Return to: Contents
© techrevu/ernest lilley 2003
Editor: Ernest Lilley
Shipping/Mailing:
500 Talbot Hall Rd.
Norfolk, VA 23505
Phone: (757) 093 5589
Email: contact@techrevu.com

TechRevu reviews technology related devices formatted for desktop, PDA, or Email distribution.

Our Other Sites:
SFRevu (Science Fiction Reviews, News, and Interviews)

CeBIT Landscape: How Big Is Big?

And discovers the technological wonders of

Research Heaven: Hall 16.

CeBIT Landscape: How Big Is Big? (originally published in Byte.com April 5, 2001)

European show etiquette is quite different from American, I'd even venture to say more civilized.

It's almost flattering to be invited in for a sit and a drink and a talk about a company's upcoming products. Even tiny American booths got into the action, with products more familiar to certain passersby. I knew it was time to go home when I nearly cried at the sight of Costco chips and salsa.

As genteel as the show may appear, CeBIT is 100 percent cutthroat. Archrivals eye each other across a thin, cigarette-butt strewn stretch of unpadded gray-speckled carpet, and attract the wandering jaded eyes of prospective customers with pulchritude, amplification, light, and sound. Since CeBIT combines the best of what would be about eight different tradeshows in North America, this is for all the marbles. (In the Hollywood dramatization, there'd be sandbags in the aisles and open gunfire when customers weren't looking.)

The show is so large, occasionally grand, that passageways through the middle of booths are unremarkable, and necessary for traffic flow. Some even have second-floor overpasses between their halves, a Habitrail for potential bigwig clients.

How Grand? Try the Fujitsu Siemens Computers (FSC) booth, where I was feted to a lovely salmon lunch while gently educated about its upcoming laptops and servers. The FSC spread was as genteel as anyone could want, but there's a general frenetic atmosphere. In The Wall Street Journal (reprints were sitting on the conference table), FSC announced its continuing efforts to be Europe's largest computer manufacturer. (Compaq, though, was not to be easily dislodged.)

And "booth" is too bloodless a term. There are five elevators in the center of the FSC stand in Hall One, they're part of the building's structure, but have been worked into the booth itself. These lifts lead from the main show floor to (1) the second floor of FSC's permanent booth; (2) downstairs to the storage areas; (3) up to the fourth-floor FSC reception area.

The fourth floor itself is a bustling little village; it contains the general hospitality area, the press area (with salmon and wienerschnitzel lunches), FSC's conference rooms, and the like. It's all permanent, and used maybe two weeks out of the year. Despite its brevity, CeBIT is for all the marbles.

Not that everything is grand. Some things are merely large. That fourth-floor landing also leads to what's been termed the ninth circle of hell, the roof of Building One. It has dozens of ocean-shipping-container-sized buildings atop it, past which a cold wind blew and a freezing rain fell. These smaller buildings contain PR firms, small meeting rooms, more storage, and sundry. There's also a doublewide for the FSC cafeteria--with hundreds of staff on the stand, even infrastructure has infrastructure.

And, CeBIT is big: Building One is the biggest structure, but there are 31 permanent buildings on the Messe grounds, supplemented by dozens of one- and two-story tents (up to 4,000 square feet each) for CeBIT. Roughly, products are segregated by function, with some haphazard rubrics like "Information Technology" inhabiting the gorgeously vaulted Building 11, plus 12, 13, and three or four others. Want to see raucous retail paeans by German cell-service providers? Outbuildings 28 through 31 are your elbow-jostling destination. Not interested in the vibrant world of banking? Skip Building 18, though expect to walk through it at least once to get somewhere you do want to visit.

There is even a separate building just for research and technology transfer, Building 16. Mind, the strange calculus of the Messe makes this building seem small by comparison, though it's quite large enough on its own to be a respectable conference anywhere else on the planet.

This organization method breaks down with the necessity to put every interested party in some booth. Building 10 is nominally "Network Computing," but with the collapse of interest in that subject, it was a catch-all for latecomers of all sorts.

This organization scheme also breaks down with the overly hyped, long overdue forces of convergence.

Lucent booth / RWE power line modem. / The University of Rostock sponsors a student robot completion using components from Lego Mindstorm and others.
Swatch wants us all to get with "the beat", decimal time standard.
How many computers can you cram into the tiny Smart Car?

Research Heaven: Hall 16 - (originally published March 28, 2001, Byte.com)

One of the organizational units is research and technology transfer, Building 16, showing work from various universities and research organizations throughout Germany and the rest of Europe. Since it is right across from the Press Center, I made it my first stop. Pure research isn't the sort of thing you'll see at Comdex, another reason I willingly endure the hassle of getting over here. Hall 16 wasn't well trafficked, even by journalists sniffing for good stories, all the more for me.

