Last night the PDX DITA faithful met to discuss recent noise about where DITA fits into a world where new developer-friendly tools are proliferating and gaining more functionality, and where some CMS vendors have been suggesting that CMS functionality makes DITA unnecessary and perhaps obsolete. To be fair, everyone at the table is working in DITA so we’re not free of bias, but it was interesting to talk about which use cases work for DITA and which don’t, and some of the reasons why other people don’t find DITA compelling.

A few themes emerged from the discussion:

–DITA has a lot of complexity and a big learning curve, but most of us have experienced the same problem with any tools as soon as the complexity and scale of projects increases. For a small project without a lot of content reuse, or multiple editions and versions in flight at once, or just a few writers, most tools do pretty well; when you start scaling up, you get into challenges and the workflows get less simple.

–It’s hard for people who don’t maintain information at scale and over time to understand how maintaining it can be as challenging and complex as maintaining code. Because most writing is ad hoc, created by non-specialists, and serves only one audience at a time, people who aren’t content or tech docs professionals are attracted to tools that present writing as text without apparatus. (A number of attendees had experience with showing someone in another job function the Oxygen XML editing interface and watching them recoil.)

–But those of us who are responsible for maintaining thousands of topics with overlapping content and producing and maintaining output for multiple audiences are kind of cynical about the value of hiding the complexity; it’s lurking underneath the covers anyway, since authors inevitably become responsible for the implicit context and future of what they write. (This is why I, personally, roll my eyes a little about the idea that DITA “frees writers up to just write”: as soon as you write something you’re making some tradeoffs about, for example, how audience-friendly versus future-proof and reusable it is. That means you probably have to know something about what’s going to happen to your text after you write it, whether you are yourself tagging any topic references or running filters against it, or not.)

—Tom Johnson alleged that developers don’t like XML and this led to a long digression about what developers will and won’t do when providing content, and whether this is any kind of blocker to developers’ creating documentation. There was some consensus that large repositories of developer-created content probably won’t end up being created in DITA (unless there is some kind of form-based interface intervening) but some attendees had had good experiences working with developers who wanted to single-source UI text and documentation text and saw XML as a logical choice for UIs to consume.

–Attendees had some useful responses to the allegations that a CCMS can replace DITA itself: when you do that, you’re back in the world of help authoring tools because insofar as the logic of the CCMS overrides the logic of DITA, you’re going to end up with functionality that is not portable once you take your content out of that CMS. For example,  if you rely on the conditional processing of your CMS for filtered output, you will have to recreate that tagging in your DITA source if you start over in a different CMS. This creates the kinds of uncomfortable relationships with tool vendors that we have talked about before, where there is a cost to moving away from a tool, or you’re waiting on bug fixes to the tool to solve your problems. (I think I have that right. We use SVN and have never used a CMS.)

–There were many tacos. So many tacos.

As usual, it was a great and collegial discussion. Thanks to everyone who participated!