Several things about Comdex have improved since last year. The traffic is not one of them. Any street within two miles of the Convention Center was as good as any traffic jam on the Southeast Expressway. Parking, normally near impossible, was complicated because Mercedes reserved half the main parking lot to set up a slalom course for Comdex attendees. The downside was that I had to park a half mile away. The upside was that I got to drive a CLK. Not a bad tradeoff.
One of the improvements I referred to above was that Comdex finally moved the conference sessions back from the Sands to the Las Vegas Convention Center, where the best exhibitors are. That makes it possible to shuttle between conference sessions and the show floor more than once a day. The advantage of this move is hard to put into words. A twenty-minute bus ride separates the Sands from the LVCC. You'd go to a session, have an hour of slack time, and end up wandering the Sands looking at row after row of exhibits from Korean motherboard and power supply manufacturers. This year, I was able to hop over to the show floor in a couple of minutes, visit a few booths, and gather information in between sessions. It worked out great for those of us trying to cram as much as possible into the show. But I do hope the Koreans don't get too lonely.
Through a combination of really awful Comdex commuter traffic and too much time spent on yesterday's report, I missed a couple of keynote speeches from the wireless conference today. Not a big deal; I'll catch them on tape and report on them later. Since I got to Vegas, it seems that every other conversation I overheard has been about wireless (The rest were mostly about whether or not to hit on 17.) But the buzz I’ve been hearing about wireless for the last couple of days became a roar in the e-mobility conference that started today, and the main lesson learned from most of the day spent in presentations is that every vendor has an agenda, and that they’re all different.
Mobile phone manufacturers are sold on the value of WAP, while everyone else appears to have written it off. Some companies are convinced that 3G, or third generation, networks providing 384 kilobaud wireless bandwidth (about the speed of a middling DSL connection) will be available within 15 to 18 months. Others are just as convinced that we’ll be working with 19.2 kbaud for a while longer than that, and they’re coming up with creative workarounds to do e-commerce using those low speeds. Disagreements rage about pricing models, the value of wireless streaming media, how big a screen is needed for browsing the Web, the proper form factors for wireless devices, and whether people want a single device that does everything or several more specialized gadgets. There were even disagreements about whether the push to wireless was being driven by business or consumers. But there are also significant areas of agreement among virtually all of the people I heard today.
In the next section, I’ll try to summarize the important points that were discussed in today’s sessions. I will have to leave some things out, or this will be way too long and I’ll be up way too late. Once I get back home, I’ll make my session notes available to anyone who wants all the details.
Here are the main points that were discussed, in decreasing order of consensus.
When it came to listing the most important missing links for wireless service, there were a few virtually unanimous points of agreement. The most important of these is location services. Everyone talking about wireless services explained how much better their services would be if they could figure out where the user actually is. Phone manufacturers cited implementation problems with E911, the emergency mobile phone locator system. Some felt that inaccuracy caused by cellular network limitations would delay completion past the government-mandated rollout in October, 2001. And even E911, as currently specified, isn’t quite good enough for all services. The government spec calls for a phone to be located to within 100 meters 67% of the time, and within 300 meters 95% of the time. This is accurate enough to allow emergency services to find a person, or to tell someone where a nearby ATM or restaurant is. It is not close enough to provide walking or driving directions based solely on the system’s idea of where the phone is.
Some companies are basing location-sensitive services on which cell tower is closest to the phone, which can place someone within a few blocks in a dense city. Commercial applications are being rolled out now using this information, on the theory that it’s good enough for some things and it’ll be way too long before anything better is available. This theme was sounded many times during the conference: work around network limitations to bring wireless applications to market as early as possible. Return on investment, even for primitive applications, is often under a year according to the editor of Pen Computing magazine. Don’t wait for improvements that may be a long time coming if the network that’s out there now is good enough to do anything at all.
