A partial archive of discourse.wicg.io as of Saturday February 24, 2024.

Proposal: Comments element


Many sites now contain user-generated content that is added to the primary content of a page, usually called “comments.”

Many users want control over whether they see this content; it is often viewed as distracting, distasteful, and in some cases, damaging. “Ignore the comments” has become a shorthand for acknowledging that allowing people to add their thoughts to more well-researched content often results in an overall lowering of quality, as well as a waste of time for users.

The Web platform should give users control over whether they see content that they might perceive as low-value. A new HTML <comments> element would allow sites to mark up this kind of content so that browsers and other user agents could automatically remove it, if the user expresses a desire to do so.

This is better than site-by-site preferences, which are onerous.

Initial implementation in browsers could be done in extensions. The exact semantic of <comments> would be need to be carefully described, so that it was clear that if the primary content of a site were user-generated content, it would not apply (but comments on that content might be in-scope).


If a web site doesn’t care about your user choice, they simply never use this element. I don’t see where adding yet-another-element that does nothing except provide some mythical context for UA’s to work against is the solution here.

Along this line, the full initial experiment can be done in user-land. Make a custom element for this mythical “comments” functionality. Make an extension to look for it and do whatever action is needed on web pages. Then go promote that sites start using it to work out the finer details. Once some data is had over how this should work and if there is a true positive user benefit, then things could move forward in terms of standardizing it if desired by the web community.

All the tools are available today to go collect data to help prove your theory. That’s the best path forward rather than throwing an idea into spec-land and seeing what happens.


I’m not sure @mnot really intended to imply that you should throw this immediately into “let’s skip all that” - he even brought it here and suggested that initial implementations could be extensions, for example… Maybe what he is looking for is people interested in this as a problem to help thoughtfully come up with a thing that would ultimately potentially yield such a proposal in the first place? idk, I probably shouldn’t speak for Mark, he can speak for himself, but at least I didn’t read it that way.

I wonder if @danbri (https://twitter.com/danbri) has comments about this because if there is any kind of substantial thought/cowpathing it is almost certainly through schema.org – I have some rough http archive reports containing what appear to be non-standard elements (dasherized tags), though they are flawed, there are thousands of them and 0 of them contain any hint that they are about comments or discussions, and quickly searching things like AMP I dont see that sort of thing there either so it’s kind of hard to think of where you begin past schema.org.


Discussion also continuing here: https://github.com/schemaorg/suggestions-questions-brainstorming/issues/223