api.Element.name entry doesn't make sense as used on MDN

[foolip]

https://developer.mozilla.org/en-US/docs/Web/API/Element/name says:

name gets or sets the name property of an element in the DOM. It only applies to the following elements: <a>, <applet>, <button>, <form>, <frame>, <iframe>, <img>, <input>, <map>, <meta>, <object>, <param>, <select>, and <textarea>.

Note: The name property doesn't exist for other elements; unlike tagName and nodeName, it is not a property of the Node, Element or HTMLElement interfaces.

https://developer.mozilla.org/en-US/docs/Web/API/Element/name#Browser_compatibility however is not a useful table for this page, because BCD correctly reflects that there's no name attribute on Element in any browser.

@ddbeck how should one handle a case like this? Should the MDN page instead link to the tables for the elements that do support it?

[ddbeck]

Yeah, this is super weird. There isn't an Element.name attribute, but there is an HTMLFormElement.name and HTMLImageElement.name and so on.

And looking at the history of the page, @sideshowbarker's comment on the most recent edit is probably right:

The Element interface actually has no 'name' attribute (this entire article should be deleted…)

My hunch is that we should make a brief check to make sure there was no historic Element.name (e.g., that some previous DOM spec doesn't contain that attribute).

If it did exist, then we could slap a deprecation notice on the page and refer people to the list of individual HTML*Element.name pages (example).

If not, then we could archive the page and drop the Element.name data from BCD.

[saschanaz]

DOM1/2/3 had no Element.name definition:

• https://www.w3.org/TR/DOM-Level-3-Core/core.html#ID-745549614
• https://www.w3.org/TR/DOM-Level-2-Core/core.html#ID-745549614
• https://www.w3.org/TR/REC-DOM-Level-1/level-one-core.html#ID-745549614

[ddbeck]

Thanks for checking that, @saschanaz.

My earlier suggestion was slightly incomplete. Here are the steps to fix this:

• Archive Element.name page
• Drop the feature api.Element.name as irrelevant
• Add features for the real api.HTML*Element.name attributes

[foolip]

Thanks @ddbeck, that sounds great! Will you check back here when the page has been archived? I can try to do points 2 and 3 from your list.

[ddbeck]

@foolip Yes, I'll update this issue as soon as I've archived the page. 👍

[ddbeck]

@foolip OK, on the advice of @chrisdavidmills, I redirected the page to https://developer.mozilla.org/en-US/docs/Web/API (instead of archiving or deleting). We're now good to drop api.Element.name from BCD

[foolip]

I've sent #6751 to remove the api.Element.name as irrelevant.

Here are the 17 HTML element interfaces that have name attributes in the HTML spec's IDL:

• HTMLAnchorElement
• HTMLButtonElement
• HTMLEmbedElement
• HTMLFieldSetElement
• HTMLFormElement
• HTMLFrameElement
• HTMLIFrameElement
• HTMLImageElement
• HTMLInputElement
• HTMLMapElement
• HTMLMetaElement
• HTMLObjectElement
• HTMLOutputElement
• HTMLParamElement
• HTMLSelectElement
• HTMLSlotElement
• HTMLTextAreaElement

The name entry exists on most but not all of these, and there are other consistency problems to work through, like whether they should be marked deprecated or not.

[queengooborg]

It looks like we have name property BCD entries for all 17 of these APIs! I guess all that's left is to ensure the statuses for these features are consistent, correct?

[foolip]

@queengooborg that sounds like a good idea! The entries on HTMLAnchorElement, HTMLEmbedElement, HTMLFrameElement and HTMLImageElement are now set as deprecated, but the rest aren't. I doubt that's right.

[foolip]

@queengooborg #14481 only fixed the standards_track status, not the deprecated statuses.

[queengooborg]

Which ones are still inconsistent? @ddbeck reviewed the PR and mentioned that four of the APIs are deprecated so the subfeature should also be deprecated.

[foolip]

@queengooborg it's deprecated on HTMLAnchorElement and HTMLImageElement even though the parent features aren't.

[queengooborg]

Daniel had noticed that the name property was put into the "obsolete" section of the spec for these two interfaces, see #14481 (comment) and #14481 (comment).

[foolip]

OK, no mistake then. Sorry :)