A second theme that was widely sounded was communication with personal information sources. A service that books you a restaurant reservation over the Internet, for example, is more useful if it automatically enters the reservation into your PDA’s calendar. For an enterprise, tying a wireless device to data on the corporate intranet enables group scheduling of mobile workers, integration between email and applications, and many other synergies. According to Metcalfe’s Law, the value of a network increases as the square of the nodes. This is true of information sources, and the linking and synchronization of related items in different places should dramatically enhance the value of all the information. This was the core of Palm’s MyPalm.com announcement on Monday night.
Linking data is also the purpose of new products like Oracle Mobile Internet Platform. This product allows a corporate user to decide dynamically which personal intranet page items will be displayed on his or her mobile phone menu. In keeping with Oracle’s commitment to application servers, they offer the development kit for Mobile Internet Platform as a Web-based service. It can be reached at http://www.studio.mobile.oracle.com and used for all parts of the process from development and testing to production hosting – an interesting service model.
Another company that is mobilizing traditional websites is EveryPath. This company’s objective is to create a wireless interface to a website that is not tied to any specific device or display method. EveryPath starts with a company’s conventional website, develops a generic site design for small devices, adds the website’s business logic, and feeds it through an abstraction layer that makes the adjustments needed to render the page on the target device. If a new device is introduced, only a new abstraction layer need be added. The core site itself does not change.
There seems to be widespread agreement that security, or at least the perception of security, is one of the problems holding back e-commerce on wireless devices. This problem extends to wireless LANs as well as mobile phones. The message of the presenters was simple. Existing security is better than people think. Also, standardization will allow standard security protocols like SSL to be ported from the wired world, something that should happen fairly soon.
I was surprised and very happy to see the topic of access control addressed by more than one presenter. The issue is how to allow people to decide, intelligently or automatically, which contacts will get through to them. Nobody had answers, but everyone considers uncontrolled flow of messages to be a serious problem and is working on tools to address it. The tools range from automatic message filtering and blocking to displaying caller ID-style context information (the new Psion will include a display the size of a tie-tack that permits a user to glance at the identity of a caller without even taking the phone or messaging device out of his or her pocket.)
One presenter brought up presence advertising as a means of control. Presence information, which allows callers to easily determine whether a user is available to respond to messages, can be set to selectively advertise a recipient’s availability. For example, a person can be accessible to family at all hours but to work colleagues only during business hours, with the change in access occurring either on a schedule or manually.
There was a surprising amount of dissent about the smallest screen that could be useful for web browsing. Small cell phone screens are useful for simple query and response transactions, such as “You’ve been outbid. Do you want to submit a higher bid (yes or no)?”. They can also direct users to a small number of standard websites like Yahoo and news servers. Their value breaks down when a user needs to visit a less popular site that is not on the phone’s menu of services. The combination of typing a URL on a telephone keypad and reading a page that has not been formatted for a phone’s display is enough to stop most people from using the Web in this way. People offering services pointed out that much of the value of the Web was in finding a site with exactly the right information to fill a specific need. This becomes near-impossible on a phone, and many presenters described a backlash to WAP among users in countries where the service is widely available. Japan’s i-Mode service, they say, is successful because it stresses simple information delivery and does not claim to be a window to the Internet.
The view of Nokia’s representative, with a vested interest in sale of WAP-enabled phones, differed from his panel-mates from Handspring, HP and Psion. He stressed the idea that Internet-enabled phones will outnumber networked PCs by 2003, and felt that the phone might become the primary device for dealing with the Internet. This sparked a great deal of debate with the rest of the panel, whose vested interest is in selling large-screen devices that communicate with the Internet through simpler, and less expensive, phones. Both sides have valid points, but the reaction of Europeans toward phone-based surfing seems to have been summed up by an Englishman I met on the show floor on Monday. “WAP is crap”, he said. That’s all.
This topic generated lively discussion. Message adherents pointed to the runaway success of messaging services in Europe and especially in Japan, where kids learn to tap out words on a telephone keypad before they can write. Messaging, in the form of email, was the “killer app” for wired networks, and instant messaging on PCs is taking the world by storm. In surveys of users and potential users of wireless text, access to email and messaging is the #1 feature desired. Further, messaging is far less sensitive to low bandwidth than Web-style services. Pagers run perfectly well at 14.4 or even slower, and message delivery within a couple of minutes is acceptable in most applications short of conversational instant messaging. To work around bandwidth limitations, E-commerce models are emerging that are based on messaging rather than WAP or web transactions. Yahoo! Everywhere is an example of a site that uses this kind of model for transactions. The Palm Store at MyPalm.com will also use message-based transactions.
