Ironically, one of the biggest stumbling blocks to efficient global content management has been the pain of “translating” CMS requests for multilingual content into formats that translation management systems (TMS) can fulfill. Where interfaces do exist, they typically take the form of connectors supplied by a TMS vendor and specific to its own TMS. Such purpose-built CMS-TMS interfaces are expensive for the TMS vendor to build and maintain. Tellingly, it’s always the TMS supplier building and supporting the integration; due to the low volume of TMS sales in a fragmented market, CMS vendors like Documentum or Interwoven let their ever hopeful TMS “partners” do the heavy lifting.

Toronto-based Clay Tablet has been working to remedy this problem with an any-to-any connector. This week the company announced a partnership with SDL for its TMS product, adding to previously discussed integrations with across, Language Weaver, and upcoming ties to other language technology suppliers. We spoke with Clay Tablet’s CTO Ryan Coleman about his company’s technology, partnerships, and target customer.

• Technology. Clay Tablet has taken a message-oriented middleware approach to linking any CMS to any TMS. It offers an open-source “producer” interface for the CMS user to generate translation requests, a “provider” interface for the TMS side, and its platform — what we could call a “translation request broker” (TRB). This TRB acts as a traffic cop between the CMS requester and the TMS provider — it receives requests, sends them to the provider for action, monitors progress, supports reporting, and so on. Coleman says that his team can integrate any system as long as it exposes its application programming interface (API). The across and SDL TMS APIs were reasonably well documented and supported. However, Coleman’s 2005 pre-SDL integration with TeamWorks (the offspring of the Uniscape/Trados GXT) required his engineers to closely instrument the client/server connection, monitor the request-response traffic, and reverse-engineer the private API.
• Partnerships. Clay Tablet started life as a service provider that productized a custom multilingual CMS. It quickly realized that the market for global CMS was past its sell-by date, saw a looming traffic jam in TMS, and realized an opportunity in connecting any CMS to any TMS. The number and variety of CMS solutions presents a huge challenge for the relatively low-volume TMS providers. We believe that the TMS companies will keep writing their own connectors for critical connections (obviously, SDL will manage its Tridion tie), but will leverage technologies such as Clay Tablet’s for the others. This approach will free up resources for enhancing TMS functions to incorporate the trio of translation technology, workflow, and business management services.
• Buyers. The Clay Tablet middleware solution speaks volumes to the tech-savvy manager accustomed to service bus architectures. Coleman says his engineering-centric prospects note that “every other enterprise system talks to every other system; why don’t these translation management systems do the same?” We expect the biggest take-up for Clay Tablet 2.0 will be in the high-tech shops that have led the charge to TMS products sold by across, Idiom, and SDL. Meanwhile, LSPs with their own solutions should consider Clay Tablet as a way to integrate their proprietary workflows with commercial and homegrown CMS products. Further, CMS vendors should take a hard look at partnering with Clay Tablet to open up their workflows to translation services.

Universal connectivity won’t happen overnight. However, this integration with leading TMS providers is welcome news for growing the translation software market from the paltry US$100 million in revenue that we saw in 2006. International buyer demand for their own language will increase the interest of information publishers and aggregators in supplying multilingual content at websites, call centers, and user guides. If Clay Tablet succeeds in finding suitable APIs, it can do a lot of the hard work in eliminating the painful inconvenience of buying TMS solutions based on which CMS they support — or worse yet, paying for each and every integration in larger firms like HP and Sun which use multiple CMS products. Over time, the network effect kicks in when any content manager can interact with any translation tool. At that point, the growing economy of scale will induce other vendors to plug in, thus supporting the creation of a richer global content ecosystem than we see today.

This blossoming of universal access into a much bigger ecosystem has happened before. Relational databases were small players in the early 1980s until suppliers like Oracle (intriguingly née “SDL” for Software Development Laboratories) and Relational Technology (later Ingres, now part of CA) began opening their engines through a succession of APIs like Dynamic SQL, CLI, API, ODBC, JDBC, database SOAP services, and XML-Data. Similarly, the banking industry grew through SWIFT (the Society for Worldwide Interbank Financial Telecommunication, 1973) with worldwide data processing, communications link, and a language for international financial transactions. In the United States, the ACH (Automated Clearing House) allowed banks to communicate and more efficiently process payments. Let’s hope translation management follows this connectivity path to a bigger ecosystem.