In the music business where I spend part of my time, there is a truism about studio work:
Fast, good and cheap: choose two.
There are many situations in which optimizing for two goals compromises a third, and small devices is one of them. The three opposing goals in this case are portability, usability and function. Optimizing usability and function gets you a nice laptop or even a desktop. Optimize portability and usability, and you get a Palm. Optimizing portability and function is a dead end that gets you a product that does good things but is so hard to use that no one buys it.
Itís this third area, the intersection of portability and function, that makes designing a good small object so difficult. The continued popularity of 10-plus pound laptops (I always count the full package, not just the machine) demonstrates that people are willing to compromise true portability in favor of desktop-style capability. The success of the Palm shows that people will pare down their expectations of what they can do with a device if they can carry it around with them all the time.
But I canít think of a single portable product that succeeded in the mass market - or even with early adopters, for long - if it was hard to use. Case in point, if a little far afield: just about every cell phone sold in the last couple of years is capable of sending and receiving full text messages, but hardly anybody uses this useful feature. Why? Itís too hard to enter text on the telephone keypad. But people have been happily sending messages on BlackBerries, Motorola Timeports and other devices for years. Devices like the Handspring Treo and its competitors from Kyocera and others will, in my opinion, unlock the demand for mobile text and set the stage for integration of mobile and IM that should be a boost for the telecomm industry.
Iíve been repeating this mantra for fifteen years: there are only six sizes that matter in portable devices. They are:
An obsolete category that mostly covers old luggables. Other than Dolch I donít know anyone who still manufactures computers like this, although Iíve got a nice old Datavue and an IBM PPC in my collection that show how it used to be.
This covers most modern laptops. These devices, particularly when combined with their power supplies, outboard or removable drives, and other peripherals and cables, take up so much room that they really need their own carrying case. They are smaller than category A machines, so their carrying cases look more like briefcases than sewing machines. The adoption of this design was the gating factor that moved portable computing into the mainstream. Having flown with an Osborne 1 in the early days, I can vouch for why this is true.
Every once in a while thereís a flurry of these machines, like the Thinkpad 500 and the original Compaq Aero, and more recently the Thinkpad 240 series and others. They always fail. They enable people to carry a computer to more places than a standard laptop can go, but are hamstrung by short battery life, small keyboards, small screens, outboard peripherals and other design issues. You could carry one of these computers easily, but cabling together the power supply, CD-ROM and who knows what else was an operation most people quickly got tired of. You could also use the machine by itself without peripherals, but the battery would invariably die in the middle of the day, or the middle of the flight. This category is a good example of how people reject otherwise perfectly functional, portable machines because theyíre too hard to use. I also think tablet PCs, which are being designed to fit this size category, will probably fail for some of the same reasons (and, until handwriting recognition becomes perfect, for some different ones too.)
This is where the magic starts to happen for me. Itís hard to measure how much more useful a computer becomes when itís always with you. The obvious examples are calendar and address book applications - who wants a calendar that canít leave your desk? - but other applications benefit too. Word processing in your pocket? Who hasnít wanted to jot down a few notes while on the run? With a word processor you can do that easily, or write a whole memo while waiting in the doctorís office, or take meeting notes in real time for immediate storage and distribution, or read and edit a document you received online. You also eliminate the little scrap of paper that get lost or washed just after you wrote the worldís best business plan on it. Similar cases could be made for spreadsheets, databases (address books are just a start), project planners, contact managers, offline email readers, or any number of other applications.
This category, of course, stares straight into the jaws of the usability trilemma. A pocket is a cruelly small space for a computer designer to work within. Many have come close, and with good designs, but have failed because an extra half inch of length or width or a few millimeters of thickness made them category C machines, banished to a briefcase and eventually a desk drawer. These include the Poqet PC, the Newton and most of the original Microsoft CE designs (the ones that were small enough for a pocket, like the Cassiopeia, were simply terrible, unusable designs.) Machines that just made the cut, like the Psion 3 and 5 series and the HP 100 and 200s, found a niche among small device addicts with a high tolerance for funny stares (Is that a 5mx in your pocket, or are you happy to see me?) But the niche hasnít been big enough to sustain any of these products, unfortunately, or to give them a foothold in the larger market. Most people treat these devices like category C and eventually ditch them in favor of a laptop. Or notepaper.
Women have a slightly different threshold here, since you generally donít wear clothes with pockets but do carry handbags. There is much less standardization in womenís handbag sizes than in menís shirt and pants pockets, and women forego handbags entirely in many situations while men are rarely without pockets. In any case, a successful device will have to appeal to both men and women. So Iíd use the manís pants pocket as a minimum requirement for portability here, with the shirt pocket as a goal if it can be done without sacrificing too much usability and function. Iíd guess that the original Palm would have failed if it had been too big for a shirt pocket.
