-window. Don’t get me wrong, handwriting is, IMHO, the best possible text entry system for a tablet computer. It’s not the fastest (a trained typist can type many times faster than a handwriter, with the (possible?) exception of a dedicated stenographer), but it’s the only input method where the user’s eyes see both his hands and the screen.
Provided the HWR (handwriting recognition) is fullscreen and as close to natural writing as possible. Newton had both: You could print your letters or you could write in cursive (the latter method requiring the system a little getting used to your handwriting style) and you could write anywhere on the screen. It is a joy to use and to this date I know many people (myself included) who keep a Newton just for taking notes.
Obviously, we cannot have Newton’s HWR on the Itablets, right? Weelll… I thought so too. There is a company, PhatWare, who sells HWR software for Windows and Windows Mobile, CalliGrapher and PenOffice. They use one of the core HWR engines, ParaGraph, of the Newton (the other one is Rosetta, and it found its way into Apple’s Inkwell technology, where it is sadly withering away) and it really works almost as well as the original stuff. I have a copy of PenOffice installed on my Windows tablet PC and the HWR recognition is indeed as good as the one on my Newton (of course, the Windows UI is utter crapola when attacked with a stylus, but that’s another story). So I sent PhatWare an email, asking them is they would consider porting their products to Linux and I got a reply.
This is what they wrote:
Dear Ms. Jansens,
I apologize for the delay they have not responded till today.
Apparently, there is a Calligrapher/PenOffice engine version for Linux, but it is not in retail. It is only sold along with devices, via OEM.
Since I am not familiar with this device market, I can not tell you off hand what manufacturers they were talking about, but I will try to find out and get back to you.
Thank you for your interest in Phatware products!
(I let the “Ms.” Jansens slide because I was so excited. The “they” in the mail is the Marketing and Development department)
This basically means two things:
- There is a “real” HWR solution for Linux out there. It’s not open source, but given the total absense of OSS HWR solutions, I’d be really quiet now if I were an OSS fundamentalist, and
- it’s completely and utterly out of reach for us users. Only Nokia could possibly have enough of a strongarm to get a license from PhatWare, which means that it’ll probably never happen.
That’s me ranted out for now. Conclusions:
- The Itablets screen layout is so wrong for a tablet device.
- Nokia could have done some research and procured themselves a license for CalliGrapher for the Itablet, which would have made at least one annoyance go away.