On 1/12/10 5:43 PM, "Dana Spradley" <dana...@oracle.com> wrote:

Not trying to argue this point at the moment--there's no point, since this aspect of the spec is not going to change.

But I will observe that, having initially reacted very strongly *against* the imposed DTD (at the time) coding patterns in DITA, I now recognize them as one of the most singularly important aspects of DITA's design, part of what makes DITA much easier to adapt and use than any other standard XML application I'm aware of.

It is specifically because I *never have to think* about how to structure the DTDs and because I can be sure that any other DITA-knowledgeable person can immediately know what to look for in any vocabulary module I create that makes it orders of magnitude for me to implement, deploy, and maintain DITA-based client-specific applications. That coupled with the now-being-realized ability to automate the authoring of the declarations themselves makes that aspect of DITA completely different. It allows me to provide task-specific solutions to clients who, two or three years ago couldn't have even talked to me because the design and implementation cost would be prohibitively high.

It is one of the bits of genius in the DITA architecture that distinguishes it from all other standard XML applications.

I am saying this as someone who practically thinks in DTD syntax. I have embraced the DITA patterns and they have set me free.

Cheers,

E.