Human factors are a big deal in category D. What makes the Psion work as a long-term device is the well-engineered keyboard, which enables effective six-finger typing at 30-40 words per minute. The unfortunate compromise was giving away good screen contrast to lengthen the battery life, but the relative failure of the Revo (rechargeable batteries, excellent screen, but a smaller keyboard and screen in a smaller package) shows that Psion was closer to the mark with the larger series 5.
This is meant to cover devices like the BlackBerry. Itís not meant to include those belt pouches that hold a Palm, HPC or other category D devices. Belt pouches are just a sneaky way to add an extra pocket to a pair of pants, and until Ethan Hawke or Tom Cruise is seen sporting one in public theyíll be seen as far more nerd accessory than fashion accessory.
The size of these devices seems to make them unusable as full-function computers, but maybe thereís some creative strategy that could make them work. BlackBerry-style keyboards are good for message-length text, but I wouldnít want to write a two page memo on one. Heads-up displays and folding keyboards seem like a kludge to me, and risk market rejection because theyíre too hard to use. But a component-based approach, using Bluetooth, could be an answer.
This category is not suitable for standalone devices, but could be used to move some functions out of the main device (more later). These could be key fobs, wristwatches or other items that can be worn without calling attention to themselves and hurting mass-market acceptance.
People are not likely to use a device like this as their only computer, particularly with the sub-sized keyboard and screen. To enhance usability, movement between the portable device and a regulation PC should be as close to effortless as possible. I see three aspects here:
Transferring files in either direction should be painless. I like the Psion approach, which adds the Psionís disk and directory structure right into Windows Explorer, and allows automatic format conversion when files are dragged and dropped on the other machine. Better would be an approach like Microsoftís Windows Briefcase and HPC synchronizer, which would allow for automatic synchronization of active files. Even better would be the ability to edit files directly on the other machine, as if it were a network server. The goal is to make file movement easy while avoiding the need to reconcile multiple versions of the same document on different devices.
Bluetooth is the way to go here. An unobtrusive serial or USB cable is a good second choice, but with Bluetooth thereís no cable to lose, and thatís a big advantage. Having the device communicate through standard network protocols is also a big win, since it avoids having to install software on other peopleís computers to get connectivity while traveling.
Itís becoming harder all the time to conceive of a useful computer that canít connect to the Internet, and a functional, pocketable and usable Web browser and email device might be successful all by itself. I still havenít missed connectivity on my Psion (good thing, since the networkís not here yet to support it), but I do have the Motorola Timeport P935 to take care of minimal communication needs. I donít carry a cell phone.
Network connectivity is where an outboard device makes the most sense. It gets a fairly bulky and power-hungry component out of the main box. It makes upgrading easy as networks evolve, so the obsolescence horizon of the main device isnít tied to the obsolescence of the telecomm network. It also allows people to use their device, and all their files, in places with incompatible communication standards. Tri-mode is great, but go to Europe and your phoneís a brick. Move the transceiver outboard and you donít have to sacrifice your whole kit to maintain connectivity.
If possible, a transceiver that fit into Category F, something the size and form factor of a watch or an automobile remote control, would be ideal. But a Category E belt-mounted transceiver, which might double as a cell phone for use with a handset (see the Motorola Accompli 009), would work too.
It goes without saying that all file formats have to either be the same on each machine, or that excellent in-line translators need to be written.
For me, the dream is a system that will fit on my person (category D or E) with enough capability to allow me to leave my laptop at home for short trips. This means:
∑ Voice recorder: Iím used to this from the Psion. Itís great for remembering things in the car, where typing is impractical. Much of this document was lined out in voice notes on my drive back from vacation yesterday. Decent sound, like the current CE spec, would make the system an audio device too. I donít normally carry a portable audio player but would probably use that feature if it was built into my main device.
∑ Video out, which would make the device a standalone presentation machine if a PowerPoint-like application was available. If the box also had keyboard out, and support for an external pointing device, I might be able to live on it for short consulting stints. Maybe the screen of the device itself could double as the pointing device.
∑ 802.11b, to take advantage of high speed networking in the increasing number of places itís becoming available. The right place for this feature is in the outboard communication device.
∑ A non-slip case finish. I canít count the number of times the 5mx has slipped out of my pocket. I can say that now since itís out of warranty. Psion got the case finish right with the 3c, but for some reason they didnít carry through on later models. Other basic system savers: good screen protection (I broke my screen once by walking into the corner of a handrail, right through the top of the case), minimal weatherproofing.