DITA versus DITA-OT!

For years, almost everybody in the world of DITA has at some point confused DITA and DITA-OT. Who owns what? How are they developed? As one of the few working on both, I may also be responsible for at least some of the confusion. So, here is my view of the differences. Also, diagrams.

Table 1. What is DITA? What is DITA-OT?
What is DITA? What is DITA-OT?
  • Open standard maintained by the OASIS standards organization. It's a document standard - an interchangeable way of encoding content.
  • The standard defines the language as XML elements and attributes: what does <section> mean? What is the syntax for @conref?
  • The standard defines a number of features: how do you resolve keyref? How do you filter based on attributes?
  • The standard comes with grammar files (DTD, XSD, RelaxNG) that enforce the rules.
  • Open source project developed by "the community". It's software - you use it to do something.
  • It is an implementation of DITA. Its purpose is to process DITA content and DITA features according to rules set out in the standard. It is not the only implementation.
  • It ships within many (but not all) DITA products, such as editors and content management systems, as a way to integrate publishing support within those tools.
  • Like most DITA implementations, it ships the standard grammar files as defined by OASIS. To DITA-OT, these are much like other required packages (Ant, Saxon, etc) that are bundled as-is to make it work.
Figure 1. Venn diagram representing DITA (the standard) and DITA-OT (the open source project). They are not the same.

Venn diagram: DITA in one circle, DITA-OT in another. No overlap.

Table 2. Who makes DITA? Who makes DITA-OT?
Who governs DITA? Who governs DITA-OT?
  • The DITA Technical Committee at OASIS owns the standard - the DITA TC is responsible for developing and publishing new versions of DITA, along with occasional committee notes that help inform about or explain the standard.
  • Every TC member must be a member of OASIS (at the corporate level or as an individual).
  • Currently 19 voting members, and a lot of non-voting members. I'm one of the voting members.
  • Background of TC members is primarily in liberal arts; last survey done in 2014.
  • Anybody can participate in the DITA-OT project.
  • The more you participate, the more say you have in the direction of the project.
  • Currently, the three most active contributors are Jarno Elovirta (not an OASIS member), Robert Anderson (me) (OASIS member representing IBM), and Roger Sheen (not an OASIS member).
  • Between the three of us, we have a mix of training in language and communication (German, Russian, multimedia and printing), computer science, and fine arts.
  • While I'm the only person squarely in the middle of DITA and DITA-OT, Kris Eberlein (chair of the DITA TC, OASIS member representing Eberlein Consulting LLC) and Eliot Kimber (DITA TC voting member) also participate in the DITA-OT community. Most other current and prior DITA-OT participants are not affiliated with the OASIS DITA TC.
Figure 2. Venn diagram showing overlap between DITA developers and DITA-OT developers. Yes, that's me in the middle. With a baby and a lemur (but not a baby lemur).

Venn diagram: DITA and DITA-OT membership with slight overlap, and me in the middle.

How does the OASIS DITA TC direct work on DITA-OT?

TL;DR version: it doesn't.

The OASIS DITA TC sets rules for how DITA works, such as how to process the @conref and @keyref attributes. DITA-OT and every other DITA implementation must follow those rules. In that sense, DITA-OT (like any other DITA implementation) is guided by the OASIS DITA TC.

The OASIS DITA TC has no control over other aspects of DITA-OT, such as default styles or processing parameters, just as it has no control over how an editor or CMS renders elements that use @conref.

How does DITA-OT get updates to DITA itself?

TL;DR version: it doesn't.

For each new version of DITA so far, DITA-OT was the first implementation to try and implement a lot of new features (not all of them, but certainly many). Experience implementing those features led to feedback on the specification. For example, with DITA 1.3, Jarno Elovirta (DITA-OT's lead developer, and again, not an OASIS member) sent in many public comments as he tried to test out and implement the early draft specification. Some of that feedback led to changes in DITA 1.3 – but Jarno himself had no vote in whether or not his comments were accepted. Whether to accept his suggestions was entirely up to the voting membership of the OASIS DITA TC.

Of course I personally play a confusing role as both a developer of DITA-OT and an editor of the specification at OASIS. Without question, my experience working with DITA-OT influences how I evaluate new feature proposals at OASIS (as I think it should). At the same time, my responsibility at OASIS is to represent my employer (IBM), not DITA-OT. As editor of the specification, my guiding principle is to ensure that the DITA specification is defined in a way that is not dependent on (or a barrier to) DITA-OT, or any other DITA implementation.


There are a lot of reasons to confuse DITA with DITA-OT.

  • Both came from IBM at the same time.
  • The names are, admittedly, quite similar.
  • Both are "open" (I've heard longtime users mistakenly call DITA "open source"; I've also heard DITA-OT mistakenly called an "open standard")
  • Both rely on volunteers (or on companies that volunteer employee time). Come join either one! Or both!
  • DITA-OT is bundled with so many DITA tools that some have suggested it should be a standard, or a "reference implementation", regardless of other implementations. Possibly of interest: I've heard that suggestion from participants in the OASIS DITA TC; more often, I've heard it from people who do not participate in either group. I've never heard that suggestion from current or recent contributors to DITA-OT.
  • Most DITA authors are working with one or more vendor tools. Very often, those are marketed as DITA tools that support DITA using DITA-OT. It's never easy to tease apart such differences without getting really deep into the details.

With all that, it's not surprising that DITA and DITA-OT are confused. Regardless: they are not the same. There is no official connection between the two. While there is some overlap in purpose (support content using DITA), and at times in participation (ahem), neither group can or should direct the other.