Canon Cat Leap Keys
Skipping to the end – the answer is I don’t know – but I think it’s worth considering!
I’ve recently been considering if point and click ‘desktop’ / ‘window’ interfaces stifle innovation because they seem to be all encompassing and universal on all modern operating systems. Even though we have some desktop variations the metaphor still seems to be pervasive for most everyday tasks. It may be that this is not a problem, just as we use a desk for most writing tasks and a pen and paper and a filing system – these are part of everyday life and are no longer questioned – these tools are familiar and comfortable and so are now just ‘invisible’ to us. However, what if our reliance on the desktop is misplaced – what happens if there are better ways of interacting with applications and their data. We can see that systems exist – like zigzag – but these is pretty complex and difficult to use and implement, showing no traction outside research.
I’m, not sure if a move back to text based systems or at least a complimentary system would not be a good thing. Point and click has a tendency to make developers in some way sloppy. As interfaces are visual and point and click is all that is required interfaces have become ever complex without much complexity being hidden behind a more simplistic GUI.
Things are not all bad – indeed the Web maybe where most of the novel interface work may be occurring. As an example take the Facebook status update which allow the user to type the @ symbol before a friends name is typed, which brings up a selection list. This means the user is not required to stop typing, select a listbox – scroll up and down what could be a hundred choices – and then select a friend – but just select using the cursor keys from a much shortened list. Indeed this reminds me in some ways of the Canon Cat ‘leap’ keys:
You moved through your data using two extra keys called Leap keys located in front of the spacebar key and by typing strings of characters. The Cat jumped to the next occurrence of that string. Raskin claimed that the Cat’s Leap-key search method to scroll from the top to the bottom of the page took 2 seconds, a mouse took 4 seconds, and cursor took 8 seconds. Larger documents increased these search ratios.
The Leap keys also controlled text selection (indicated by hilighting), deletion, copying, and moving. If the selected text was a mathematical formula one keystroke with a special key calculated the mathematical result and the answer appeared on the screen with a dotted underline overlaying the original formula. If the selected text was a computer program written in either FORTH or 68000 assembly language, then a special key let you execute the program (I don’t think many Cat users did any Cat programming). You performed mail merges by selecting columnar text data and pressing another special key. Repetitive command sequences could be automated by assigning commands and text strings to the Cat’s numeric keys. One special key let you dial a selected telephone number either for voice or modem communications. Data received from the built-in modem flowed into your text as if you had typed it.
I’m not sure if there really is an answer here, or if we should just consider hybrid solutions, but for me typing using symbols to enable special faster interactions seems logical, the less I take my keys and attention from the keyboard, the faster my interactions seem to be; especially when I am cognitively focused on a single task.
Bugs Become Features... 
I would point to the rise of multitouch as evidence that innovation isn’t particularly being stifled. While still in the direct manipulation paradigm (and very much an advance toward less indirection), I remember seeing the early videos about it less than a decade ago and then BOOM, it’s everywhere, and rather well done.
There’s still a considerable text oriented ecosystem where new ideas can be incubated (emacs!!! :).) But consider the omniaddress bars in e.g., Chrome. By noting that having two “modeless” text boxes actually enforced a (disruptive) mode switch, they found a solution (textual disambiguation with incremental confirmation) that really changes how you interact with the “non-click” web.
Related of interest: http://blog.vivekhaldar.com/post/3339907908/the-cognitive-style-of-unix
Hi Bijan, you may indeed be right – but I’m not sure. The sheer amount of research work coming out of the likes of PARC with the Star OS, to me, represents such a huge leap forward that very little recent work comes close to matching it in both scope or magnitude. Incremental developments along the way do occur but I’d say even multitouch doesn’t represent the kind of leap Xerox made back then. And today, multitouch represents an input paradigm as part of a GUI not a replacement.
So there’s several different questions that need to be sorted out, e.g., the “global” rate of innovation (now vs. then), the amount of innovation reasonably *possible* (now vs. then), and the effects of given innovation (now and then; e.g., dominance).
Xerox Parc was pretty much working in an empty field with relatively little competition. So anything they did was likely hugely innovative and likely hugely influential. *Comparable* innovation might be occurring all the time without having comparable *influence* (and, of course, you have to measure “time out”; arguably, Parc innovations didn’t really take hold until Windows 3).
PARC was also unusual in making advances on many fronts at once. Again, this is partly a moment in history situation (new area; little division of labor, etc.) and perhaps partly the singularity of that team.
Consider non-monotonic logic which really started in an 1980 issue of AI (where the three “big” paradigms were published). There’s never been an issue of any journal wrt non-mon logic that is remotely similar to that issue (it’s really astonishing). But there’s been tremendous innovation and increase of knowledge since then.
Oh, and multitouch as not-replacement…well…yes, but then that’s a surprising metric for innovation, i.e., that the prior art is overwhelmed. I think this artificially distorts comparisons. (So, really, we need an innovation & impact metric to make the discussion sensible.)