Standards are like free trade: They allow enterprises to specialize in what they do best, they lower prices, they inspire new services, and they promote the growth of a more integrated and inclusive culture. Also like free trade, the logic of standards is not instantly obvious. It takes patience to grasp why it is good for companies to stop competing and innovating, give away access to their customers and settle on a single set of unchanging rules.As a result, the interest in both free trade and standards rises and falls with time. Perhaps by no coincidence, they were in fashion together during the past decade: There was a consensus that free trade (such as NAFTA) and standards (HTML and others) were both good things that ought to be extended whenever possible. We caught this spirit in 1992 with an article ("Getting IT Together," March 1992) that declared that interoperability, cooperative computing and integrated information environments?all synonyms for standardization?would define the next era of computing.And we were right. During the \u201990s the spread of data and product standards laid the ground on which networking, the Internet and, finally, e-commerce were built. Part of what made that period so fertile was that standards packed a double return when applied to computers. They not only lowered the cost of information exchange?their traditional benefit?but also raised data transfer to a point where computers could recognize the content they were handling and deal with it without human intervention. The most famous example is probably HTML, but there are many others. One underreported but critical instance is the spread of the Standard for the Exchange of Product Model Data, or STEP, through the manufacturing world. STEP is meant to fix a major headache that comes with the complexity and dense information of the modern production cycle: the constant need to translate data, usually by hand, into the different forms and definitions required by each separate step of the chain?from design through manufacturing and beyond. The National Institute of Standards and Technology estimates the cost of all this effort, measured in delay and errors as well as man-hours, at many tens of billions of dollars.STEP addresses the problems with a product-representation language that establishes common terms and functions throughout the cycle, from engineering services to machine shops to foundries. Parties can pass designs around this circle as necessary with each value being entered only once (the value of a bolt STEP identifier, for instance, would allow anybody using STEP to find all the relevant information about a particular bolt in a product). The standard was released in 1994 and is now embedded in most computer-aided design\/computer-aided manufacturing (CAD\/CAM) products. While not every engineer writes in STEP yet, it is widely adopted in fabrication-intensive industries such as aviation, automotive, shipbuilding and furniture manufacturing. As we noted before, data standards are not just machine-readable but machine-comprehensible, and this fact soon invited research engineers to raise their sights. Products typically pass through two stages of engineering: design and manufacturing. The engineers working in the second stage write the programs that tell the manufacturing machines how to do what the design engineer has ordained. If the design engineer orders a half-inch hole at a given location, the manufacturing engineer calculates exactly what that means in terms of drill head pressure, drilling times, degree of precision, angle of approach and so on. Imagine the enormous savings in time and money if the machines could calculate their settings from the design itself, automatically and interactively. Engineering scientists have been pursuing this goal of point-and-click manufacturing, as it is called, for 20 years or more.So long as we\u2019re dreaming, wouldn\u2019t it also be nice if designers had real-time access to running subtotals of the materials\u2019 cost and estimated manufacturing times created by a given design change, or if they could simulate operations dynamically to see if moving parts interacted destructively? Achieving any of these ambitions requires building a system in which a CAD program can query specialized databases about a wide range of topics, including materials, machining sequences, prices, physical properties and so on, and incorporate the answers it gets from one database into the queries it sends to others.STEP permits users to write databases with this high degree of interoperability. The idea naturally follows using the Internet to sell STEP-compliant data to manufacturing enterprises around the globe. (STEP Tools is one of the companies that has embarked on this business model. Its databases were used in a General Dynamics-sponsored demo last August in which tank parts were milled directly?point and click?from a CAD file.) As these databases get more useful and numerous, manufacturing everywhere will become smarter and faster and cheaper. It will be a permanent revolution.STEP is an example of the kind of success that gave our 1992 article its upbeat tone. If we wrote the same story today, however, that tone would need to be less rosy. Lowering the cost of information exchange does not benefit everyone. It hurts the plans of companies that hoped to sell products at the old prices, for instance. During the past few years many groups have begun to respond to the concerns of traditional vendors by strengthening the relative position of privately owned versus collectively owned intellectual property. Both copyrights and patents have acquired new protections and extensions. And there appears to be a growing tolerance of private ownership of data standards. (Even the W3C, the consortium responsible for developing standards for the Web, announced in August that it was thinking about endorsing proprietary standards.)The new emphasis on intellectual property has resulted in patent protection being sought for methods that would once have been left in the public domain. And governments are more reluctant to see national strengths given away. "Standards are often built on a collective agreement by the companies involved not to collect royalties on their [intellectual property] positions," says Dave Lindbergh, a consulting engineer for Polycom, a conferencing systems vendor in Milipitas, Calif., and an old hand at the standards game. "A single holdout can wreck the whole process. It\u2019s a serious problem."Like free trade, the benefits of standards are no longer beyond challenge. If we wrote our 1992 article today, we would not be able to assume the same level of consensus we did then. Increasingly, like globalization?the new name for free trade?standardization has to win its battles one at a time.