Thursday, 28 November 2013

The HP TC4200: Your Grandfather's Tablet

It is often said that mankind fails to learn from history. I believe that this is not true. The real problem is that we learn too well. Rather than facing each new problem with fresh eyes and an open mind, we dig through the past and dig up the same old solutions; we meet new problems with the same old violence, the same injustice. We work the same old game, and grow fat as the world struggles to bear our weight.

I have often wondered how The Future will remember us, how history books yet unwritten will describe the present day. I surmise that the internet will ensure that writers of The Future will have no shortage of facts, and so there will be none of the blank spaces that litter the histories of ancient Egypt, and the unrecorded histories of the Polynesians, and of cultures that left no written material behind. The Future will have no shortage of facts. It will have a superfluity of facts, a clashing mass of big data, and perhaps this will be just as much a barrier to understanding as the alternative. How will historians assemble these facts into a coherent narrative, one that reflects the world as it was, rather than an equally-plausible narrative that never happened? The broad outline of history will probably remain intact, but with such a wealth of data historians of The Future will be tempted, probably compelled to fine-tune their portrait of the past; to create a microscopically detailed recording of events that history would once have smoothed out, like pebbles on a riverbed. The result will be like staring at white noise, it will have detail but no meaning.

The distant past of 1993. Will styluses take off? No.

It is redundant to say that nothing is ever forgotten. The things we truly forget become unknown unknowns; we do not know that we have forgotten them. But in an interconnected world there is usually a record of a thing's passing, if not the thing itself, and the great churn of events occasionally brings long-lost artefacts to the surface. I am reminded of all the lost motion pictures from the early days of cinema that eventually turned up in a dusty attic or abandoned monastery; fragments of novels, fragments of symphonies discovered in abandoned notebooks, lost paintings conjured from beneath the surface of new paint with X-rays.

Today we're going to have a look at tablets. But not modern iPad-style tablets, except in passing; I'm going to talk about the PC-based tablets that predated them. It's hard to remember nowadays, but there was a time when tablets were a dead end. Nobody wanted them, they were a solution in search of a problem. At first they were a well-meaning but misguided attempt by technological visionaries to capture a future that never happened, and then they were just a cynical attempt to jazz up a declining laptop market.

One such relic of the distant past - a Hewlett-Packard HPTC4200 from 2005 - fell into my hands recently, and I was curious to see how it compared with my Asus Transformer. Hence this post.

A HP TC4200, powered by a 2.13ghz Pentium M, running Windows XP Tablet Edition. It's a convertible laptop with a rotating screen.
Performance-wise it's a mirror-image of modern tablets. The graphics subsystem is rubbish, and so it's useless for games, but even an elderly Pentium M can hold its own against modern tablet CPUs, so for browsing the internet or number crunching it's just as fast as a tablet.
But it's hobbled by the 1024x768 resolution. The graphics card can output 2048x1536 on an external monitor - ironically the same resolution as a retina iPad - but this defeats the point of a portable tablet.

In the 1990s and 2000s PC tablets were often called pen computers, slates, or sometimes convertibles. Journalists didn't want people to confuse them with digitising tablets, which were something else entirely. PC tablets emerged from the pages of science fiction into the real world in the early 1990s, and are still with us - the Microsoft Surface Pro is essentially their modern heir - but there's a melancholy air of failure about them.

And in time the terminal replaced libraries.

For decades tablets were the next big thing. Punch cards and printouts were replaced in the 1960s with keyboards and monitors, and then mice in the 1970s; and text-driven interfaces had given way to graphical desktops, which were not necessarily ubiquitous even by the early 1990s. PC gamers, for example, were still accustomed to using DOS for the serious stuff such as Mechwarrior II until well into the Windows 95 age. The next logical step - or so it seemed - was the pen-driven portable computer, with some kind of interconnected online operating system that would recognise handwriting (and probably speech), so that businessmen could schedule meetings simply by jotting mtng John next Tue re: Fisher account 15:30 and then John would get a message to say that he was due to have a meeting and perhaps one day meetings could be held digitally and where would it end, eh? Eventually the computers would arrange all the meetings and then they would learn to talk to each other and human beings would become HELPLESS SPECTATORS to their fate, as in The Forbin Project.

I can just remember the period. The 1990s, that is, not the future of The Forbin Project. Tablets were the next big thing in the same way that quantum computing is the next big thing today, e.g. they were exciting but no-one expected them to actually exist for some time. How would the computers talk to each other? How was this supposed to work on an aeroplane? Who was going to pay for all this new hardware, and would the handwriting work? What use was a pen-based interface for editing spreadsheets? Were people expected to enter Lotus 123 formulas with a pen? Isn't that a disaster waiting to happen? And so forth.

