Sep 05, 2006

In my recent review of XML editing tools, I looked particularly at their support for DITA, especially integration of the DITA Open Toolkit. The DITA OT is a reference implementation of the OASIS specification for "ready-made metadata" in the DITA DTDs and Schemas. Why is the DITA OT important and likely to affect your choice of a CMS in the near future?

A major promise of XML has always been to make content more portable. This client benefit may strike fear in the hearts of some vendors, who depend on their client content being "locked in" to proprietary formats. Microsoft is perhaps the biggest beneficiary of lock-in. And the greatest source of content is still their market-leading suite of Office products, especially Microsoft Word.

XML CMS vendors are gambling that clients who migrate their content away will be balanced by others migrating to their XML-native system, and that the XML marketplace will grow larger overall, a win-win situation benefiting every vendor. This "co-opetition" is one of several aspects of DITA that may change the way content is managed in the future as well as the way content management tools are sold.

Nick Kuppers of X-Hive in the Netherlands points out a major productivity benefit when clients have to migrate content because of corporate mergers. Airline industry clients have million-page manuals that might have to be rebranded and adapted to new equipment realities in months, not years. Open standards like DITA contribute directly to the business bottom line.

In the past, each vendor hoped to sell his or her latest new revolution in technology, "pushing" it through marketing. With DITA, vendors find customers coming to them pre-sold on structured writing, asking what the vendor is doing to support DITA. DITA is a "pull" marketplace.

Buyer Buy-in
But do customers really know what they want? And what they really need? Just because interest comes from the grassroots does not mean it is free of hype. Are they all just after "the next big thing?" Even the best tools cannot solve people and process problems. You may have to fire or retrain your twenty-year veterans who want to write introductory narratives and other heavily context-dependent content.

When customers come asking specifically for topic-based structured minimalist writing, says Rick Schochler of Innodata-Isogen, a systems integrator specializing in XML CMS, they are truly educated customers. They are doing the buying instead of him doing the selling.

Will all the vendors use the same Open Toolkit code to process content that has been marked up with the common DITA DTDs? A "reference implementation" means they could choose to implement their own equivalent code. But the open code is growing rapidly by community contributions back to the OT. If vendors "fork" their code, they will have to adapt it themselves to every new community contribution that their customers demand. It's in their best interest to contribute their own improvements back to the community pool, and many are planning to do so.

Chip Gettinger of Astoria Software, which has been pursuing XML content management since it spun off from Xerox many years ago, says the OT "out of the box" can manage 80% of the average technical documentation project. Where it already has been specialized, like software documentation, which IBM originally built DITA for, it can handle 100%.

Hands-On Management
What Astoria and other XML CMS vendors are doing is putting an attractive, highly usable, interactive, and soon completely web-based interface on the Open Toolkit. They are fully integrating DITA into their menus. They are also moving OT processing away from the desktop to fast remote servers.

The author/editor opens a project, typically seeing a DITA Map, a sort of virtual table of contents look into the content. Suzanne Mescan of Vasont says they connect DITA Maps and the DITA collections (Tasks, Concepts, References, and Topics) with bi-directional links, allowing users to simply drag-and-drop content to build new documents.

When the writer clicks to publish to the Open Toolkit, dialog boxes ask for the desired publication channels for output: Help, HTML, PDF. They also ask for the destinations for the different results. Should files go back to the main repository, be sent via ftp to web servers, and so forth.

At this point, says Debra Boczulak, product manager at XyEnterprise, a contributing tech writer can step away, the entire publication process controlled by a master configuration file, with editors notified, etc. Or a sophisticated hands-on manager can watch interactive reports while the OT builds the publications, dealing with any warnings or error messages on the fly. A complete audit trail logs the process for later analysis.

I visited XyEnterprise recently, invited by JoAnn Hackos, whose classic works on user and task analysis, minimalist topic-based writing, online communications standards, and single-source documentation publishing are a driving force behind DITA. Hackos was conducting her DITA Boot Camp in the Boston area. Camp attendees had the DITA OT in their laptops, typing in command-line instructions to publish their content and watching hundreds of lines of output flying up their screens in the command prompt window.

They came to appreciate, as I did during my XML editing tools study, the benefits of the XML CMS vendors integrating the powerful DITA Open Toolkit into their sophisticated GUI tools.