The next segment we will be talking about technology drivers. What are the enablers the technology sector is providing for changing the way people can globalize, can reach forward, and so forth. I will speak for a while and then turn over the stage to Armando Garcia who will present some very specific technologies in the media area and then we will finish the segment and at that point there will be a break.

Fundamentally technology changes the relative costs and factors of production. By giving easier access to more information, by providing crucial data, providing it on time, by making it possible to have far more complicated calculations, using far more complicated software to apply to problems using the ability for computing to do much deeper analysis at very low cost. There are possibilities of brand new changes. Expected results are reaching out to customers - of course there is the risk that the customers can reach out to new suppliers and learn about more deals. Herr Pichler described the good part about reaching out to customers, the reverse side is that they also have choice and you need to be ahead of them. It's easier to reach out to new suppliers - of course they too can reverse the story. And so this is a continuing battle when the technology makes all the experimentation, all of the search far easier, and also it is possible to work more easily with new partners.

So let me just try to introduce just a little bit of reality. As a technologist I love talking about Internet years, web months, everything happening immediately and it really is true. You can tinker with software over the weekend; you can put up a new version of something, just so long as it doesn't matter very much. If the software is just something you are playing with, if you are letting people try, it doesn't matter - any good programmer can bang out a new version and put it out tonight. If instead you have mission critical software, something that your enterprise really depends on then we are talking about something that takes months, maybe a year. It is perfectly normal to have a three or six month testing cycle for any important change. A change to any important software no matter how easy the change is, you may have to go through a whole test cycle to make sure that nothing else breaks.

If, even more, you are transferring whole business processes, going to direct sales for example, changing whole practices, you know this is going to take more than a year to introduce. To really change how things work in an organization - or even more complicated, between organizations - it is necessary to be very realistic about this. Similarly if you are changing the entire computing system, if you are replacing what you were doing before with something new, not just making a moderate bug fix or a small extension, but really changing it out in reality it takes several years. Your betting your company on any such change nothing will happen fast because of the price of a bad bet. If you are instead going even further changing customs, changing the mechanisms of a whole economy, getting into the issues Dr Banneman raised, this takes multiple years to play out, perhaps even decades. And certainly any change that is going to change the way society works with fundamental political habits is a generational change. It takes decades to play out. And it's important to realize which changes you are talking about - which things are a quick hack over the weekend and what requires the passing of a generation in society.

In this segment we are going to be discussing six directions in which technology has a significant impact: the introduction of inexpensive and reliable communications, probably the biggest single technology driver that technology provides for global activity; better software; more reliable services; the ability to speak with many other people in the way they need to be spoken to; the opportunity to use far more complicated and wonderful media, exciting video and so forth; the ability to have much faster relationships between companies, more dynamic inter-relations, a change in speed; and the ability to manage huge amounts of information to have many complicated business processes changing and underway.

First let me talk a little about communications. Basically there have been enormous improvements in the available technology over the last five years. Finally the revolution is beginning to happen on the communications front. The costs for providing quality data, quality video, and so forth have fallen precipitously - huge factors in the actual costs. Prices however have moved rather more slowly, and they are also moving at different speeds in different countries. It is important to recognize what the cost factors are. Eventually competition will set in and the prices will be following them. I'm not going to predict when various price factors will change, but the costs are a universal datum to consider.

Let me separate communications for the moment into long haul and local. The biggest changes that most people can see are in the backbone connections, the amount of cost per bit per second. There are pure technology issues - there are changes in optical fiber technology, the ability to send a signal thousands of kilometers without an amplifier, without a repeater. This changes your choices on how you do the installation; it has huge implications for trans-oceanic cables where every amplifier installation is extremely delicate. The lasers and the electronics at the ends of the fibers - there have been major breakthroughs in efficiency, power efficiency, total bits per second - so you can put a lot more bits per second through every channel on a fiber. The advantage of this is you don't actually have to pull out the fiber, you just change what is on the ends. Since the timescale for installing the fiber itself is usually multiple years, because you have to dig trenches or put things under the ocean and big planning, changing the ends means somebody changes a couple of things in the office. Big changes here, orders of magnitude literally. The next revolution called DWDM, Dense Wave Dimension Multiplexing - what this means is multiple colors, multiple channels on the same fiber, so not only is each channel of the fiber getting more bits per second, but you are getting to put more channels onto each fiber. The net result here is that we are talking in a few years of having commercial possibilities of a terra bit per second on a single fiber. A terra bit is a million Meg bits a second - a lot! This is already possible in the laboratory it is now a matter of evolution, it is a product upgrade equipment. This takes no crystal ball on whether it will happen - only when and how the prices will fall. Just to give a normalization of what a terra bit means, that is equivalent to a million VCR grade TV channels or a hundred thousand top quality video channels on a single fiber. We have a lot of capacity coming. Voice channels, ordinary telephony I won't even discuss - one of these carries all the voice of the US on a single fiber if all were going from one point to another. So the capacity for the voice just disappears into the noise as it were.