And then, in the early 1990s tablets became the hottest thing ever, with developer conferences and endless articles in the popular press, but people got sick of them almost overnight and they became something of a joke. IBM's pioneering ThinkPad 700T had such a minimal impact on the real world that, by 2005, the writers of PC Magazine had completely forgotten that it existed:

In reality the X41T was the first ThinkPad tablet for twelve years, but not the first ThinkPad tablet ever; the very first ThinkPad (1992) was a tablet, and it was followed by the 750P convertible in 1993 and the 360P convertible PC in 1994 (as well as several direct successors). Notice the use of "convertible" and "slate PC" and the $1,900 price. Contrary to the review, the TC4200 doesn't have a built-in DVD/RW drive.

Contemporary hardware from GRiD Computing and Toshiba also died a death, to the extent that GRiD is forgotten nowadays. Toshiba still makes tablets but nobody cares about them. For almost a decade tablets languished in obscurity until Microsoft decided to revive the concept in the early 2000s, but tablets were no longer The Future by then. CNET's contemporary report on the launch of XP Tablet Edition has a subdued tone and points out that the tablet PCs of the period were conservative designs, expensive, and with limited support. Microsoft was simultaneously hedging its bets with Pocket PC, a kind of less ambitious successor to the Windows CE-based palmtops that had failed to set the world alight in the late 1990s, and it's a testament to Microsoft's hugeness that it could afford to fail and fail again. The industry's final attempts to push the PC tablet, in 2005 and 2006 (with the "ultra-mobile PC"), were met with a resounding shrug from one and all. When I think of the failed 2005 tablet boom-that-wasn't I automatically think of this image, which puts me off my food.

And yet the idea of reviving interest in laptop sales with a new form factor was fundamentally sound. A couple of years after XP Tablet Edition flopped for the last time Asus invented the Netbook, which sold in great quantities and kept the PC market (and indeed XP) afloat for a couple of years. Netbook sales thrived while tablet sales dived because netbooks jibed with the PC market's drive for a small, cheap computer that could connect to the internet; tablets were bulky and expens-IVE, and that didn't JIVE. The public simply didn't GUYVE a damn about them because they didn't help people to LIE-V... oh, it's no use.

Andrew Bujalski's Computer Chess disappointed me, but I'll review it in the next post. Here we see an early prototype of the iPad.

The Purpose and the Price

For years there was a fundamental question as to whether Future Tablet would be a new device or a rehash of existing technology. It would of course have to be portable, which was a challenge in itself in the early 1990s. Laptop PCs had only recently evolved into their modern form. And there was the OS. The ThinkPad 700T and GRiDPad had the option to use GO's PenPoint OS, or DOS plus a pen-driven version of Windows 3.1 ("Windows for Pen Computing") but it seemed silly to graft a pen interface onto such an unpromising foundation as Windows 3.1.

The hardware requirements for DOS and Windows support were challenging in a mobile context, even for conventional laptops, and the compatibility benefits were very slight for tablet use. My recollection of Windows 3.1 is of using it simply as a launcher for Word and Excel, and the Windows version of Fractint. It managed memory and printers and so forth with slightly less fuss than DOS, but there was nothing particularly pen-ny about it. You didn't even need a mouse to use it, you could use keyboard shortcuts for almost everything.

PenWindows was originally devised as Windows/H, and was intended for Windows 3.0. It was eventually launched as Windows for Pen Computing. It was a commercial non-entity, but had the effect of deflecting attention from PenPoint; GO went bust, Microsoft did not, and in that respect Windows for Pen Computing was a roaring success.
Presumably Microsoft assumed that if PenPoint had been a success in pen tablets it would not have been a stretch to port it to conventional laptops and eventually desktops as well, at which point it would be a threat to DOS and Windows.

Short of an expensive Total Pen-Driven Office and Communication Suite(tm), in the early 1990s the pen interface came across as a gimmick in the PC context. For graphics applications it made more sense to buy a separate digitising tablet, and a larger monitor, and of course neither of these things were portable. It seemed therefore that pen computing was an opportunity for a fresh new start using new hardware, a new architecture, a whole new interface paradigm. Apple decided to take this route with their Newton MessagePad, which drew nothing from the existing Mac OS or Macintosh hardware. It ran a wholly new pen-based operating system on chips from ARM, at that time a minor player in the computing world.

