Comdex 2002: It sounds like Bill Gates, but is it actually the fat lady singing?
Greetings from Las-Vegas-on-the-Ipswich, as I monitor Comdex Fall from home for the second year. My hat's off to the people who turned down my invitation to fund my trip to Comdex - my audience was ahead of me on this one.
It became clear a couple of weeks ago, when I got the show's preview magazine, that the proportions of this Comdex would be less than epic. The exhibitor list told the story. It listed less than a thousand exhibitors, heavy on the software startups and Pacific-rim peripheral manufacturers who populated the overflow exhibit space at the Sands Convention Center in past years. Gone were most of the big names of the industry, the companies that funded booths bigger than your house and generated much of the excitement on the show floor. Even many mid-tier companies were missing. Of the major PC manufacturers, only Hewlett-Packard and Toshiba have a presence this year. Palm has a booth, but Handspring declined to use the Comdex platform even though it is working hard to establish its Treo line of text-enabled cell phones. From early reports, it appears this Comdex will be hard pressed to match the attendance numbers of the 2001 show, which was held under draconian security measures just two months after the September 11th attacks and in the face of difficult economic and travel restrictions.
Monday morning on the radio, I heard technology columnist Larry Magid say that the troubles at Comdex were indicative of a trend away from the mega-expo and toward niche-specific shows. I think he's wrong - the attendance figures for CES, held last January in Las Vegas, and CeBIT, held in Hanover, Germany in March, were down by a much smaller margin than Comdex. But more than that, I hope he's wrong. The industry needs big trade shows, which create a generalist's melting pot where experts from different parts of the business have a chance to meet and interact. How many alliances and products have come out of meetings that would not have happened outside of the pressure cooker of Comdex?
As a paying customer of Comdex, though, what I'll miss most is the chance to talk to industry execs, product planners, engineers, and everyone in-between. Other than the major keynote speakers like Bill Gates and Scott McNealy, every presenter was available one-on-one conversation after a conference session, and over the years I've run into everyone including Bill Gates - twice - cruising the show floor. As for the product engineers, some of whom spend fifty-one weeks a year in the basement labs of their companies and the fifty-second in Las Vegas, Comdex offers a rare opportunity for an ordinary geek user and buyer to meet those folks and get an honest assessment of the problems inherent in a new technology. Comdex was expensive, crowded, unmanageably huge, and an invaluable window into the next year's developments in all areas of technology. If it fades away, we'll have lost something important.
Unfortunately, Comdex itself probably will fade away. Its growth was based in Metcalfe's Law, which describes the exponentially increasing value of communities as they grow. (For a good example of this, compare eBay to its competitors.) But as Metcalfe gives, it also takes away. Comdex 2001 had 1600 exhibitors and about 125,000 attendees, down about 40 percent from the previous year. Fewer vendors talked to fewer potential buyers, and the exponential effect of the shrinkage was just as powerful as its effect on the show's growth for two decades. Those who stayed away in 2001, and those who attended, looked at what they might expect from this year's show and just decided they didn't need to come back.
Comdex's decision to include consumer-oriented exhibitors also took a toll on the show's core appeal, blurring the line between itself and the Consumer Electronics Show (CES). The Comdex name itself is short for Computer Dealer's Expo, and the show grew on the strength of its orientation toward corporate buyers and serious tech product developers. Nevertheless, in 2000, Comdex brought in consumer companies, and closed their largest, closest and cheapest parking lot for the duration of the show so Mercedes-Benz could set up a test-drive track for their cars. Did more E320 sedans show up in data centers? Unlikely. Did the loss of focus, along with the added inconvenience of finding parking - and paying $20 a day for it - make a few extra people decide to shop their trade-show dollars elsewhere? More likely. The likely beneficiary of Comdex's fall will be CES, a show that was always about selling gadgets to consumers.
I've seen this happen before. In the early 1980s, the National Computer Conference went from the largest computer show in the world to oblivion in two years. The catalyst then was an upstart trade show called Comdex. Now, as Comdex follows the NCC's arc into history, there are several catalysts: the change in the general economy and in the tech world, the reduced number of new technology introductions relevant to corporate buyers, and most importantly the refusal of Comdex itself to change with the times. When the time came to decide where to spend this year's reduced marketing and travel budgets, people found they had no compelling reason to subject their bodies and their company budgets to the Comdex ordeal.
