Sunday, November 15, 2009

DITA in the real world

First, in case you've missed it, a little background about DITA. It stands for Darwin Information Typing Architecture. It came out of IBM and has a strong theoretical and practical background. If you look back a few days in the blog you'll find some more background.
Initially DITA was seen as stuff for geeks. If you used DITA you must have been a geek. One of the great things about DITA is the Open Toolkit. It does all the heavy lifting for you, donated by IBM but open source and regularly and aggressively maintained and updated. Without the toolkit, in my view, there would be no DITA. The toolkit has allowed vendors to build tools around it, allowed users to generate output and generally turned DITA into a practical and usable technology.
Right now is an interesting time for DITA however. It has substantial uptake but, and it's a big but, it isn't by any means mainstream. I think there are two misapprehensions that are holding back adoption. I don't think these misapprehensions go away but I think they are easily managed. In short the misapprehensions are:
  1. You have to be a geek to use DITA. You have to be able to write ANT scripts and get your hands dirty;
  2. Nobody can author DITA content without a heap of training. Why can't they just let you use Word?
There's no doubt in my mind that DITA is ready for the mainstream, neither is there any doubt that there are substantial organisatonal benefits to using DITA. I'd go so far as to say that every organisation of more than 50 employees can mount a business case for DITA. Many smaller organisations can as well, particularly specialist organisations.
There are also some un-explored or at best under-explored use cases. The most compelling for me is proposal management. Most proposals generated by services organisations are replete with content that has been used in the organisation before. Most proposals would contain at least 60% of content that appears in other recent proposals. Examples include service descriptions, staff CVs, capability statements, solution outlines, descriptions of technologies, case studies, references, legal disclaimers...
Bid time is always busy and prone to errors, staff are trying to manage, often massive, proposals and copying and pasting material from everywhere. Formatting goes awry, the wrong content is copied, outdated content finds its way in and close to deadline time things get messy.
A further concern is that much of the material is written by high value staff. When you find those staff writing again from scratch, or perhaps worse pulling material from their own private archives rather than the organisational repository you know that your costs and quality are under siege.
A study I commissioned at a large services organisation showed that there were a number of major benefits to using DITA in the bid/proposal area:
  1. In excess of 60% of material in any proposal/tender could be treated as WOOO content - write once, once only. Put that another way - in excess of 60% of bid writing costs are wasted!
  2. Bid and marketing staff are continually under pressure from sales staff to prepare mini proposals, one offs and unsolicited proposals. Many of these requests can be turned into "roll your own" opportunities. If sales staff are given the tools to access the repository, select the required topics, order them and then produce them in a Word document then you have serious improvements in efficiency. What's more you lower the cost of responding to clients who say "what have you got to offer in this area?". It become cost effective to respond.
  3. Architects and other high value technical staff can focus on value added activities and fee earning rather than re-hashing the same content multiple times;
  4. Sales staff can generate "print on demand" services brochures and can tailor these brochures to the particular situation it is intended for.
  5. Quality control, approvals and version management are suddenly achievable and easily so.
  6. Bid staff have the capacity to manage greater numbers of bids, their job satisfaction is improved through reduced pressure, improved outcomes and a focus on adding value rather than scrabbling through content.
  7. Oh and the legal team can roll their own too by storing their clauses in the system.
Wouldn't it be nice? When can we do that? The answer is that the tools are available right now and this system can be implemented for a software license cost of under USD$20,000 and a services cost of less than that. That's a DITA solution, CCMS, proposal wizard, authoring software...the lot.
You'd pay for that in 3 months in even a modest sized organisation.
But let's get back to the main issue. How do you mainstream DITA? The simple answer is that you don't have to be a geek. What's more, authoring in any of the current crop of tools - XMetal, Oxygen, XMLmind, easy and what's more it's a relief! Once you are free of Word you'll wonder why you didn't do this before. The DITA DTD is not that complex and for little money you can specialise it just for your organisation.
The biggest single step though is the adoption of a CCMS that manages the content and the production of the outputs. This takes all the worry about links, searching, running the transforms and more away. That single step abstracts you from the geeky heart of DITA and makes it no more difficult than using Sharepoint, for instance.
Our little company is there already - CCMS, single sourcing, multiple outputs, massive content re-use.
The next killer DITA app? The one that, to my knowledge, doesn't yet exist? That's easy. When I have a Web Content Management system that reads from my DITA repository then DITA will have made a further massive step.
I'm talking about getting an outcome similar to what I can get with eZ Publish, Joomla, Drupal et al. At the moment I can generate HTML from DITA. What I want is a system that dynamically renders DITA topics to the web. I want a GUI that lets me place topics on a page - by reference and then renders those topics to the web. If I change the topic, I want that change to be dynamically rendered on the web.
In fact this isn't a Web Content Management system. Rather it's a Website Management System. The content is managed elsewhere. I simply want a GUI templating/content placement engine.
Nobody, to my knowledge, does that now and that includes Alfresco and Hippo.
There's the DITA challenge. But before that let's start implementing DITA more broadly, it's easy.


  1. Getting up and running with DITA looks quite challenging to many individual and small organizations like you say. I agree that a DITA based blog/cms publishing might help mainstream it more amongst technical communicators who are already aware of DITA.

    "Selling" DITA to a general audience is harder, they just want to publish something and likely don't care about the storage format. If such an app existed, its challenge would be really trying to demonsrate and prove how reuse and site management would be dramatically improved.

  2. G'day Brett, thanks for the comment. My experience is that it's no use selling the technology, as always it's the benefits that sell it. There's also an interesting gap. Senior management get the benefits, the technical team get the benefits, but the people in the middle just want a quiet life.
    The critical issue that sells it is the WOOO/ubiquity story. Why would you write the same material again each time you prepare a bid, for every brochure, for sections of your website...?