Dr. Arati Prabhakar
National Software Summit Conference
I was pleased to get the invitation to join you because I think you are at the beginning of a process of deciding where is it that this community really needs to go. The outcome of that process is a very critical one. I wanted to start by talking a little bit about the bigger picture of where the information revolution is going. This is a community that is creating that revolution and you are critical to making sure that it is possible for that revolution to continue and to grow and to flourish. Then I wanted to talk about software's role within that revolution, from a perspective outside of this community but from a broader view. Finally, I wanted to give you information about some of the things that we're doing at NIST, in particular working with the software community.
So let me start with the big picture of what's happening in the information revolution. If you're like me, you're sick to death of all the jokes about road kill on the information highway, so I thought we might try to explore a broader analogy that I think will give us better insight into what the nature of this revolution is. I have tried to understand how we can wrap our minds around the breadth and the pervasiveness and the extent of this revolution, and the best thing I can come up with is to try to learn some lessons from the last major transformation in our economy and in our society, the Industrial Revolution. There's a very useful analogy in thinking about the transformation that occurred in that time. The content of our revolution is very different, but I suspect that the implications in terms of the pervasiveness and the way the revolution transforms every kind of business and every kind of life in our society, offer some very important lessons.
So if you take a look at where we were in the Industrial Revolution 100 years ago, I think I can argue we're at about the same place in the information revolution now. The business of industrialization had been going on for some time. Railroads had been around, mechanization had been around, and the industrial process had really begun some decades before. But about 100 years ago, with the advent of interchangeable parts, we were really entering a new era in industrialization. It accelerated the pace in industrialization. It created the opportunity for economies of scale. It transformed the industrial enterprise from the business of building large, carefully handcrafted systems to a business where people could specialize in one part of that system, yet all the parts would somehow magically fit together and the entire system would operate. From that came the kind of "customer-supplier" relationships -- the structure of our industries as we know them today. So I think it was a very important knee in the curve of industrialization.
Similarly, when you think about where we are in our information revolution today, obviously computing and communications technologies themselves aren't new. We've had some form of these technologies for decades, but today we are going through a transformation that is qualitatively different. It has to do with the amount of interoperability, the usability, and the pervasiveness of these technologies. Today it's possible to think about the context of these changes from a very different perspective than even five or 10 years ago. For many years, when you talked about the information revolution, it was computer scientists and electrical engineers talking among ourselves about it. But today, when we talk about this transformation I find that we are talking with teachers, we're talking with people in the health care business, we're talking with the financial community, we're talking to people on the manufacturing factory floor. It is a pervasive business that is transforming from a technology-driven revolution to one that is drawn by users, shaped by users, ultimately driven by users. I think that in hindsight, it will turn out to be the inflection point in our revolution.
Let's think for a minute about how the Industrial Revolution transformed our economies and our societies. Today there are so many things around us that we take for granted that are a result of that kind of change that happened, driven by technology, transforming our economy and through that transforming our broader society. Consider the labor force. If you think about our work force when this county began, we were 90 percent agrarian. Ninety percent of our work force was employed in the business of creating food from the land. That number today is somewhere in the vicinity of 2-3 percent. I think it's very worth noting that in that transformation from 90 percent to 2-3 percent, we didn't give up our ability to make food. In fact, we still pull more value out of the land and create more food and more resources than even this nation needs. We're a major exporter in agriculture. So that's an interesting change in work force, but we need to look one level deeper at what it meant in terms of our capability. Now think about what's happening in the manufacturing economy. Manufacturing employment these days is about 16 percent of our work force. It's been on a down slope for several years. I think all signs are that it will continue in that direction. A critical question for us as we think about where we're going is what does this mean about our manufacturing capacity? So far it has continued to be the case that we are a major producer even though the employment involved in that job is shrinking. So what's growing? While all these other trends have been tailing off parts of the work force, "knowledge" workers are the part of our economy that is booming. The people that are involved in the business of information are more and more the center of our work force. Obviously that's the direction things are going to go in the future.