It wouldn't have taken a rocket scientist to see this coming, not after last September's attacks and especially not after last November's show. Some adjustments in the product - better online content for a fee, for example, which might have extended the powerful Comdex brand into a year-round, Internet-based portal for new technology. A launch of such a service during the actual trade show, when Comdex awareness is at its highest, could have given Key3Media a platform for growth. But the stewards of Comdex decided to play 2002 as if it was 1999. As a result, with news stories pointing to their likely bankruptcy filing, they probably won't get another chance.
My first surprise in this year's online Comdex visit came when I tried to watch the keynote speeches. Where last year's show posted complete, unedited video streams of the 90 minute speeches, the site this year is limited to a single excerpt between three and eight minutes long. And so far, none of the three excerpts I've watched - Bill Gates, Carly Fiorina and Scott McNealy - have had any kind of product-specific content. The editors have singled out whatever optimistic statements these three have made about the industry and presented them as the only content worth publicizing. I will try to digest the news stories from these speeches and present the truly interesting things in a future mailing.
Even in these straitened times, Comdex is still the venue for many new product announcements. Dell, which has no corporate presence at Comdex this year, used the Microsoft booth to roll out its new Axim X5 PocketPC. The base system arrives with 32M of memory, slots for SD and CompactFlash cards, along with the usual complement of PocketPC software. They offer one distinct advantage over competitors: a removable lithium ion battery that can be charged outside the unit, supporting laptop-style hot swaps and extending the effective battery life considerably.
The Axim is virtually indistinguishable in its specs from the industry-leading color Compaq iPaq, or any other mid- to high-end competitors, so Dell competes in this more-or-less commodity market on price and services. An Axim broken during the one year warranty period gets the typical Dell treatment. Dell will ship a replacement unit the same day they determine that the owner's initial Axim is broken. The owner returns the broken PDA in the same packaging. Downtime is kept to a minimum, an important advantage for a device as critical as a PDA. None of Dell's competitors offer similar coverage.
The entry-level Axim sells for $199 after a $50 Dell rebate, just over half the cost of similar models from other vendors. Another $100 buys the advanced Axim, which adds a docking/charging cradle (the cheaper model, presumably, offers a cable instead), a faster processor, more flash ROM, and 64M of memory. $200 is often cited as the breakthrough price at which consumer devices find wide acceptance, and Dell's Axim is the first PocketPC to compete with the simpler Palm PDAs at this price point - although they will probably sell more copies of the more realistic, and still reasonably priced, advanced model.
Since the physical layout and software of all PocketPCs are basically the same, manufacturers must compete with small feature advantages, price and service. Dell, with their direct sales model and excellent service reputation, is in an excellent position to lead their competitors in both areas. Their $200 price also allows a PocketPC to compete on the same level as Palm-based PDAs for the first time. They may have been late to the party, but Dell has proven they know how to show the market a good time. Check out the Axim at http://www.dell.com/us/en/gen/topics/segtopic_axim.htm .
---
On the almost-too-weird-for-words front, watch company Fossil has developed what looks from the pictures to be a workable wristwatch-sized PDA. It's runs the full Palm OS, complete with stylus and Graffiti handwriting recognition. Some changes had to be made to the user interface because of the small screen, but a watch may be the perfect form factor for the Palm OS, which stresses simplicity over features. With a 33MHz Dragonball processor, 2M of memory, 160x160 screen, an IR port for beaming and a USB port for computer connections, this appears to be a fully functional Palm PDA. No details yet on price or ship date, although Fossil already sells a proprietary PDA watch that downloads information from a PocketPC for $145. Check out Fossil's offerings at http://www.fossil.com/tech/default.asp?ID=tech&load=true .
---
Nextel's Tuesday announcement was significant for two reasons. First, they are offering a version of RIM's BlackBerry cell phone for their IDEN network. BlackBerry users have access to IDEN's Direct Connect feature, allowing the units to connect as walkie-talkies to other Nextel customers. Nextel also offers email monitoring and all the other text services BlackBerry users expect.
But even more interesting than the device was Nextel's announcement of a $50 monthly plan for unlimited wireless text. This plan and T-Mobile's $40 unlimited text plan for the Danger Sidekick may be the beginning of a trend to address potential buyers' uncertainty about the cost of wireless text services. This is a trend that should catch on. Read about Nextel's offering at http://zdnet.com.com/2100-1103-966245.html .
Every now and then, a flawed idea comes along that's so powerful, so compelling, that it has to be tried even though everyone knows in advance it's going to be a loser. Sometimes the idea is so powerful that it erases the collective memory every few years and demands to be tried again.