That's the good news on the backbone, on the connections between the big cities or between the countries. Locally, traditionally people talk about the problems of the last mile and how do you get that signal from the central office into the house. In reality this should be viewed as the opportunity of the first kilometer. How do you get all this wonderful stuff from the house to the network and back? There have been finally breakthroughs in the last few years. For more than a decade we have been talking and hearing about the possible opportunities of better technology and finally they are happening. Most of unfortunately still connect using a ordinary analogue modem at 28,000 bits per second - theoretically some of them will go at 56,000 bits a second but you never seem to get that in practice. There is a theoretical maximum on a phone channel of 64,000 bits a second and that's all you can get running on an ordinary voice channel. However, many people use ISDN, which is and 128, kilobits a second is available. There is the visual subscriber loop technologies which move up an order of magnitude or more - suddenly using the same copper plant but not using voice channels you can talk about transmitting mega bits one way, hundreds of kilo bits the other way on an ordinary voice channel between you and the rest of the network. Suddenly you are talking about using the cheap copper line into your house to carry video grade communication.

If you don't like that one you can use cable modems - special devices that attach to coaxial cables that you are using already for some of your TV reception. Once again you are able to talk about mega bits coming into the house and easily transmitting hundreds of thousands of bits back. There are good engineering reasons in both cases why there is an a-symmetry - in the short run this is not going to be a big problem only if people insist on sending a lot of high quality video from home will the a-symmetry start being an engineering challenge. Now these technologies have existed in the literature in the laboratory for a long time and my colleagues who invented ADSL have been promising that is was going to everywhere next year since 1985! It's finally happening. In the US the expectation is that in the middle of 99 there will be at least one million homes that have access through either ADSL or cable modems. We have finally got from the point of it's a toy and there's little deployment experiments to it will be available in a large fraction of the country. It's still a difficult engineering problem - every installation looses money because of problems but this is just a matter of learning curve. I think that we are going to see over the next few years in the US certainly major successes here. There are European exercises already on use of cable modems; homes are being equipped appropriately. We are finally beginning to see the possibility that even in the home not just in the business you can talk about mega bits a second.

Let me talk now about software and service - new possibilities, new capabilities, things you can do reliably using workable systems in technology. Fundamentally, better software is coming available. New technologies, better packages, more experience embedded in them. You don't have to be the pioneer doing everything first yourself, you can simply buy a shrink wrapped product or some service installed by experts that will provide many of the capabilities needed for web business, new digital world, e commerce and so forth.

