Do you have doubts about DITA?

by Gordon Dennis

May 2011

Perhaps revisonary and indeed reactionary thoughts are inevitable whenever change is mooted . Concerning DITA, there's certainly a lot of ‘motor-boating’ going on (this may be identified simply by its characteristic sound of "but but but .......").

It’s time to reflect for a moment that saying "No" is sometimes the simplest answer – some people actually believe that it’s always the best answer. After all, it absolves you of responsibility and puts off the time when you have to do something - hopefully until such time as someone else must bear responsibility.

On the other hand, saying “No” is also often not the best , most effective, or most efficient way of running things.

Not that DITA is without issues. For example, the eLearning capability in DITA 1.2 is hardly what's needed; come on, we need to be able to generate SCORM compatible eLearning materials from DITA topics - directly and simply !

Also, the DITA Open Toolkit is an untamed beast – I’ll come to that in a minute.

Having said all that, though, DITA is actually much better than its critics give it credit for: let's take a look at the competitive field:

Dita itself, comparatively the new kid on the block, having been born in the 21st century.
DocBook which as its origins in the late 1980s
S1000D which has even earlier SGML-based origins

Can DocBook hack it?

DocBook is rooted in printed documentation; almost its entire raison d’etre in fact.

Topic handling capability is a bolt on attempt to catch up with DITA, but at the end of the day, in an age where people want information on demand and take all kinds of mobile devices for granted, it just won’t do anymore.

So what about S1000D?

On the face of it, the most comprehensive approach available. Version 4 is not just XML only (finally biting the bullet and ditching SGML), but is also XML schemas only.

It has the best Applicability facilities in the industry. And S1000D is a member of a whole family of standards, including S2000M for materials management; S3000L for procedural documentation; S4000M for maintenance programs; and S6000T for training applications.

Two problems with S1000D

Problem 1 is the Learning curve: if you thought that DocBook - or DITA - had a steep learning curve, take a look at S1000D - steeper, longer, tougher.

Problem 2 is the Cost of ownership. Currently available S1000D toolset prices are astronomically high. The Defence industry legacy has inflated prices beyond reach of most commercial organisations - even the largest; the Nimrod MRA4 syndrome writ large.

Because of steep learning curve, S1000D eployment takes months and years. It’s a great way to employ costly consultants, so the total lieftime cost of ownership is very high.

XML and the eLearning conundrum

In reality, most tools that provide SCORM capability are base on proprietary formats

A classic eLearning tool, one used all over the world is Articulate. This is closely integrated with PowerPoint (itself a proprietary format) but is in no way open.

None of the three standards compared here can act in this scenario. You can' take in a DITA topic; a DocBook Section; or an S1000D Data Module into Articulate and generate SCORM compatible output in the way that you can generate, for example, a Windows Help file (.CHM) from those sources.

Back to DITA !

So, as I said at the start, DITA is actually much better than its critics give it credit for

It's the only one of the three standards considered here where "minimalism" is cited as a virtue and as a central part of the design paradigm. And as we all know, we should strive to "keep it simple, stupid"
You can actually explain what DITA does to a non-technician and have he/she up and running in a matter of hours, not weeks and months.
It’s not “big bang” – it’s inherently scaleable so you can start small, delivering benefits very rapidly, then work up in terms of sophistication of the solution.

We need to make a clear distinction here between DITA itself and the DITA Open toolkit.

DITA itself

Is a mature, simple and yet powerful way to split information into topics.
In our industry that makes compliance with standards such as CAP676 much easier.
Is one of the most straightforward ways around to re-use XML content.

DITA Open Toolkit

Is definitely untamed - configuration is very critical, as is the file system structure.
No really good tool available for end user means that configuring the OTK is a technician's job. This of course makes it more expensive, introduces delays and requires careful documentation.
Note that there is an alternative to using the OTK. Aimed at medium to large processing throughputs ‘TopLeaf’ is available from Koala and provides the user with a far less ‘technical’ user interface.

So what needs improving in DITA?

A wider range of topics from the current 3. Some S1000D lessons learned could be applicable !
eLearning facilities need a step input, not just a ramp up. We need to be able to generate SCORM compatible eLearning materials from DITA topics
OTK must be tamed!

The last word

The problem with all three standards is that none of them are "all things to all men". That's because they simply cannot be; so attempts to "prove" that any one of them does have that characteristic are doomed to failure.

So we should stop trying to do this and concentrate on business imperatives rather than technical elegance; on benefits rather than flashy features; and on opportunity created rather than problems encountered along the way.