A handful of lint errors in chapter-3-1-19

[robinwhittleton]

I‘m running lint from the master branch so maybe this is newer, but I’m getting a couple of errors about the dl:

s-061 │ Error │ chapter-3-1-19.xhtml │ <dd> element without block-level child.

and:

s-062 │ Error │ chapter-3-1-19.xhtml │ <dt> element without exactly one <dfn> child.

@acabal: got any input on the second?

[vr8hub]

Interesting. I didn’t get either of those here (I’m current from a release perspective). And you and Alex and I discussed at the time what this should look like, so it looks like what you and he suggested. :)

I don’t believe there’s anything in the manual for this, so let me know what I need to change and I’ll take care of it.

On Jul 27, 2020, at 3:46 PM, Robin Whittleton notifications@github.com wrote:

I‘m running lint from the master branch so maybe this is newer, but I’m getting a couple of errors about the dl:

s-061 │ Error │ chapter-3-1-19.xhtml │

element without block-level child.

and:

s-062 │ Error │ chapter-3-1-19.xhtml │

element without exactly one child.

@acabal https://github.com/acabal: got any input on the second?

[robinwhittleton]

Yep, I think we may need to make some linter updates.

[acabal]

That is not yet in the released version of lint. But, the manual outlines glossaries: https://standardebooks.org/manual/1.0.0/7-high-level-structural-patterns#7.11

<dt> needs a <dfn> child, and <dd> needs at least one block level child like <p>.

[vr8hub]

Right, but this isn’t being used for a glossary. You and Robin and I discussed this several weeks ago (it's in the War and Peace thread); we’re using the dd/dt objects to format an alpha-to-numeric code. I can put the p’s on the dd, but there are no dfn’s.

On May 23, 2020, at 11:20 AM, Alex Cabal alex@standardebooks.org wrote:

Yeah I think that would work.

<dl> <div> <dt>a</dt> <dd>1</dd> </div>

<div> <dt>b</dt> <dd>2</dd> </div> </dl>

dl div{ display: inline-block; text-align: center; }

The CSS ended up being quite a bit different, but the html is pretty much as Robin and you described:

<dl>
    <div>
       <dt>a</dt>
       <dd>1</dd>
   </div>
   <div>
       <dt>b</dt>
       <dd>2</dd>
   </div>
…
</dl>

[acabal]

OK. I forgot about that. I believe I updated the manual after the big glossaries discussion on the list, which was after we discussed this. I think we should still put <p> in <dd>, so that it's consistent; that's in the same spirit of requiring a block level child in <li>. But since we're not defining anything here, we can exclude <dfn>. I'll update lint to only look for <dfn> in actual semantic glossaries.

[acabal]

Adding <p> shouldn't require changing the CSS at all, I don't think.

[vr8hub]

Fixed. No, it didn't require changing the CSS.

This is a curious/education question. Why do we require a block level child on things like <li>, <dd>, <header>, that don't themselves require one? Is it just consistency for the times we do need one? Of the three, <li> makes sense, because e.g. endnotes are often more than one paragraph. But <header> and <dd> (almost?) never are.

[acabal]

Basically for consistency so we can always expect at least one child. This makes things easier to style and post process, and removes the decision of whether to use a child or not. <header> often has multiple children though I don't think there's a rule for that; maybe there should be one. <dd> is much more rare but having more than one line for a definition is a reasonable assumption and it does occur in the corpus.