Cobol, the language that became the standard for commercial software development from the early 1960s right through until the late 1980s, is now seen as something of an embarrassing reminder of the past, what with the Y2K problem (a boon to all those once- mothballed Cobol graybeards, fast coming out of retirement and buying their nursing homes charging consulting fees to fix selfsame Millennium Bug). But Cobol proved enormously useful in creating a lingua franca for business software development, especially after IBM’s historic unbundling in 1969, which created the independent software industry overnight. Now some proponents of electronic commerce are talking of the need for a new Cobol to achieve something similar in their fledgling market – only this time it is not going to be solely an application development tool, but part of an architectural framework. CBL, Common Business Language, may get off the ground
By Gary Flood
in the near future, hand in hand with other proposed mechanisms, if the Palo Alto-based CommerceNet consortium of 500 firms and organizations worldwide (though 150 is the actual figure for hard-core committed members), according to its founder and chairman, Jay Tenenbaum. Tenenbaum, speaking at this week’s Business Online 97 e-commerce conference in San Francisco (CI No 3,190), believes that for e-commerce to really take off we need a much more aggressive definition of openness than we have been used to putting up with. The ultimate example would be if a customer wanted to standardize on some future United Parcel Service (UPS) protocol for all e-commerce shipments, but occasionally chose to use its rival FedEx as well, then UPS would have to swallow the idea and make sure it could talk FedEx’ code. In his view, this is the only way to create what he calls a massive economy of online services and Internet markets (iMarkets), all linked through a common framework which enables them to utilize and add value to each other’s services. To get sight of why Tenenbaum wants to do all this, think of all those zillions of Web sites out there which know nothing about one another. Say you’re in the market for a new loan, so you use some sort of search engine to look up what may be available on line. And after a few wrong turns and some pages in Swedish, you may find what you need. But maybe you wanted the loan to buy a car. So you repeat the process. And then you remember you need some car insurance, and so on and on. It would make much more sense to either have more intelligent search engines that don’t just perform blind search but could group a number of relevant hits together for you, or some enterprising young buck should group a number of such sites together, creating virtual businesses or sets of businesses – hence, iMarkets. The reason the first solution doesn’t really work right now is that there is no universal catalog for storing all the ever-growing mountain of information out there, no one ueber-cyber White Pages; the reason the second doesn’t work is that Web sites are atomized, unable to intercommunicate, and the only way that two can be connected is by hand-coding links. To solve not just one but both problems, CommerceNet is proposing eCo (from e-commerce). Interoperability – a framework that will define the components, interfaces and processes – is the job of eCo and CBL, the new Cobol. This framework has to be broad enough to gain wide acceptance and support within the industry, engage the user community, and be conducive to the development of effective standards, Tenenbaum admits, but believes that the strategy CommerceNet is using – trying to create a buzz among both vendors and end-users to create this (still vaporware) eCo framework is the way forward. So eCo, to be based on emerging standards like S/MIME and IIOP/CORBA (Common Object Request Broker Architecture), is a modular, object-oriented architecture, and will consist of applications and services that model real markets and business processes, with the Common Business Language (CBL) enabling applications
to communicate using messages and objects analogous to those used in real commerce, with an extensible set of interface specifications, class libraries and network services so applications can be quickly assembled from existing components and subsequently reused themselves, and finally, a layer of middleware that insulates applications from each other and from platform dependencies. Every eCo system application would be a network-accessible service, those services to be in turn provided by distributed objects; and furthermore eCo system objects would have to be bilingual, able to respond to both everyday http requests from browsers and CBL messages from agents. The eCo System framework will also have to leverage commercial Internet platforms such as IBM Corp’s CommercePoint, Netscape Communications Corp’s ONE, Oracle Corp’s NCA, and Sun Microsystems Inc’s JECF. But CommerceNet believes it can do all this by not trying to create new standards but emphasizing interoperability (aggressive openness, remember), piggybacking on existing de facto standards like Java, IIOP, protocol negotiation, gateways, and mediators. In addition to eCo and CBL a new meta data layer should be added to each and every Web page, so that both new and legacy sites could identify themselves, enabling the speedier compilation of that proposed universal catalog to build up a more useful picture of what’s actually out on the Web, and also to make it easier for those potential iMarket entrepreneurs to build their virtual Malls. Ambitious, certainly, and very much dependent on appealing to the better natures of e-commerce vendors. To which criticism Tenenbaum has an answer: Some people don’t want to be commoditized, especially those that have been building up their own distribution channels [for e-commerce]. But if they won’t, the number two’s in each market will do it for them. Eventually, buyer demand will drive the market this way. It’s also worth pointing out that CommerceNet is not too good at being proactive with the press and getting its agenda out there for discussion, with a rather confusing Web site and no apparent press release collateral actually on eCo. Despite such challenges, CommerceNet is pushing ahead regardless, with a so-called Mall 2000 demo project underway for summer rollout, intended as a showcase of eCo and CBL potential. That date, incidentally, reminds us that Tenenbaum says CommerceNet people like to pronounce CBL as Cobol because the end of the century is near and we figure we need a new Cobol after the year 2000. Believers in the need for such a new lingua franca are referred to https://www.commerce.net.
Content from our partners
