Highlights of the 2008 DITA CMS Conference
By Eric Armstrong on Apr 13, 2008
The 2008 DITA CMS Conference was informative, educational, and in many cases surprising. My personal highlights include Daisy, DITAStorm, WebWorks ePublisher, and more...
Here's a whirlwind summary of stuff that, to my mind, had a good chance to compete for "best of show". (For links, see the resources section at the end.)
A WYSIWYG Wiki that can produce books in PDF and single-page HTML. For a full review, go here.
The online, browser-based editor for DITA is all grown up. It now handles all DITA elements and is available in a standalone (desktop) edition, as well as a shared (server) edition. Cost is approximately $400/seat, in unit quantities, which makes comparable to other reasonably-priced editors. There are four versions in all, so you can get the combination of features you need for CMS-integration, workflow management, and version control.
- MarkLogic Server
Still the standard for dynamic content delivery. Instead of having to pre-publish every possible version of your docs, the system creates the version a particular user needs when they visit the site. They might interactively create the document they want, or they may get a document
- WebWorks ePublisher
It's a bit pricey, but it gives you HTML help outputs with popups and menus. It can also iron out some of the rough edges in the DITA Open Tooklit, and make it easier to customize your outputs without having to code XSL transforms.
A Java-based translation memory system built on the open source OAXAL standard for translations. It lets authors use the equivalent of translation memory to reuse text that has already been translated, both to ensure consistency and reduce translation costs. Includes the capacity for multi-lingual spellchecking and glossary managment.
- Staged Production (Keith Schengli-Roberts, AMD)
A terrific talk that documented a quantifiable increase in document quality, at greater speed and lower cost, in addition to localization savings. He found that the focus on individual topics that DITA encourages, coupled with the lack of redundancy, led to more frequent and more thorough reviews. (Small amounts of material often, rather than big documents at the end of the process.) As a result, he found that the time to produce documentation was 40 to 50% of time previously spent, deliverables were 225% larger, and that the average time for project completion was two thirds of previous levels. (The difference between 50% authoring time and 66% project delivery time was no doubt due to the additional feedback acquired during the review cycles.)
To get those effects, his organization created a serious change to their processes. Their goal was to achive the "4 C's"--documentation that was complete, correct, clear, and consistent. To do that, they built "technology transfer" into the project development plan. On average, each developer had a week allocated for communicating information, and the project wasn't considered complete until that task had been performed.
As Keith remarked, "You get what you reward". So making documentation part of the process was responsible for two of the four C's: Complete and Correct. It was then the authors' responsibility to make the communications Clear (and concise), with the editors acting as a final check to ensure Consistency.
Keith also had an interesting viewpoint, best expressed as "WYSIWYG considered harmful". The idea was that, with older desktop publishing systems, 50% of your time could easily be spent on formatting, to make things look right. Editing the tags directly completely removes all temptation to do that. (On the other hand, it's easier to orient yourself in a document that has some visual decorations. But I have to admit that I spend a fair amount of time getting the "look" right in my HTML documents. I'm sure I would see something like a 15% percent increase in productivity if I never had to worry about that.)
- Conversions (Joe Gollner, VP, Stilo International)
A really nice analysis of DITA conversions, covering both processes and problems. The speaker displayed an excellent grasp of ontology- and taxonomy-construction. The combination suggests that he would be an excellent choice for setting up DITA-based system that captures and exposes organizational knowledge.
- Code Reviews (Carolyn Inkster and Shannon Rouiller, IBM)
If you wanted to find out how to maximize DITA quality, this session was a good one. And if you ever wondered whether semantic tags were really worth the trouble, it also devoted some time to justifying them. Writeup here:
- Introduction to XSLT in DITA (Steve Anderson, SalesForce.com)
Great introduction to XSLT and XPath expressions. Also explained the concept of a "shell". Supposed XSLT file X defines transforms for A, B, and C, and that you want transform B to behave differently. You put your new transform in file Y, and then create a new file that imports X and Y. That's the shell. Then, when you do your build, you reference the shell transform instead of the original transform, as you normally would. (For example, instead of specifying the standard dita2xhtml.xsl in your build, you would specify the shell that imports both it and your overrides.) With that system, your overrides are preserved when you download new versions of the DITA Open Toolkit.
- Influencing Success (Andrea Ames, IBM; Meryl Natchez, TechProse)
This was a brilliant talk that highlighted the importance of carrying out a change management project in parallel with your production project. Parts of that process included setting expectations, overcoming reluctance, defusing resistance, and keeping management informed. It also emphasized the importance of having a vision, expressing it frequently, and above all maintaining a positive, inspiring attitude. The talk was an inspiration in its own right.
- 2008 DITA Conference
- Daisy Wiki and Book-publishing System
- DITAStorm browser-based editor
- MarkLogic Server for Dynamic Content Delivery
- OAXAL -- Open Source Standard for XML Translation
- Stilo International
- WebWorks ePublisher
- XML-INTL system for OAXAL-based translation managment