Security problems - we all know they are complicated, we all know they are expensive. The realities are many of those have been addressed appropriately. You can once again get practical installations that can provide adequate security - smart cards, biometrics, and so forth - to give you the identification, the cryptography to give appropriate control. It's not trivial. This is an area where no amateur should proceed because it is so easy to get security wrong but there are people who know what they are doing and you can view it as a problem that is just underway. There are new capabilities that you can get support for - I will come to this in a moment - but the problems about how to put a catalogue up on the web, two years ago you were a hero if you could manage to do that right. Today you can call IBM or any number of companies and say I would like to put up a web page, and if you just want to put up a web page with a dozen items on it, you can have it in twenty minutes. If you want to put up something a little more sophisticated it'll take a couple of days. If you want to do something really fancy, that's hard. Similarly, accepting payment is relatively straightforward assuming you just want to do electronic funds transfers, checks, credit cards, standard software does all this, does it reasonably. Probably the big challenge coming is integration of the new with the old. As I alluded earlier on some of the timescales, if you are trying to take an on going set of business processes, on going systems and make the new world, the new electronic tickets, the new electronic offers fit in smoothly and seamlessly. This is difficult and this is going to be one of the big challenges for all the successful companies. Life is much easier if you are just starting up fresh and you don't have an existing customer base, you don't have an existing set of processes - of course, you don't have an existing business or existing money, so there is an advantage of being established and know what you are doing. However, as you add these new capabilities integrating with the old world is increasingly difficult but is probably the success factor for many organizations. And then as you start saying that the electronic means of communication with the customer, with your suppliers, with your partners goes from being an interesting opportunity - what I normally call 'selling tee shirts on the web' - to being the fundamental of the new business. Part of your new business model, suddenly you will no longer accept 'the system crashed again' as that's okay. You don't say the telephone system crashed again, that's okay. Once you start saying this is business critical you have to install true business critical software, hardware, monitoring everything else. Suddenly your requirements become absolute not just maybe. In some sense, these become the two biggest practical challenges to moving out to the next stage of a really integrated new e business.

I am just going to dance through a number of topics - I don't have time to go into appropriate detail - almost every line here is worthy of a one-hour lecture.

Payment - they are the issues of accepting electronic cash, credit cards, debit cards, Giro accounts - all of the standard techniques. There is a whole new world of being a bill aggregator offering bills for other people doing it electronically - this is probably more important in the US than in Europe because Americans still don't believe in electronic funds transfers but I just note that at the moment, AT&T basically pays me $4 a month not to receive a phone bill from them by getting it electronically, they literally cut my bill by $4 a month. This is just the first step for real companies setting out in the direction.

Payments in the business to business world are a radically different subject than business to consumer. The big companies don't usually pay with credit cards but they do pay with electronic funds transfers and purchase orders. It is a very different business process and this becomes a business process integration problem not simply a little gateway. The progress is large here. In the next few years you can expect these things to be nailed.

Catalogue sales I mentioned. If you just want to put a text catalogue, you can have that in ten minutes, if you want to have something with pictures, its not much harder with a scanner, if you want to do a great job as Dr Garcia will talk about later, if you want to have 3D images, if you want to have video, if you want to have much more complicated linkages now you are pushing the technology, but it is coming. It will really affect the way customers read what you are offering and the extent you can customize the offer can figure for the individual will also clearly affect what they think about dealing with you.

Transactions - those of you who are familiar with the computer science of a transaction - this is a nice well defined world, you do something, it's done in a minute, rather like going to a cash machine it either gives you the money or it doesn't and its over either way. In this new world where I say maybe I would like to buy a ticket to Bonn, please reserve it for me and I'll come back in a day and tell you whether I am buying it or not. Suddenly I have a transaction that is up in the air for a while. Even worse if I say I'd like to buy a ticket to Bonn but actually I am only going to do that if I can get the hotel at a good price and if it's available at the weekend rate, so the multiparty interaction is up for several days, it's a whole new world of the interaction. Managing that, no losing anything, is technologically a problem but the progress is pretty clear embedding this in software so you don't have to think about it is going to be the next step in a few years so that in a few years these complicated transactions, these very much more flexible interactions will start being a given just as the quick database dip - yes, no, my balance is okay - is standard today.

Other functionality - on line negotiations, auctions, requests for prices, brokerage. You all know the amazing success of a few companies like Onsale and Ebay and so forth which have already started to radically structure the world of amateur collectibles. Well to some extent they are going to start effecting core businesses also. We certainly predict that auctions and competitive negotiation are going to become increasingly important part of e business. Basically, fixed prices are a convenience when the communication costs are high, when it's really annoying to have to go back and say can you give me a better price or can we negotiate on this deal. If it costs me no dollars and no cents to send the packet saying no, give me a better offer, no I don't want to pay that price going to some place and can we get a better fare. All this changes the power; it changes the negotiating capacity on both sides. The power of the individual rises, also the power of companies dealing with other companies' changes. It is no longer reasonable to say let me call the purchasing agent to buy $200 worth of pencils. You wouldn't do that, you'd just say I'll take the ordinary price, but if a computer is just going to talk to another computer why not have it advertised and find out who can get me $200 for $180 on pencils. This is going to be a big change - it will affect the procurement worldwide, it will affect marketing and sales worldwide. You can have anybody contribute just so long as they can do the logistics. It will not be profitable to send pencils from around the world to save 10 cents to get them delivered tomorrow. On the other hand, getting information sent around the world could very well be profitable. So there are big issues here on optimization, what the real costs are what it would cost to fly something versus sending it via glass fiber? The world is changing; the software is available to support some of these changes. One of the big areas of hard work over the next few years will be optimization of the supply chain, how you manage that better