But although the MessagePad attracted an enormous amount of press attention, the marketing blitz didn't translate into huge sales. The machine helped to found the PDA market, but didn't even dominate that, because it was simply too expensive; prices started at $800 for the launch models but that didn't include the optional fax modem, without which the MessagePad was a very expensive filofax. In 1993 $800 was cheap for a laptop, but the MessagePad was far from being a laptop replacement, and for the genuinely practical uses anticipated in the LA Times' contemporary report it was overkill.

Apple had the right idea in starting from scratch with a new OS, but made the same mistake that a lot of innovators make when entering new territory; they were too ambitious. The company seemed to recognise this - the MessagePad's original prototype was larger and envisaged to cost thousands of dollars - but even so the end result was simply too elaborate for the intended market, who for the most part opted for a newer generation of small laptops. The Newton became a novelty for the wealthy, and novelties date badly.

The next big thing in PDAs was the Psion Series 3, a clever little clamshell device that was prized by journalists for its tiny but functional keyboard, and the next big thing after the Series 3 was the Palm Pilot, which sold for about $300 in 1996 and was marketed as a simple electronic organiser that you could hook up to your desktop PC. It's notable that the original press release goes out of its way to avoid any mention of handwriting recognition (Graffiti, the Pilot's text entry system, is buried quite far down the page). In retrospect the industry's emphasis on handwriting recognition seems a terrible mistake, and although the magazines and columnists were convinced that the idea had merit, I can't remember a single person in real life who wanted to interface with their computer with writing, even if the handwriting recognition worked 100% perfectly. Which it didn't. There's the separate issue of using a pen at all, rather than fingers, but having used both formats I'm in two minds as to the relative merits of each; the average Joe clearly enjoys touch-sensitive screens, but stylus input is very precise on smaller screens and works surprisingly well as a mouse replacement on a laptop.

I can understand why the industry became fixated on handwriting recognition. Human beings have been storing and communicating information by scratching marks on a surface for thousands of years. But typewriters had largely displaced handwriting as a means of business communication by the 20th century, and it's as if the inventors of pen computers had forgotten this. Handwriting recognition never became the dominant interface form in any context and is nowadays moribund. Modern tablets - including PC tablets - use finger-and-thumb-driven touchscreen interfaces and onscreen keyboards that you press with your fingers, and the applications and OS are built to cater for this.

The problem is that lengthy writing is awkward with a pen, and smaller memos are almost as fast to write by pressing virtual keys. Reviews of the MessagePad tend to predict that handwriting recognition would be fixed in time for the next MessagePad, so that you would never have to amend your input. But even if the problem had been solved, the need to search the internet for non-English words and names defies natural language recognition. A glance at the Google search history I generated whilst writing this article throws up looney toad quack / loonytoad quack / tsang tsz-kwan* / gilead share price / pilot 1000 / windows/h penwindows / jerrold name / cortex a5 dhrystone, and what would a perfect handwriting recognition system make of that jumble of nonsense? Besides, I find it hard to recognise some people's handwriting and I'm a human being and can understand the context of the words. I suspect that truly effective handwriting and speech recognition would require a computer capable of passing the Turing Test, because written and verbal communication are so peculiar to the human experience.

* She reads Braille with her lips; I was thinking about unusual input methods.

The dedicated PDA eventually died off. Mobile phones assumed the same functionality, with the added and crucial bonus of not having to connect them to a separate wireless unit. The last survivor of the PDA age was the Blackberry, which had emerged from the world of text pagers. Unlike the MessagePad, the Blackberry 950 email PDA did one simple thing well, in this case send and receive email wirelessly. It had a little keyboard. At roughly the same time SMS texts were taking off, and it seemed that overnight everybody under the age of 25 was capable of rattling off semi-coherent text messages just by waggling their thumbs like so (waggles thumbs). As the saying goes, "the street finds its own use for things", and texting was massive in the UK in the early 2000s. The market wanted to push handwriting recognition; the street wanted texting, and the street won.

The fact that people took to such an odd, unintuitive interface as multi-tap SMS whilst spurning handwriting recognition perhaps says something about how misguided handwriting recognition had been. It was an example of technologists using the lessons of the last war to fight the next one; the world had moved on, and as I write these words handwriting itself is a dying form. My own handwriting skills have atrophied to an extent that I now dread having to fill out forms in case I make a mistake. For Christmas cards I simply draw little triangles and squares and hope that my family understand. If God had intended for man to communicate with his fingers, he would have put mouths onto our fingertips so that we could talk with them. Or he would have invented the typewriter, which he did. I like to think that up in heaven László Bíró rests in peace, knowing that his invention was the ultimate evolution of the writing implement and will never be surpassed or replaced, because it was the end. There are few objects more symbolic of the 20th Century than the plastic ballpoint pen. A device that used modern mass-production methods and modern materials to facilitate the recording and communication of information at a price affordable to all; it was the purpose and the price that mattered, not the method.