In the GMD-Forschungszentrum booth, the W3C was promoting its standards and would-be standards. It was offering tutorials on SMIL, WAI, P3P (an online privacy standard), and SVG (Scalable Vector Graphics). It was encouraging to see this group promoting what it does, ere it slips into irrelevance.

The research on display was a mixed lot, some impenetrable to the English-only speaker, much of it from a German research consortia that had dragged its star players out for a good hawk. Most of the exhibitors were principal investigators (PIs) and their graduate students, all as apparently uncomfortable in business suits as U.S. computer-science professors would be. But like any other researcher, once they figured out I was truly listening to what they were saying, they warmed and started into the arcana of their subject -- some things are more universal than others.

The Deutsches Forschungszentrum fur Kunstliche Intelligenz GmbH (DFKI, the German center for research on artificial intelligence) was showing several cool projects. Cool as best as an English-only journalist could tell from the German literature and English explanations. By far the most interesting was FairPay, an attempt to algorithmically prove correctness in e-commerce systems, in particular crypto and message authentication. Program-correctness proofing fell out of favor in nearly all academic circles because it is exponentially more difficult to prove large programs. The FairPay approach is top-down, which I took to mean that it proved function-level correctness, rather than by individual lines of code. This also greatly reduces the amount of effort needed to proof programs. FairPay also appears to be usable to prove correctness in existing systems, not merely ones under development, which previous attempts at program correctness did not.

If this is as significant as it appeared, then program proofing might make a major comeback. Most e-commerce systems are rickety and ill-planned, and could certainly use more rigorous development methods. I encouraged the FairPay PI, Oliver Keller, to publish in ACM and other like-minded publications.

Another interesting DFKI project on display was CASA, a demonstration system for integrating online professional auctions with support for logistics and other pricing criteria. The particular demonstration was forestry auctions, unsurprisingly in conjunction with the German Ministry for Forestry, coupled with auction tools for bidders.

Cartographic Programs On Display - The first project that caught my eye was CDSAT: Change Detection by Analysis of Satellite Images, cartographic/photo interpretation software that algorithmically detects changes between pictures in multiple spectra. Mortimer Canty, the PI, was Canadian, so language wasn't a barrier. His software uses wavelet technology (which has also fallen out of favor) to normalize images and emphasize differences.

Contributing editor Eric Pobirs saw a demonstration of (deep breath) An Open Global Infrastructure for Spatially Aware Applications out of the University of Stuttgart, a Geographical Information System (GIS) that combines virtual objects with representations of real ones, so as you "fly" through a city, you see real 3-D images of terrain and buildings, coupled with computer-based knowledge of how they relate.

From the list of partners (Institute for Photogrammetry, Institute of Communication Networks, and Computer Engineering, Institute of Parallel and Distributed High-Performance Systems), the project is very multidisciplinary. From the brochure, it's the very model of the sort of flying-through-a-city stuff you see in every spy movie.

The Rest of the Research Bunch - Much of the "research" in Hall 16 was actually commercial companies showing off ready, nearly ready, or pipe-dream technologies. For instance, RWE, Germany's largest power producer, was promoting its Internet-over-power-line service. RWE serves the rich Rhine Valley, with over 3 million rate payers. Just as U.S. phone companies have had to create arm's-length subsidiaries to sell Internet service, RWE (as a regulated utility) did too -- as "RWE Powerline." It was partnering with Ascom, which has grabbed a major share of the power-line networking-equipment market, and an ISP to haul bits to the individual neighborhoods.

Most of the world outside North America pays per-minute charges for every call, which has proved a major impediment to the spread of the Net. If the significant technical impediments to powerline networking are in fact fixed, RWE may find itself in a position to pick up a significant chunk of users, and then later offer value-added services like video-on-demand (a fact not lost in its marketing literature).

RWE's distribution scheme is like this: The Internet attachment points will be at individual neighborhood points (from context, I'm assuming on stepdown transformers), 2 megabits shared between about 150 homes. It expects a not-impossible 10 percent uptake. Fifteen users sharing 2 megabits is quite reasonable, again assuming it can deliver technically.

It was hardly the only organization trying to carve up the potential Internet subscriber pool in untraditional ways. Lucent and Alcatel eyed each other across an aisle, promoting various DSL technologies. The recent divestiture was reflected in the booth layout, with Avaya (the former subsidiary of Lucent) occupying a large corner of Lucent's layout. Alcatel was also promoting a licensed, line-of-sight wireless service. I couldn't quite get why it was superior or even competitive to AT&T's unlicensed non-line-of-sight service, but it was crowing about design wins in China and Europe.

Where Next? - I could have spent an entire day learning about cool future research projects, but the siren song of ready and near-ready products tempted me away from academia. In future reports, I'll talk about more mundane things like PCs, wireless devices, and handhelds, but it was great to learn about future projects a little off our usual focus.