The questions of virtual enterprises, dynamic businesses starting to deal with companies in a very close way on a short instead of a long timescale - this is also a technology question as well as a business one. The technology is now becoming available to allow short-term videoconferences. I can decide to connect up to you by putting a video camera on my computer, a video camera on yours using the internet in between. I don't have to spend a month getting the phone company to put in a special line at each end; I don't have to rent time on a satellite. If I am quite willing to take medium quality, I can just do it today. Improvements in the networks as I mentioned earlier mean I will be able to have better quality in a small number of years. So two companies that don't normally deal with each other suddenly can. New families of marketplace is not simply I will go to your web site and decide if I am going to buy something, but a more complicated portal, a more complicated marketplace where people deal with each other all the time, these are also part of the new technology. New standard capabilities over the next few years will allow you to set up a far more complicated relationship - multiple parties, long term connections for professional niches for individual businesses and the technology is coming for all this at a regular industrial level not just at research level.

Customization - this has been touched on already a couple of times. This is going to be a very big theme, most people believe in the moving business. Right now you get personalized or customized offers only in special high value markets, some place that wants to charge you a lot of money, is prepared to customize and personalize for you, mass markets you watch it on TV. The cost factors changed in this one by factors of 100 or 1000 because of computer capabilities, new mathematical out rhythms, ability to use new data. There are wonderful new opportunities - the best marketing minds will figure out how to really take advantage of this information. The information itself exists, the computing will exist. Not just targeting the customer, repackaging for the customers' needs specific groups of information or offers that appear to be aimed at you. In the long run, even the manufacture of goods as needed. In the information world there are cases we are already seeing in insurance companies that are willing to create a specialized policy for a customer request. Why do auto insurance policies only come with fixed amounts of deductible - 500DM, 1000DM? Why not 792DM? The insurance company can handle that - they can now manufacture a policy, as you need it. This involves having information, being able to compute, being able to generate. There's a whole new set of opportunities and the technology makes this possibility wide open. Now the question is who will be smart enough to find the right way to make an offer to a customer, to make a negotiation with a supplier.

Quality - I already alluded to this as a bit driver. Users only care about the service they get. They're not actually interested in whether the network is working, whether your servers are working, whether your particular software is working. All they care about is that they tried to do something and either it seemed to work or it didn't. This is a huge challenge to the technologist. The fact that the customer doesn't really care how it's working gives us a lot of leeway of fall back, restart alternatives - the only trick is it's got to happen so fast the customer doesn't know something went wrong. This is a hard problem, any of you actually involved in building systems know how hard that is to do but the good news is that there is serious progress. We know how to monitor networks, we know how to monitor re-routing, we know how to do planned failovers of one node onto another in a computing farm. There are actual possibilities of doing this right - what's essential is that the customer doesn't know that anything went wrong if it went wrong. They simply have to have the comfortable feeling that you try to do something with your business and it works. The technology here is very tough, there's lots of research issues here, there's lots of practical business process to make things bullet proof. We're getting there - the question is what are you willing to pay to make what level of quality in what level of timeframe. The customers will start demanding 99.99% availability pretty soon and if you only offer 99% they get real annoyed.

Basically on the street technology side, we probably doing real good fundamentally. The network is improving, the number of bits per second are a liability coming through, we have new software capabilities offering new business possibilities, new basic support. Another more software area, but equally interesting, is the ability to speak with others and many others that you didn't normally talk to.

