Do we really need to speak the same language?

Business-to-business integration (B2Bi) is old-hat. A robust technology to allow the systems of two (or more) business to be integrated. An archetypal example is a direct purchase order/invoicing system between business customers and their suppliers. In these transactions, orders and invoices are being exchanged automatically and electronically, the relevant data fields being extracted and populated into the relevant ERPS and other databases on each side. Simple stuff and relatively easy to achieve, once both parties agree on the channels and the document structures that are sent to each other (the ‘language’).

The challenges with this approach arise when you have many-to-many relationships in a business ecosystem with lots of different suppliers and customers. Some solutions arise by mandate – the sheer market grunt of one or a few big players that dictate to everyone else what the rules are. But without such a driving force, you either need everyone to agree on some common standards for describing things like purchase orders and invoices, or you have to engineer custom parsing B2B interfaces for each relationship. Or, you just live with the friction and continue processing a subset of your orders via email or fax.

In the case of ‘common standards’, the example of purchase orders and invoices seems achievable universally, because they are items that are, in their most raw forms, common to every business. While such standards might exist in certain industry pockets (RosettaNet for the chip manufacturing industry comes to mind), there is no universal standard for ecommerce transactions. Furthermore, the challenge becomes even more interesting (to us, at least!) when you go beyond the ‘simple’ of purchase orders and invoices to the ‘complex’ – other types of business forms processing.

The home loan and conveyancing industry sectors, groups with which we’ve had some involvement already, are ecosystems in which the electronic exchanges are far more fiddly and fraught than online purchase order processing. The National Electronic Conveyancing System (NECS) will attempt to address some of the efficiency needs in Australia. In this case you have seven or eight state jurisdictional land titles offices, each with their own similar-but-different language for describing business activities, such as transfer of land title ownership. One NECS approach could be to mandate that all state offices use the same language for common transactions. But this would require massive and disruptive changes within those agencies, some of which have been processing documents their own way for over a hundred years.

A more palatable alternative is to let the agencies continue to maintain their own language and create an 8-way dictionary - a giant look-up table, or “universal translator” that supports not only human readability, but (eventually) software integration as well. While a thorough and robust solution would accommodate modern semantic techniques and standards, it is a simple concept and certainly achievable.


Reprint from
FRICTIONLESS BUSINESS ECOSYSTEMS - NICTA