This is the power of handwriting recognition, a can't-miss idea implemented time and again on can't-do hardware. The list of companies that tried and failed to make a commercial success from tablet computers is impressive. Microsoft. Apple. Momenta. General Magic. Tusk. Go. Toshiba. IBM. You've heard of the ones with deep pockets and broad product lines, but the specialists just sank beneath the waves.
Well, kids, it's happening again. Fueled by Bill Gates' belief that tablets are the natural way to use computers, and having apparently laid off everyone who remembers PenWindows, Microsoft has launched a new offensive against the keyboard.
Why won't the TabletPC succeed? Because the keyboard is a better input technology than the pen by almost every measure - so much better, in fact, that the pen's few advantages can't rescue it. Consider reliability: Press the "j" key, and you get a "j", not an "l" or an "i". My handwriting doesn't match yours, but the TabletPCs designers assumed they'd be able to find the common ground. Microsoft has even made the decision not to allow users to train the Tablet's handwriting recognition, because they think that allowing for personal variations in handwriting will make the recognition less reliable, not more. Think about that.
Second: Microsoft uses one bad technology to rescue another. They try to cover the shortcomings of character recognition through auto-correction - by distinguishing whole words and automatically substituting what they believe the writer meant. This is bound to be unreliable, since it's built on top of unreliable character recognition. Consider what would happen if Word's spell checker took every typo you entered on a keyboard and automatically substituted, without warning, a real word based on the characters you typed. How much longer would it take you to proofread a document - in fact, how much less would you trust your documents - if you had to make sure that every individually correct word on the screen was the one you actually meant to type? Doonesbury lampooned the Newton for its handwriting recognition - remember "egg freckles"? - and any system that combines unreliable handwriting recognition with automatic auto-correction will cause its users the same kind of trouble.
The combined effects of recognition and correction problems lead to the biggest problem with current handwriting recognition programs - usability. The biggest advantage of handwriting is that you can write without looking at the paper, allowing you to maintain eye contact with your subject and generally monitor what's going on around you. Typing on a traditional keyboard in a group setting is an isolating experience, and the ability to free your eyes from the screen is one of the reasons pen input seems so compelling. But, if anything, current handwriting recognition technology forces you to spend more time looking at the screen, because you can't trust what the recognition engine will do to your words when you're not looking. Of course, you can turn off handwriting recognition and just save the digital ink, but once that's done you can get just about the same utility by buying a 59 cent steno pad and a hundred dollar scanner.
The steno pad, by the way, won't break your arm when you're trying to carry it around, or write standing up in the hall. And it won't take a few seconds to wake up, or a couple of minutes to reboot, before you can start writing. And it won't break when you drop it, sit on it, or close your overloaded briefcase too hard. And it won't cost you a few extra minutes at airport security.
But if you want an expert opinion on handwriting recognition, don't take mine. Consider Jeff Hawkins. Mr. Hawkins invented Graffiti, undoubtedly the most successful mass-market handwriting recognition tool on the planet, when he worked for Palm. He used Graffiti, along with the rest of PalmOS, with him when he left Palm and formed his new company, Handspring. But look at Handspring's new Treo line. The Graffiti pad is gone, replaced by a thumb-operated microkeyboard, except as an option on the low-end Treo phone. It's still PalmOS, but now it's keyboard-operated.
This doesn't imply that Graffiti doesn't work. The omission of autocorrect actually makes it more trustworthy than TabletPC, and nobody's going to write an annual report on it anyway. But Handspring's move away from the handwriting technology that its founder pioneered says something about the practicality of keyboards.
And in case you think Hawkins is alone, look at the other recent PDA introductions, like Sharp's Linux-based Zaurus, Danger's Sidekick, and the increasing adoption of BlackBerry-based phones by companies like Nextel, Cingular and T-Mobile. Even on such small devices, handwriting recognition is starting to look like a transitional technology that sufficed until someone could figure out how to make a decent small keyboard.
Unfortunately, the pen computing ideal is easy to imagine, while the reality that you're no better off takes some time and experience to learn. This is why the pen concept keeps coming back, and why it will continue to surface until the technology catches up with the dream. But it won't happen this round, I predict.
By the way, let me give credit to two companies that refused to rise to the bait. You won't be seeing tablets from Dell or IBM at this Comdex. Of course, IBM has a reason. The company's most recent attempt at a pen computer, the TransNote notebook, is still fresh in their corporate memory.
That's all for days one and two of Comdex. In some ways there's more news about the show than from it, but there have been a few interesting items, and there are still three days to go. I hope to check out developments in biometrics, Linux, and maybe a few more cool gadgets. Stay tuned.