The efficient use of bandwidth is important because US wireless service providers have no incentive to shift voice bandwidth to data uses. Carriers can sell all the bandwidth they have to voice users, and if they add capacity, they can sell that for voice as well. While all carriers see the value of mobile messaging, they have a hard time thinking about degrading their voice bandwidth to make room for a new service when they are running at capacity just to serve voice needs. Nobody addressed the question of when this situation would change.
Hard to believe at first, but there was some disagreement about the value of some standards. Not standards like 802.11 and Bluetooth, which allow products from multiple vendors to interoperate, but standards like GSM that stifle revolutionary change in areas that still need it. A questioner pointed out that the emerging 3G standard for mobile phones is based on CDMA, a protocol that would not have been developed if the US had standardized on GSM at the same time as Europe and Asia. That said, the vast majority of panelists thought standards were a good thing in the vast majority of cases.
Many other, less strategic points were made by conference presenters. Here are some, in no particular order:
Don’t get too attached to that PDA or mobile phone. The majority of panelists felt that the lifecycle of these devices would fall into the same range as that of digital cameras, about six to nine months. Depressingly, that includes the integrated phone/PDA devices that are expected to cost $400 to $900. This doesn’t mean that the devices will be useless or even obsolete, only that new models with new features will be released in very rapid succession. This kind of a lifespan for devices and new features underscores the need to base corporate strategies on architecture rather than specific devices.
An entire session this afternoon was devoted to the 802.11b standard for wireless networking. Through creative badge-engineering, the geeky sounding name “802.11b” has been replaced with “WiFi”, short for Wireless Fidelity. Seen in this context, the puzzling renaming of the Firewire standard for high speed data connectivity to IEEE 1394 last year makes sense. Could people keep a straight face asking for a computer with Firewire and WiFi?
The standard itself is more important than the name. WiFi is finally a wireless networking standard with meaning. Any WiFi-certified device will interoperate with any other, regardless of who made them. In the same way that wired Ethernet cards can be purchased from any combination of vendors and used on the same network, WiFi cards will simply work on any compatible network.
The interoperability of wireless networking devices is causing wireless networking to be much more common. Service providers are beginning to build wireless points of presence at “hot spots” where mobile workers congregate, like airports and Starbucks. (Nobody actually mentioned Starbucks, but it makes sense to me.) Laptop manufacturers are starting to build WiFi into many of their current models. Apple builds a WiFi antenna into every Macintosh except for the cheapest iMac and provides a dedicated card slot for the wireless network card.
Some problems still exist for wireless networking. Roaming, in which a PC is carried from the range of one network access point into the range of another, still works only if both points are in the same IP subnet. Cisco is working on a router-based solution to this problem. And WiFi uses the same 2.4 gigahertz frequency band as some wireless phones, all Bluetooth devices, and microwave ovens. When asked about the problems of interference, especially in a laptop where WiFi and Bluetooth devices may be fractions of an inch apart, the panelists pointed out that both protocols were designed to reject interference. Both would continue to work even if they got in each other’s way, although they would probably slow down. Besides, they said, Bluetooth broadcasts at a fraction of the power of a WiFi card, so they’d be the likely losers in any conflict over transmission interference. This was the same answer I got last year from Ron Sperano at IBM, and it clearly indicates there’s more talking to be done.
A final issue preventing world domination by WiFi is cost. Wireless LAN cards are currently in the $200 range, about $50 more than conventional wired cards, and access point hardware that ties the wireless LAN to the wired network is not cheap. Cost reduction is a major goal of the WiFi industry group.
There’s more to say, but no time or room to say it. I’ll save the small amount of booth news I have for a later dispatch. Tomorrow brings lots more conference sessions and probably a little time on the show floor. Bye for now.