Programmable systems are the next 'killer app'

Programmable systems are the next 'killer app'
By Larry Cooke, EEdesign
March 20, 2003 (7:46 p.m. EST)
URL: http://www.eetimes.com/story/OEG20030320S0064

In the last decade of the 20th century we saw the proliferation of digital technology not just in our communications through the Internet, but also into everyday items such as cars, phones, ovens, cameras, and exercise equipment. This explosive introduction of technology is profoundly changing the IC industry and society at large. And we, the engineers who are creating it, aren't always fully aware of its implications on future design.

The question on the minds of many people today is...what's next? What's the next electronics-based technology that will positively impact the economy? I believe that whatever the particular electronics hardware and/or software applications that drive growth and change, they will be based on the concept of meta-applications and "programmability." But before discussing that, let's take a look back to the beginning of the PC.

The PC wave
The recent and "still going" application that pushed tremendous integr ated circuit (IC) growth was the personal computer. In fact, the PC is largely responsible for broad public awareness of semiconductors. Earlier waves of IC industry growth were limited to business and military applications.

However, in the past 25 years since the introduction of the PC, a generation has grown up accustomed to interacting with computers as part of their everyday lives. But the acceptance of hardware (and resulting awareness of IC in the consumer world) was due to the introduction of graphical user interfaces (GUIs), which hid the complexity of computers from the average user while allowing them to perform tasks such as writing calculating, and storing and retrieving lists in a more efficient manner.

In that wave, the PC replaced typewriters, adding machines, paper spreadsheets, and large index card files because the alternatives were intuitively obvious. Despite these great advances, the PC and technology behind it didn't change the way we communicate.

The next wave: univ ersal digital communications
The next wave of growth involves the ability to rapidly computerize and integrate universal digital communications with everything. We are already seeing this start with the convergence of the Internet in our cars, TVs and cell phones, but that is just the beginning. It is now possible to integrate miniature mechanical sensors/controllers with wireless network communications, and computers all on a single chip.

As the cost of these devices comes down, we will see an explosion of common everyday "networked and smart" products such as: wireless-computerized sewing machines, vacuum cleaners, showers, door bells, woodworking tools, and even the clothes you wear. This next wave will be able to computerize and integrate virtually everything in our world.

While this is the direction we are moving in, the Dot Com bust showed us that society will not just accept technology for it's own sake. It must be cheaper or better and easier to use than what exists today to be acce pted. To understand this better let's look at what has worked on the Internet.

The Internet has and will continue to replace older technology in the areas of consumer trading (auctions), product sales, general information, education, and personal communication. Still, the need for integrated digital products has been limited to cameras, cell phones, TVs, stereos, and interactive games, where digital technology has provided products that are better, cheaper, and/or easier to use than before.

It is easier, quicker and cheaper to send a digital picture of your used car from your camera to a potential buyer via an online auction than buying the film, taking the picture, getting it developed, and sending it to a auctioneer by traditional mail. In general, Internet applications have been accepted when they put more control directly into the hands of the consumer.

Proliferation requires programmability
For this proliferation of the Internet to continue, we must be able to transfer knowledg e from the product expert to the consumer in ways that make the products easier, quicker and better for the consumer. This means the product experts must be able to "program" the products so the consumer can use the "programming" of their choice. For example, programmable cookware could allow consumers to recreate professional chefs' creations, if the chefs could "program" them and consumers could use the "programming." Unfortunately, there is such a wide diversity of niche products in the world that economically integrating them is impossible with the methods we are using today.

The rising cost of designing and prototyping large ICs, coupled with the increasing cost of applications software development, would prohibit the creation of most of these products. The industry today can only afford to build these complex Internet driven products when they have a proven market in the hundreds of millions.

Internet integrated niche products will only become viable when the industry creates standard meta-GUIs , a standard meta-control language and highly connectable, programmable controllers for these products. The meta-GUIs will allow engineers to quickly tailor expert and consumer interfaces, and the meta-control language will allow engineers to quickly develop product specific code for their programmable platform derivative IC chips. The programmable platform derivatives will either be minimally mask programmable or entirely field programmable so the actual product development is limited to the GUI interface definitions and the application specific hardware and code development. Prototyping costs can thus be kept to a minimum.

Secondly, the Dot Com bust also taught us that technologists can't predict what consumers will accept. We engineers need to change our "build it and they will come" attitudes. As the IC technology has spread into consumer products, we engineers presume that the average consumer would like what we like. In general, that is not the case. Many consumers are annoyed with technology. It changes too quickly, it is not understandable, and it doesn't work for reasons they don't understand.

For example, the Newton failed because it didn't work as well as the consumer wanted, but the Palm Pilot succeeded because consumers (to the surprise of many engineers) were willing to modify their writing habits in order to improve the input recognition. Lesson learned: Engineers must rely more on customer feedback and input when designing the next generation information age products. This will also require far more programmable products than we have today, because a large part of the design will be done after consumers and experts have had a chance to use it.

The attitude "IC software and hardware engineers can solve anything" will also need to change. The best products will be those that both the consumer and expert can modify, fix and use, ideally without instructions. Last week, I bought and successfully assembled a set of drawers from instructions with no words. Yet, try as I might, I can't get my home network to work when it breaks, even with instructions, and I'm an engineer. Also, I may be able to program my computer and wire my house, but I don't know how to cook fine French cuisine, so a chef, not a programmer should be the one to "program" recipes for integrated cookware.

In short, the next "killer application" will likely be meta-applications, the next generation ICs will be programmable, and success of the industry hinges on getting non-engineers to "program" the products. Clearly, we have a long way to go.

Nevertheless, the Virtual Socket Interface Alliance (VSIA) is taking steps today through its new Platform Based Design and Hardware-dependent-Software efforts to insure that the issues raised by some of the new technologies are openly discussed and resolutions are identified.

Larry Cooke is vice president of marketing of the VSIA.