Java is a key element. It delivers the capability to write secure applications on smart cards. As a result, we can: (a) have multiple, "non-cooperating" applications on the same card, and (b) load these applications onto the cards after they are issued.

With Java, we get "Write Once, Run Anywhereª", a major benefit to card issuers and developers who have been working on proprietary operating systems. Java will help reduce development costs and time-to-market for new products and services.

Smart cards have a long track record, mainly outside the US, as an enabling technology for applications involving money and sensitive information. The major applications have been in banking, where smart cards are used to provide off-line authentication and transaction security; health care, where they enable secure access to information by patients; and mobile telephones, where the GSM standard for digital mobile communications uses them for subscriber identity. Today, chip cards of all types by themselves represent a $1 billion business worldwide, split among a handful of card manufacturers.

But, considering the length of time that smart cards have been around - about as long as PCs - their limited usage, especially in the US, is surprising, until we look closer.

Widespread application development has been limited by the technology itself. With 8-bit CPUs and limited memory, typically less than 20K ROM and EEPROM, applications are written in assembly language on proprietary platforms.

Also, the business case for introducing smart cards into the electronic transactions infrastructure is not as compelling in the US. The present authorization system, based on mag-stripe cards, works well because we have reliable, fast telecommunications.

Finally, the smart card industry has been a relatively small, closely-knit group of companies and institutions - insularity born partly from limitations in the technology, but also because of the strong security requirements. Smart cards, in almost every application, represent money or value to the people who carry them and the businesses that issue them - one of the primary reasons they are so appealing. It is also the major reason that smart cards are being pulled into Internet/Intranet applications.

To better understand the future of smart cards, consider the history, albeit short, of the computer industry. One lesson is clear: It is not the computer technology itself that propels new innovations to the forefront, it is the applications. For Apple it was VisiCalc; for the IBM PC, it was Lotus 1-2-3.

Will there be one for smart cards?

Probably not just one. Smart cards can be used in so many different market segments that multiple-application cards will lead the expansion in the US. This allows the costs to be amortized over several applications, the ones we know about today, plus others that haven't even been thought of. Another lesson is that applications do not come from technology providers. The best applications are still to be written, coming from the imaginations of people like you.

The smart card and software industries are providing two important prerequisites for rapid growth in the smart card industry: standards for smart cards to work with PCs, coming from the PC/SC Workgroup specifications, and an open development environment, coming from Java Card API.

The entry barriers for smart cards are falling, making way for new applications in a wide range of markets. With Java, smart cards can join the world of mainstream computing - and after 20 years, it's about time.