Re: DITA versus semantic HTMlL?

Mark Giffin

Forgot to mention that late-model versions of the DITA Open Toolkit support LWDITA:

On 10/17/2019 8:11 AM, Mark Giffin mark@... [dita-users] wrote:

Eliot's comments are solid. You could also use a gradient approach by using Lightweight DITA. You could edit in the HTML format of Lightweight DITA (HDITA), described here:

You can mix HTML and full XML DITA topics in the same DITA map. You can also use the Markdown version of Lightweight DITA (MDITA) for even simper editing (and fewer features), all in the same map with the others. As part of LWDITA there is also a light DITA XML format (XDITA).

Although an official spec for LWDITA is not yet complete, it should be pretty stable as described in the reference above. Many tools already support it, such as Oxygen and FrameMaker.

Mark Giffin
Mark Giffin Consulting, Inc.

On 10/17/2019 6:45 AM, Eliot Kimber ekimber@... [dita-users] wrote:

The new semantic elements in HTML5 are useful but still quite minimal compared to DITA (or any comparable XML language for documents, such as DocBook, JATS, etc.).

The challenge you'll find with HTML for authoring is that it becomes hard to constrain and hard to add additional semantics except through @class or @data-* attributes, which again are hard to constrain.

One of my main clients until recently (when I got a real job and stopped doing full-time consulting) had made the choice to use HTML for authoring (they are authoring codified municipal code, so there are quite a few important semantics in the content that go beyond "section" and "article"). After a couple of years to refining their markup approach and migrating legacy content to their HTML, for which they have a very tightly-controlling schema (based on the schema used by O'Reilly), they started to realize the limitations of the approach as it became harder and harder to both express the semantics they needed and ensure that authors did the right thing.

When I came on the project I introduced DITA to them but didn't try to push it because the decision had already been made and couldn't be unmade. But at one point my main contact there said "We should have used DITA, shouldn't we?" to which I replied, "Sadly, yes".

DITA markup was originally designed to be as much like HTML as possible to make it familiar to authors already familiar with HTML, so at the topic body content level there's not a huge difference between HTML and DITA. Of course there are a lot of other differences.

Another important consideration is linking and re-use: DITA and DITA-aware editors provide unmatched support for authoring documents that have sophisticated links and reuse, something you won't find in HTML or HTML editors. One of things I did for the client mentioned above was implement a sophisticated linking UI on top of Oxygen XML. It basically replicated functionality that is out of the box in Oxygen (or FrameMaker or XMetal).

All that said, if your content is really pretty simple, doesn't require reuse, and you don't need to do a lot of linking, then maybe HTML authoring is sufficient. But for most environments it probably is not.



Eliot Kimber

On 10/17/19, 7:48 AM, "Da'ud Vyd d.vyd@... [dita-users]" wrote:

HTML5 introduces a few semantic tags. OxygenXML can render DITA or HTML to PDF using CSS. Topic-based writing can be enforced with discipline regardless of the markup system. What is the immediate advantage (if any) of using DITA in a small

business with a single technical writer equally comfortable with HTML and DITA?

-da'ud vyd

Join to automatically receive all group messages.