Let's look a little bit deeper at what some of the transformations were. I read recently about some of the changes that happened in education in our work force during the period of the Industrial Revolution. At the turn of the last century, about 10 percent of our kids got all the way through high school. The demands of the Industrial Revolution, the kinds of changes it demanded in the work place, drove the needs for education to such a degree that in 25 years that number went from 10 percent of our kids to about 50 percent. Can you imagine that in a period of just 25 years, half of the people in this country were going all the way to high school? I think we see some early signs that those same kinds of changes in education and training will be driven by this information revolution. Some early data from the 1980s already show a significant wage differential, about a 10-15 percent premium for those people who are knowledgeable and skilled in using computers. That's just the tip of the iceberg. So while we're thinking about how information technology changes, how our children learn, we also need to be thinking about what it is that they need to learn because the skills that will be needed -- that are already needed today -- to be a driver and a player in this information revolution are quite different than what they had been just a few years ago. Those are very major changes, but think about all of our social policies. Think about unemployment. Think about Social Security. Think about our anti-trust system. Think about our structure of regulations. Think about where those social policies are rooted, and again and again you'll see that they are based on the fact that we were moving, several decades ago, from an agrarian society to an industrial society to a society where a structure was necessary in order to enable the progress of the technology in our economy. There are very different needs in that kind of a society, and by the time we're done rolling out this revolution, we will again find that the kinds of changes that are needed in policy will be equally dramatic. I suspect that the answer lies not just in turning the clock back, but probably in being very creative about how we turn the clock forward and how we think about what the demands will be in the future.
This is our business at NIST, and it's the business many of you in this room are involved in, the entire technology system that we live in that has helped create this information revolution. Its shape and form are rooted in the industrial revolution, and the changes that happened over a period of decades in the earlier parts of this century. If you think about the role of private industry and technology, the role that government has played in investing in basic research in mission driven technologies, that structure also has its roots in the role that technology plays in the industrial enterprise. That, too, needs to be rethought and it has started to be reshaped as we move into this information age. My point is that in the course of the information revolution, we will find a pervasiveness of change at all levels in our society. There will be no one and no kind of businesses left untouched by the changes that are going on, and that's an important context to keep in mind as we think about the technology side of what is driving this information revolution. It's nice to know where you are going as you are thinking about what it is you are trying to do on a day to day basis.
So what is this information technology? What is the national information infrastructure? What is this information revolution made out of? It's actually a set of tools and technologies and applications all taken together. I like to think of it in the famous three layer cake way, that it's a set of networks and the protocols to make bits move around seamlessly across many, many domains. That it's a layer of middleware, the capabilities to turn bits into data, the things that are used across many, many application domains but fundamentally help us take bits and turn them into valuable information. On top of those basic capabilities are a host of different application domains. There are usual suspects -- education, health care, manufacturing, and electronic commerce -- but I think we will find that as you map those applications, there really is no area that is left untouched.
The role of software across this enterprise simply cannot be overstated. I find that I come out of the physical domain still looking in from the outside of this business of software, and I struggle with a physical technology background to grasp it. One thing I find very amusing is that there is not really an understanding of how very many different roles software plays across this information system. I think the analogy for those of us in the physical world is that software is like atoms. You can do almost anything you'd like with atoms and the function of the materials and the systems you build out of different atoms can be radically different. That's the only way I know to get at the scope of the different ways that software plays a role.
If you look at the three layer cake system, you see the role of software in the network level and the bitways level. You see the extremely critical role of software in the middleware level in terms of creating the tools that allow us to use hardware and use networks and use computers in a way that really makes them powerful tools rather than impediments. Finally, you see the role of software in each of these application areas in transforming the broad generic capabilities of information technology into something that is a powerful engine to change the way we practice health care, to change the way we educate our students, as we change the way our financial networks are operated, et cetera, et cetera.
It's very clear to me that we would not be in a position to even imagine these kinds of changes without the capabilities that this community has provided, but it is also clear that if we are to continue this revolution, if this blossoming of information technology is to continue, there are some very critical issues that the software community will have to deliver on.
Let me try to give you my own perspectives on that. In the same way that we can imagine the entire information revolution an analogy of the Industrial Revolution, there are some interesting analogies for software as well. If you think about the productivity of building physical systems before some of the ideas of the Industrial Revolution came to bear, it was a business that where productivity I imagine had been flat for decades, perhaps even centuries, with incremental improvements. That's because it was fundamentally a handcrafted business. In many regards, we face some of the same kinds of issues for software today. So much of the software enterprise today is still the business of building fine handcrafted systems with all that that implies, everything from the tremendous creativity that is possible to the fact that it's very difficult to have a level of engineering discipline and a level of quality that really ensures that the job will be done and that it will be done well in every instance. Software today, we find in one area after another, is the long pole in the tent.
