Tuesday, March 20, 2007

Interfacing

I have lately been regarding my laptop with venomous hate since that bloated make-work project for new Microsoft hires, Word, declared that "There are too many edits in the document. This operation will be incomplete. Save your work."

The dictatorial order to cease productivity must have provoked some sort of internet stroke, because I couldn't find any advice on how to fix this beyond the usual titter tatter about defragmenting hard disks. Great advice for time travelers from 1980, not so helpful here. Anyone who can turn on a computer nowadays runs scheduled maintenance, and it's mildly insulting MS recommends something like this when confronting a fatal error in one of their flagship products. Being frustrated with support, however, has become one of those computer rituals we do, like reinstalling Windows monthly and destroying AOL disks with surplus Soviet firearms. This weekend I plan to burn several copies of the Word install disk just to have the pleasure of blasting them to bits, and blasting those bits into even smaller bits.

Let us now sober and look at this thing, this laptop, this housing for obsolescence. It is amazing how little of its volume, mass, and cost is devoted to its stated purpose of computing. In its way, it is the Polish cavalry(1) of the 21st century. The whole thing is a container for human interfaces. The QWERTY keyboard is itself a legend in designed inefficiency, for if the keys were more optimized, old-fashioned typewriters would have jammed constantly under the high speed assault of typists' fingers. Meaningless today, of course, but that's information technology for you. People don't easily part from something that takes years to learn, which is a rather sensible attitude, all things considered. The toolmaker must always operate by the maxim: to make is to save.

So we must craft something so much better that no one has to learn anything. In its day the keyboard was such an invention, as you didn't need to apprentice to a typesetter to use one. The letters were on these little buttons, and when you hit a button with a letter on it, the letter appeared. Sure, eventually you had to learn to change ribbons and such, but it's still heaps better than putting together an inked press block, mostly in that you didn't have to change ribbons as often as you had to ink the block.

At the least one should not spend more time training than what was saved by switching to a new tool. Let's make that into a ratio so we can get a scale on these things:
[(Time spent on old tool)-(time spent on new tool)] / Time spent training
The typewriter is a great example of this. Training time was way less than the time it saved, so it makes money from the get-go, i.e., it had a percentage greater than 100%. Once your percentages drop below 100, you have to be very careful, because you are betting that the new tool will keep up its current rate of savings into the future, just to break even. Accountants have very sophisticated ways of calculating these things, but it is something technology people frequently forget, which is a shame, because it's how we figure out the need for a tool in the first place. To make is to save.

So whither computer? Many folks see the demise of the classic laptop in favor of something like a mobile phone on steroids- something with the computing power of a modern laptop in the form factor of a Treo or Blackberry. That's a no-brainer. The thing that takes some thinking is the interface, the thing that takes up so much space in conventional computers. If we could get over looking like a dork, HUD displays would make a great monitor substitute, but otherwise miniaturized projection systems can make any white wall into a screen. Keyboards are harder, but possible. I am reminded of a sci-fi novel in which the user, equipped with a HUD and a "virtual office", can type on any flat surface because his "virtual office" software displays a projected keyboard wherever the flat surface is. Such a system severely lacks tactile feedback, however, and it has the side effect of looking like a dorky virgin because you are wearing a HUD. I'd argue for some sort of flexible keyboard, or perhaps some way of scanning finger movements. Speech recognition is another great possibility, as most people type slower than they speak.

Eventually, however, we won't need an interface at all. When they release this system for people, we need to be careful that we don't lock people into a QWERTY system for the mind- something that locks thoughts just for the sake of an interface.

1.That's Polish Cavalry as the German and Soviet propagandists portrayed them- the actual Polish cavalry in 1939 disposed itself quite well against what was probably the most formidable invasion force in history. Poland fell so rapidly not because of its "Pollack" commanders but because they based their defensive plans around the rapid response of allies in the case of German invasion. Since the West preferred that Hitler kick Stalin around a bit, the rapid response was deferred. In proportion to its population, Poland suffered the wor

No comments: