Question

I’m trying to use the <a> tag in HTML5 more as a container as this tag can now have block elements as children, example:

before (valid XHTML 1.1)

<div>
    <h3>
      <a href="page.html" title="article title">article title</a>
    </h3>
    <p>
      text
    </p>
    <a href="page.html" title="article title" >
      <img alt="image">
    </a>
    <a href="page.html" title="article title" >
      read more
    </a>
</div>

after (valid HTML5)

<a href="page.html" title="article title" >
    <h3>
      article title
    </h3>
    <p>
      text
    </p>
    <img alt="image">
    <div>
      read more
    </div>
</a>

Does this new way of markup have any effects for SEO?

Answer 1

OK, removing pure semantics from your question (which, in my mind, does have a material impact on deciding on implementing your chosen method) and concentrating on pure “SEO” value and impact:

The first example needs to be qualified more, as if we take your example as literal, then you are linking to the same page.html 3 times. Google (specifically) only takes the link anchor value from the 1st link to any page that it comes across, so – the value for the first example is only extracted from that first link. The 2nd link (using an IMG tag with an ALT attribute as the anchor value), and the 3rd link using read more as the anchor value are effectively “ignored”. It’s important that other signals are used to supplement the first link’s true intended value, such as surrounding text, images etc.

The 2nd example (HTML5), wraps all of that semantic/surrounding content up to make the effective ‘anchor’ value from which search engines will derive the link’s intended meaning, and then as a consequence, the meaning of the destination page of the link.

Using an anchor tag as a containing wrapper for content that contains additional emphasis (the H tag), an image and an additional div only increases the difficulty that a search engine has to decipher the intended meaning of the link so it can associate it with the destination page.

Search engines (and Google predominantly) are constantly improving their crawling ability to enable better algorithmic parsing and processing of the HTML. Apart from emphasis signals (which are very low), Google mostly ignores the mark-up. The exception is of course links – so making an effort to simplify the parsing/processing by providing clear signals as to a link’s anchor text is the safest way forward. Expecting them to understand all of the differences of HTML3, vs HTML4, vs HTML5 and all of the transitional, strict and other variations of each, is probably expecting too much.

TL;DR
Possibly, but only in terms of true link value.