I like to tell the story of when I had worked closely with semiconductor manufacturing equipment companies. One after another that I visited would say, "Gee, you know, we find we build the system and we try to tack the software on at the end and it never works." Over time we're figuring this out, figuring out how to think about embedding software early on, thinking about that as a high value added critical piece of the system. A few years later when I came to NIST, we went to visit a car manufacturing facility. We were there with one of the equipment manufacturers who provided a critical measurement tool for the factory floor. The night before we went on the factory tour I said to the person from the equipment company, "Gee how do you embed software in your system?" And he said, "So much of our business now is software intensive and an enormous number of our technical people are software people -- even though we still think of ourselves as a mechanical engineering company -- and we think we've really got the software problem licked." The next day we went out on the factory floor and there was their piece of equipment and there were the auto manufacturing people talking to the equipment company. The auto manufacturing people said, "We really love your equipment. We think it's really doing fabulous things for us, but the software really is a problem. It just doesn't work fast enough. It's really the bottleneck in the whole process."
So pervasively, and over and over again, what we still find is that how software is integrated into the business -- be it manufacturing or anything else that we're doing -- is so critical to achieving success. We are very far from having that problem licked. The other parts of the information revolution have been addressed; the hardware side, for example, has been on productivity curves where the amount of capability doubles every 18 months, and we all count on that happening. Software is not yet on those kinds of curves, and some dramatic transformations will likely be necessary for that to happen.
I see some very encouraging signs, and I think they are headed exactly in the right direction. Two things really need to happen and have begun to happen. One has to do with engineering discipline in the business of building software. The other has to do with the new technologies that will enable us, once that discipline is in place, to move very rapidly forward. There is a contrast between building discipline and allowing the creativity to continue to bubble, and that's really going to be the crux of the issue. None of us wants to see a rigid, top down discipline or an approach that inhibits creativity. I would argue the kind of engineering discipline, the kind of process that really will make a major difference in this business, is the sort of approach that enables creativity, that becomes a platform for greater creativity. Tension is actually very healthy.
But once that process is in place and we see some real progress, some of the work that SCI and many of the others have been involved in has pointed us in that direction -- the question becomes, what are the advanced technologies? How can science or the basic understanding that underpins this area, feed, drive and accelerate our ability to build software systems reliably in a way that's really providing tremendous capability?
Let me finish by telling you about some of the things that we're involved in at NIST. We started our lives at NIST as the National Bureau of Standards. We date back to 1901 and since 1988 we've also been involved in some other activities, some new programs, Manufacturing Extension, the Advanced Technology Program, the Malcolm Baldrige National Quality Award. I suspect that we're like every other organization today in that information technology really becomes for us a pervasive tool in the job that we do. Since our business is part of the Department of Commerce, we are partnering with industry on technologies that matter for long term economic growth. More and more of what we care about, what we have to focus on if we are to fulfill that mission, has to do with the business of information technology. Let me give you a brief overview of what kinds of activities are going on in each of our programs, and let me start with some of the things that are perhaps the least obvious.
One of our programs is the Manufacturing Extension Partnership. Our goal here is to reach the half of our industrial manufacturing capacity across the country that is in the 380,000 smaller manufacturing firms. It's an enormous part of our capacity that's really in firms of one or two or five employees up to about 500 employees. These companies tend to be very good at the manufacturing process. They typically have not kept pace with new manufacturing practices that would make them much more efficient. They very rarely have really joined the information revolution and jumped on the technology curves that this community really is driving. One of our key roles in manufacturing extension is to help make that transformation happen.
We work through our MEP extension centers around the country. We end up working an amazing number of companies whose most leading edge information technology is that they have a PC on which they do their accounting system. These are companies that don't have modems. You can go talk to them about the Worldwide Web, feel free, but it's actually sort of not relevant yet and figuring out how to help them make the first few steps so they can become part of the changes that are happening with information technology and manufacturing is one of the roles where we are making some real progress.
There's a fabulous opportunity in the manufacturing community if you can use information technology to transform the way customers and suppliers work together, transform the way you design and manufacture and control inventory. The gains we're beginning to see can be truly staggering. But again, if these smaller companies aren't hooked to that, they will increasingly be left in the dust, they will increasingly be left out of the business where many other companies from other parts of the world really are playing. That's the gap that we're hoping to close.