In fact it almost seems as if voice communication is a dying medium, not necessarily a bad thing if you ever have to share space with a noisy commuter on the train or, horrible thought, in a packed airliner. Voice recognition was another obvious but misguided technological side-street that exercised the minds of technological evangelists for years; despite the success of Siri and its imitators, it suffers from the same problems as handwriting recognition, but magnified because speech is even less formal than a dashed-off note. I find it hard to understand the young people of today when they talk about twerking and looney-toad quack, and I'm a person myself. Not young, but I can remember being young. I remember that I was young.


Ultimately the modern tablet market is the result of two different evolutionary paths that have met almost in the middle. The PC tablet was an attempt to graft a touchscreen interface to the relatively powerful hardware of an IBM PC, in a form factor that was easily portable and ran for hours on batteries, preferably without needing a cooling fan. The iPad-style tablet was an attempt to improve the speed and power of a smartphone platform in a form factor that was large enough to browse the internet comfortably on. The two platforms faced strong technical challenges and in practice the smartphone route won, although it was a tight race.

With the Windows CE palmtops and Pocket PC and latterly the UMPC Microsoft had the right idea, and it's ironic that Microsoft is nowadays written out of tablet history. The problem is that although Microsoft had the right idea, the company's efforts were diffuse and peripheral to its core business, whereas Apple had laser-like focus, and the company has learned painfully that failures hurt and can kill. And, it has to be said, there is a probability that Microsoft's efforts in the tablet market were really designed to kill GO and latterly Palm and Handspring, and that the tablets were simply a means to achieve that end.

Scroll down and you get Steve Jobs when he was overweight and had a proper beard. The analyst's implication that Apple needed a sub-$1,000 computer wasn't as odd as it sounds today - the original iMac launched at $999 and for a time the iMac was Apple's salvation.
This was when Jobs was interim CEO. "He's not the man for the long term", says Chip Colby. "Part of what you need from a CEO and a chairman is someone with a better grasp of running a business". Which is uncanny; Jobs only lasted for another twelve years at Apple, during which the company's share price CRASHED from $7-8 a share in January 1998 to a rubbish $545 as I write these words.

In the 1990s Apple spread its efforts far too thinly, and was in danger of becoming an irrelevance; Steve Jobs brought focus back to the company. Whereas Microsoft had the money to launch a string of products without caring much whether they flopped or not, Jobs recognised that Apple did not have the resources to fail as hard or as often and was only a flop or two from going under. In his first period at Apple he had direct experience of the Apple III and the Lisa, which wasted a lot of the money that the Apple II was bringing in; the slow uptake of the original Macintosh led directly to Jobs being sidelined and then forced out of the company. By the time of the iPad the company had some breathing room, but I doubt that Jobs or anyone at Apple was prepared to launch it half-heartedly.

Meanwhile the PC tablet remains in a state of chaos. In the early days the fate of PC tablets was wedded to the PC laptop, and it's worth bearing in mind that truly practical, portable PC laptops that could run for a whole working day (plus commute) are a relatively recent phenomenon; and laptops that you could hold and use in one hand are very recent indeed. Although a few low-power x86 chips were tried out in the early 2000s - Transmeta's Crusoe attracted the most attention, powering the clever HP TC1000 of 2002 - none of them were particularly satisfactory until the Core Duo age, and it was not until relatively recently, with the falling cost of components and the rise of affordable SSDs, that the hardware problem was cracked.

PC tablets still exist, and indeed Lenovo's ThinkPad tablets are a direct descendent of the earliest pen PCs. But Microsoft has caused no end of confusion by using the desktop version of Windows for some of its tablets - the expensive ones - and an entirely new and incompatible operating system for the rest of its tablets - the cheap, supposedly mass-market ones - and a third and again incompatible OS for its phones. Rather than develop for three separate operating systems, app developers have stuck with iOS instead.

But what about the Hewlett-Packard TC4200? I've actually run out of space on the introduction so I'll deal with it in the next post. Yes, this is the internet and I can make this article a million words long if I want, but you'd get bored. Here, have a cup of coffee, with brandy in it. Sit down, have a scone. The TC4200 was part of the 2005 tablet "boom" and attracted moderately favourable reviews. Hardware-wise it was essentially a HP NC4200 laptop with a rotating screen. At 2kg it was really a non-starter as a tablet, and I surmise that HP had a very limited budget and perhaps there was a contractual obligation to Microsoft to launch a tablet, any tablet, in order to qualify for cheap Windows XP licences (say). But more about its hardware in the next post.