There are a bunch of issues that come up when you think about global interactions. There are multi-cultural interactions, the most obvious being currency - the internet today is still remarkably dollar orientated and given how easy it is to present prices in any currency that you ask for, this is clearly a change that we will see a lot of. Harder of course is multi-language presentation. Not everyone wants to read everything in American English - 75-80% of the web is in American. This will also change statically; the browsers are quite capable to saying if there is a version in French I will preferentially see this page in French. You can see a demo on line translation capabilities; you can get translation that works acceptably enough for many purposes. On line automatic translation isn't good enough for legal grade documents - you would be crazy to sign a contract translated by a computer program today without knowing more. It won't have the beautiful poetry of a wonderful marketing campaign; you still need an ad agency to make it sing. But if you want something that is good enough to be useful to figure out where this plane is going, what this book is about. Translation software exists - no people no hands. The main successes are improved out rhythm experience but to be honest most of the real progress has to do with bigger memories, faster computers. Work that you could only do on a super computer a few years ago, you can do on a $500 portable computer today. This is a theme. Heavy computing makes up for many sins.

Similarly speech output, speech input. We all know the basic forms of this. If you call a Freephone number for a company and find yourself what is called in America, voice mail jail - you know, press 1 if you never want to talk to me again, 2 if you do, do you want me to talk to you in Dutch, Flemish or in English (and it says that in English of course!) and then you go on for another five minutes and then it says I can't help you! This is all standard; this is not good enough. If you start introducing talking computers, ones that can present more information, and yet they can talk to you in a language that you like, sounds decent - suddenly this is a new world of interaction. You are talking with far more people - not just people who are far more tolerant and only a small fraction of the people you want to sell to or buy from are that nice. Technology is going to move rapidly towards making if much easier much better not just possible. There is also with speech recognition, knowing who is talking - this for example is very useful in video conferencing being able to put up a little footnote saying this is Stu Feldman talking and then a picture of me and then as it shifts across an eight continent connection you do this. The technology is starting to be available in limited ways; it is also useful in security, also for 'Thank you Mr. Feldman we are able to offer you this.' This is technology that is working accurately today.

Speech recognition. The important fact is to realize that there has been huge progress and practice on speech recognition, speech understanding. It's nowhere near perfect; it never will be perfect because human beings aren't perfect at this job either. If somebody talks to you bunch of words in isolation you will make a surprising number of errors - a standard experiment. However, computers are getting quite good at recognizing reasonably large but still limited vocabularies, being able to handle people that they deal with a lot and if you can craft the task, computer understanding is actually adequate for many, many purposes. You are already hearing this on standard telephone calls - say or press 1 is a standard thing I get all the time when I make a free phone call. This area, once again, huge progress in understanding how the auditory tract works, deep understanding of language and lots of computing. Basically the breakthrough was when the hundred mega hertz PC became common. That turned out to be the breakpoint between adequate speech understanding. Since you can't buy a computer that slow anymore virtually any computer is adequate for doing reasonable voice understanding, doing it better will just be a matter of progress. We are already here at a very good level. This means you can talk to people who don't like clicking or can't use a mouse or a keyboard for either physical reasons or preference reasons. You suddenly are talking to people their way much better.

So the new world most amazingly gives you the possibility of using data, using images in a far more convincing fashion.

Audio and video are now common on the web. I assume all of you at some point or another have clicked on an icon, waited a moment and had someone talk at you off a web page. It's not standard but it happens all the time. There are plenty of web sites that have low grade but useful video. We were doing it for the 1996 Olympics for one of our web sites, far better in later sites. This is happening - nothing special but the important issue is it goes from mediocre or worse to acceptable and good as we start putting more bandwidth. So going back to first topic as the network improves, as the network connectivity to the home improves suddenly the quality of the video goes to good. Immediate first use will be video conferencing - we are already seeing the impact of that. Improved product information, fancier stuff available on web sites. You are starting to see some web sites which basically say if you have something better than a 56Kb modem it will send some great stuff. More complex media will have impacts, the relationships between the pieces, the synchronization is crucial and this is all happening and basically why to do people do this? Fundamentally people like talking to people, people like looking at images. We are wired by evolution to enjoy communicating with other members of our species for understanding pictures of the world.