A little closer to the business of advanced technologies, we've got several activities going on within our laboratories and in the Advanced Technology Program. As the old Bureau of Standards, we were created at the turn of the last century in response to the need for a measurement infrastructure to support the Industrial Revolution. If you're going to have interchangeable parts, the parts all had better be measured on the same system so they play together, and creating that common language for commerce in the business of measurement standards has been our goal within the laboratories. It's a job where we have delivered on the initial promise very effectively. Today, with this information transformation that is happening, it becomes vital that we understand what role we need to play in this new transformation. Some of the analogies again hold if the key issue here is creating a degree of interoperability, scalability, and usability that transforms information technology into something for users. A critical question will become, how do you create that common language? How do you create the benchmarks? How do you create the test beds that ensure that transparency in the system is happening? We have done some early work in that area, but to really try to go after that job, we're in the process now of creating an information technology laboratory at NIST. It will be essential for us to build a much stronger relationship with the entire technical community, the industries and the broader technical community, if we are to deliver in this arena as we have for the manufacturing domain.
Within the Advanced Technology Program, this program functions by funding high risk projects in companies on a cost shared basis. Our goal in investing here for taxpayers, for our part of the cost share, is those broad, enabling technologies that are beyond the reach of private investment, but which can really make a big difference in terms of downstream technical, and more fundamentally economic, opportunities. Again, information technology plays a major role. We have within ATP a set of 11 focused programs. Each of these was defined and proposed by the industry and launched over the last year. If I give you some examples out of those 11 focus programs, it becomes very apparent how important the role of information technology and in fact particularly software is. Out of those 11 programs, five are quite explicitly connected to the business of information technology.
One is health care information infrastructure. The dream here is to address that 20 percent of our trillion dollar health care investment in this country every year. About 20 percent of that is administrative costs, information transfer. Automating and simplifying that portion of health care costs across all the domains of health care could really make a dramatic difference in costs. We imagine that it's possible to squeeze many, many billions of dollars out and not only can it be done without reducing quality, it can be a great accelerator for health care quality. That's the goal of the health care information infrastructure program in ATP.
Another program is digital video and information networks. There are two streams of video flowing right now. One comes out of the broadcast world, the other comes out of the multimedia computing based world. These are very different environments. We're at a point today where in the next few years they will either converge in some ways so that they are interoperable or they're going to cross and continue to diverge. The goal of the digital video program within ATP is to try to see if some of the technical underpinnings for that kind of convergence and interoperability can be laid at this point. We have a program in digital data storage trying to dramatically accelerate data storage capabilities in the digital domain. The digital revolution is opening interesting opportunities in the data storage community where today so much of the competition overseas has a tremendously strong position that it's really much more of an analog-based technology.
American companies are positioned to really leap ahead and provide some of the basic tools. We have a fourth program called Technologies for the Integration of Manufacturing Applications, trying to understand how to create the enabling tools that would allow factory floor manufacturing systems to be created, software systems to control manufacturing on the factory floor. How can you create those systems in a way that really meets the needs of manufacturers in a way that avoids the business of building fine handcrafted custom systems that are not scaleable, that are not extendable, that are not interoperable with other systems? Those are four programs.
The fifth program in ATP is probably the one that's most directly linked to the software community, and that's a program on component based software. Of all the focus programs we have in ATP, this is perhaps our most risky, our most aggressive program, because its goal is to create the opportunity for the business of components in software to be viable. It's really directly focused at this concept of getting to the point where the business of software is really a set of businesses, with each organization specializing, focusing on its particular arena of expertise and then generating components that can be assembled, that can be pulled together in effective systems. The key concepts in this arena, if they work, really do offer the potential to change in a dramatic way the effectiveness and the productivity in building software, so critical particularly for large, complex software systems.
So the Advanced Technology Program, which only started in 1990, is a program that's open to all areas of technology. Its content is determined by which industries really want to do something. We want to make something happen here. I am delighted at the amount of information technology content within the ATP. I think that's natural and is exactly what I would expect. It's a place where we are able to participate with the software community in building some of the capabilities that are really going to drive us into the future.
Let me just summarize by saying again, if you consider the big picture, everyone in this business is engaged in as exciting and as vital a job as I can imagine today. It's truly a business where history will look back and say this is a period when so many of the critical changes, so many of the critical new directions, were established in what will be the information revolution. At the same time, that means that the stresses and the challenges are absolutely enormous. It is very possible that without solving some of these problems we will be in an environment where there's enormous potential -- but where we simply don't overcome the technical hurdles that allow us to continue this great explosion and capability. I'm very pleased to be here to talk to you all. I look forward to continuing to work together on these hard problems.