[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [dita] Potential stage 1 proposal, let steps element nest
For what it's worth - Jang's right that the constraint would be no more difficult. Where today you update the content model of <step> to disallow <substeps> as a child, you would update it in the same way to disallow <steps>.
I've gotten a couple of responses off the list that had either an ambivalent or positive take on this.
Based on those responses, I think some may have understood my initial suggestion as "just get rid of <substep> but keep the <substeps> container". It's not what I meant, but I suppose that would be another alternative -- if <substeps> had <step> as a child instead of <substep>, it would resolve most of the issues I laid out. That said, my original intention was to get rid of both the <substeps> container and the <substep> element, replacing both by allowing <step> to have <steps> as a child.
Regards, Robert D. Anderson | |
E-mail: robander@us.ibm.com Digital Services Group | |
11501 BURNET RD,, TX, 78758-3400, AUSTIN, USA | |
Having just sent in regrets for next week, I'm going to stir the pot with a controversial idea.
I've been asking IBM authors what they want from DITA 2.0. One of the more common responses so far is "get rid of substeps" -- meaning, just let <steps> nest.
The reason we have <substeps> is so that task can enforce (what DITA 1.0 designers considered) good information design. That is: if you've got three levels of nesting, that's too much for a single task, and you should break the task apart. If <steps> nests, that allows task to have 3 or 5 or 20 levels of nesting, just as with <ol>.
I've heard several counter-arguments in the last few days:
- There is no semantic difference between step and substep, so it's frustrating to have two sets of elements
- It prevents copy/paste or drag/drop if you try to turn one into the other
- If a step in one task is reused as a substep in another, you cannot conref the whole step - instead you have to conref every child of the step
- The limitation isn't even effective, because people just put <ol> inside the substep's <info> element
- If an author / a team / a company want to limit nesting, there are other ways to do that, such as Schematron rules or build / formatting rules
Given the design history and potential migration cost, I didn't think the TC would look kindly upon this idea, but then the arguments started to pile up, and I wondered if the rest of you were hearing the same thing.
Curious as to your thoughts...
Regards, Robert D. Anderson | <ecblank.gif> |
E-mail: robander@us.ibm.com Digital Services Group | <17306364.gif> |
11501 BURNET RD,, TX, 78758-3400, AUSTIN, USA | |
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]