Norm Walsh has blogged a very extensive analysis of how to implement some of the features of DITA in DocBook.
I've not blogged on this topic as much as I would like, simply due to workload and time constraints. I have been looking at DITA quite a bit. At Flatirons Solutions, we have implemented very DITA-like solutions for customers using a combination of DocBook and a content management system (Documentum and IBM Document Manager). Simplified DocBook, IMO, is more compatible with the basic elements of DITA. In fact, I wish they had re-used more of the element names from DocBook, instead of introducing slightly different tag names with the same basic semantics...
DocBook sections are very much akin to DITA topics. That is the typical chunk level we have recommended in our implementations. Then, you have a "master document" that is an assembly of chunks and very much akin to DITA maps. Different master documents can use the same chunks in a variety of orders, in the same way that you can have different DITA maps. The master documents can be at the book, chapter or article level, depending on your business needs.
The biggest advantage of this approach, is that you can leverage existing DocBook tagged content (just chunked at the section level) and take advantage of the wide variety of tools for converting, editing and publishing DocBook. DITA is still relatively new, and has limited support compared to the tried and true DocBook model. You can even assign DITA topic types (reference, concept, task) in the role attribute, or add a DITA topic type attribute through a DocBook customization layer!
FYI, for those attending XML 2005, there will be several presentations and tutorials on DocBook, and Flatirons' CTO, Eric Severson, will be presenting on a DITA implementation we have done with BYU.
Follow-up: See also Eve Maler's interesting post on this subject at:
http://www.xmlgrrl.com/blog/archives/2005/11/02/topic-oriented-architecture/
See also: DocBook
No comments:
Post a Comment