Contrarian Minds: Jon Bosak

How a simple new standard, known as UBL, could take over the world.

By Al Riske

15.Jul.04--The odd thing about UBL, the universal business language that Distinguished Engineer Jon Bosak has been shepherding through the standards world, is this: If you weren't paying close attention, you'd think it was already a fact of life.

But the dirty little secret of electronic commerce is that it's not nearly as automatic as many people think. Most of it doesn't even happen on the Internet. Over half of what we call e-commerce is still transacted over old-fashioned, proprietary, and expensive electronic data interchange (EDI) networks. Even for companies that use the Internet, there are still a lot of fits and starts where data has to be rekeyed because of incompatible business systems. In many cases, it's little more efficient than doing business by fax.

With UBL, Bosak says, "we're talking about one company having its back-office software automatically issue a purchase order to another company, have it go over the wire, and have the other company's business software understand that."

Sounds a lot like a Web service, doesn't it? Well, it is. UBL is the basis for a simple kind of Web service -- automating purchase orders and invoices -- which is why Bosak loves it. It's also why a lot of industry analysts and visionaries have been saying it's not enough. It's the wrong approach. Pay no attention.

Years ago, Bosak was the driving force behind XML, the extensible markup language that most people agree is now one of the most important building blocks of Web services. A modest, soft-spoken man, Bosak will tell you that others did most of the technical work, but anyone who was there knows that XML wouldn't exist if not for his vision and leadership.

Bosak contends that UBL, which is based on XML, is exactly what the world needs right now. Not too little, and not too much.

"It's my claim that there is an enormous need for this, that real businesses will seize upon it and deploy it widely. I've been saying that for four years now, and it hasn't been tested, the part where I say it will take over the world," Bosak says. "In the next year, we're going to see about that. Either I will turn out to be right, in which case I will be hailed as a prophet, or I will turn out to be wrong and be hailed as a doofus."

The naysayers take a longer-term, theory-driven view. A standardized data format won't solve the whole problem, they say, because the form of the data will differ depending on how business processes are constructed. To solve the whole problem, you have to look at process standardization.

"The basic problem there is one scope and time," Bosak says. "It's true that, in theory, the grand unified theory that makes all this work 30 years from now will be based on process standardization. But as a practical fact of how standards work, it is far more difficult to get a group of people to decide on a standard for processes than it is to get them to decide on a standard for data."

Setting process standards is more difficult for technical reasons and, more importantly, business reasons.

"There's a limit to how much standardization real businesses actually want in their processes, because their competitive advantage lies in the fact that their process is a little different from yours -- and it's different in ways they don't want you to know about," Bosak says.

"So UBL is saying, 'Okay, we're not the final solution to anything. We're not attempting to be theoretically sophisticated here,'" Bosak says. "All we're trying to do is take what has been standard practice for several centuries now, and automate that to the level where you get a real return on your investment. This is the next step forward."

Bosak is clearly a pioneer, but he focuses on the practical rather than the far-out.

"None of the stuff I've done for the past 15 years has been new. I wouldn't be interested in it if it were," he says. "I like sure things."

In the standards game, where consensus is everything, that's not just smart, it's essential. Stakeholders typically have solutions in place that work within their shop or with a small set of partners, so they don't want something really new. But they do want to be able to do more business online.

Since the order-to-invoice cycle is central to business and contains only what's needed -- nothing more -- Bosak sees it as the perfect candidate for standardization and automation.

He's not alone. When Bosak talks to business leaders around the world -- and he's been standing in front of a different audience every couple of weeks for the past two years -- they all say, "Yes, that's exactly what we need."

"The problem they have now is that a company like GM, for example, has been dealing automatically with its suppliers using EDI for many years. The system works fine, but the cost of implementing EDI means they are only trading with other multinationals," Bosak says. "So they do not have an ideally competitive situation. Small suppliers that would compete with their big suppliers can't get in because they can't afford to implement EDI."

UBL's big advantage is it's far cheaper to implement. It's an open standard and it works over the Internet, not an expensive private network.

"The customers like it," Bosak says, "and when you find something the customers like, you're on to something."

So what's the holdup?

In one sense, there is no holdup.

"We already have XML languages that have been defined for use in vertical industries," Bosak points out. "There's the RossetaNet for electronics, CIDX for chemicals, FpML for financials. The Open Applications Group is a set of XML documents that's been widely adopted in automotive."

The problem, he says, is when you try and do something across industries. The various XML languages don't translate well.

Contrary to popular understanding, XML isn't really a markup language, Bosak says. It's a language for defining languages. So there are many XML languages, and while they share a common alphabet, they may be as different as German and French.

"The only practical solution seems to be the definition of an interchange format into which the others can be mapped. This is not an easy solution," Bosak says, "but in 30 years of markup language design, no one's come up with a better one." Bosak believes that very large players that have to interoperate with more than one industry -- the government, for example -- will take UBL and say, "If you want to deal with us, this is what you'll use."

In fact, the government of Denmark was first to publicly announce adoption of UBL, even before OASIS -- the Organization for the Advancement of Structured Information Standards -- made draft 1.0 available in May.

To Bosak, UBL is a really obvious approach to a specific problem. "The more special-purpose something is, the simpler it can be," he says.

The advantage of simplicity? Look what happened when HTML came on the scene.

"I was an expert in electronic publishing at the time and HTML caught all of us completely flat-footed because it was too simple -- much, much simpler than the systems we were using," he says.

But that simplicity enabled HTML to take over the world of online publishing, and Bosak believes UBL can do the same in the world of business.

In other words, history may be about to repeat itself.