W3C home > Mailing lists > Public > whatwg@whatwg.org > September 2011

[whatwg] <comment> element

From: Shaun Moss <shaun@astromultimedia.com>
Date: Mon, 05 Sep 2011 18:55:02 +1000
Message-ID: <4E648E66.5010909@astromultimedia.com>


On 2011-09-05 6:36 PM, Odin wrote:
> On Sun, Sep 4, 2011 at 10:43 PM, Shaun Moss<shaun at astromultimedia.com>  wrote:
>> Yes, but this is not semantic!!! Comments are not articles. They are
>> completely different. Comments can appear in reference to things that are
>> not articles (such as status updates), and therefore would not appear inside
>> an<article>  tag - so how would the browser recognise them as comments?
> It is semantic.
>
> Comments *are* in fact articles. You're thinking of it in the wrong
> way. Article is not a newspaper article, but something that would make
> sense to stand on its own.

Please explain to me how it makes sense for a comment to stand on its own.

To an HTML author, especially a newbie, an article *is* a newspaper 
article, and this is entirely distinct from a user-submitted comment 
related to the article. Semantics isn't just for robots, it's for 
humans, too - a fact that seems to be frequently overlooked. Giving 
elements obscure, unobvious meanings in the spec is a kludge, plain and 
simple. For example, to the WHATWG and W3C the <b> tag now basically 
means "different but not important". The <i>, <u> and <s> tags have 
similarly gained bizarre new definitions. To everyone else on the 
planet, however, <b> means bold, <i> means italic, <u> means underline 
and <s> means strikethrough. This may come as a surprise, but 99.9% of 
HTML authors don't read specs. They look at a tag, and think, now what 
would I use that for? "Ok, so the <table> tag is for tables, right? I 
guess <article> is for articles, then. Oh, it's for user-submitted 
comments, too? <confusion>wtf?</confusion>"

>
> So, a *nested* article is defined to be dependent on the outer
> article, but still it is it's own content and can be syndicated as a
> individual content piece that's related to the parent article.
>
> It makes perfect sense and is quite beautiful and doesn't require a
> whole slew of tags. It's very nicely done.

We don't need a whole slew of new tags. We need *one*: <comment>, and 
/maybe/ one more: <comments>, to wrap them.
I'm not sure of the reluctance to introduce new tags that are useful and 
have a clear purpose. There are 10 tags just to describe tables. I'm not 
saying we should do that. I'm saying: here is a very obvious use case 
for an element that (imho) is not being met by the current spec.

>
> And comments /are/ syndicated. Just look at WordPress. When I read
> blogs in Liferea, I get the blog posts, as well as each individual
> comment loaded from the syndicated comment-stream from that particular
> blog post.
>
>
> <article>
>    <h1>HTML5 is great</h1>
>    <p>Yup. It is.</p>
>    <footer><p>By Me</p></footer>
>
>    <article>
>      <p>You're so correct!</p>
>      <footer><p>By Ben</p></footer>
>    </article>
>
>    <article>
>      <p>Better than butter, I say</p>
>      <footer><p>By Adam</p></footer>
>    </article>
>
> </article>
>
>
>
> Perfect is the enemy of good. Cue in xhtml2. :-)

This is a false analogy. HTML5 is better than XHTML2, because it reduces 
errors and is more in line with browser behaviour. If XHTML2 was perfect 
it wouldn't have been canned. I'm not talking about anything like that. 
I'm proposing a new tag to represent an extremely common feature of 
millions of websites.

Shaun
Received on Monday, 5 September 2011 01:55:02 UTC

This archive was generated by hypermail 2.4.0 : Wednesday, 22 January 2020 16:59:36 UTC