name not allowed on a li element

[daniel-montalvo]

Ref: https://validator.w3.org/nu/?showsource=yes&doc=https%3A%2F%2Fw3c.github.io%2Faccname%2F

The HTML checker complains that there is a name attribute on a li element. Can we use another method for this that passes the HTML validator?

[pkra]

They don't appear to be used for anything.

Looking at 9c71070, it seems like they were left in place by @cookiecrook during #186. Maybe to facilitate review or perhaps for historical value? In the latter case, we could add a data-historical- prefix.

[daniel-montalvo]

Thanks for this detailed explanation, @pkra

If @cookiecrook does not object, I'm happy to take this on. I am taking an overall pass at most of the ARIA modules to make them ready for transition to next maturity level.

[daniel-montalvo]

According to @jnurthen this was left for old links to keep working. One can do https://github.io/accname/#step1 and be taken to the "Initialization" step under the "Text Equivalent Computation" section.

[pkra]

According to @jnurthen this was left for old links to keep working. One can do https://github.io/accname/#step1 and be taken to the "Initialization" step under the "Text Equivalent Computation" section.

It doesn't though -- only for a tags. Putting the names on some anchors in the list items should work though.

[cookiecrook]

Yes. Ideally we'll need to leave non-linked anchors so that older incoming links don't break:

<li id="newNamedStepID">
  <a name="oldNumberedStepName">Labelledby:</a>
  etc etc
</li>

[cookiecrook]

Cool URIs don't change (or break)
https://www.w3.org/Provider/Style/URI

[cookiecrook]

@w3c/aria-editors Should there be a GitHub hook to include validation as a PR check?

[daniel-montalvo]

Not sure what we would check here in addition to what specprod already checks.

If you have a look at https://github.com/w3c/accname/actions/runs/6706842952 that is checking links and markup, including the HTML validator, which now fails on <li name="">

[cookiecrook]

Ah. I didn't know it was already validating.

Now I'm wondering why that didn't get flagged on my original PR #186.

[daniel-montalvo]

@cookiecrook We have a much lightweight set of linters for the PRs. I realized the other day because I wanted to make sure we could publish on TR, for winch markup validation and link checkers are compulsory.

[jnurthen]

We can't use <a name> as respec will try to resolve these as references

[cookiecrook]

We can't use <a name> as respec will try to resolve these as references

Do you have another resolution in mind? I suppose we could drop the old permalinks.

[cookiecrook]

Oh I know. We can move the old numbered name to an id on the <em> that is the first child of each li

<li id="newNamedStepID">
  <em id="oldNumberedStepName">Labelledby:</em>
  etc etc
</li>

The <em> should already be there... Just move the attr to the firstChild and change the attr name from name to id.

[jnurthen]

I created a PR with empty spans with IDs - see above. Seems to validate fine and works

[jnurthen]

Oh I know. We can move the old numbered name to an id on the \<em\> that is the first child of each li

<li id="newNamedStepID">
  <em id="oldNumberedStepName">Labelledby:</em>
  etc etc

The \<em\> should already be there... Just move the attr to the firstChild and change the attr name from name to id.

Not every line with the name attributes has an em element to add it to. Decided adding a span and being consistent was more important

[cookiecrook]

okay... let's continue in the PR