The Real Deal About rel="me"

Earlier today, Barry Schwartz published a piece over on Search Engine Roundtable called Google Does Not Use rel=me Microformats. The article got shared around Twitter a lot, and those tweets surfaced on the microformats IRC channel. In it he mentions a few things that I would like to respond to, that is probably best done in this format.

For context: the article quotes a tweet where John Mueller writes he is unaware of Google using rel="me" anywhere.

Authorship Irrelevance

Schwartz writes about the demise of Google surfacing individual authors. I am sad about that too as it was a great way to show how individuals make up the web, even amongst those big multi-user blogs. But he seems to miss out on the fact that rel="me" was never used for authorship. At least I have not been able to find anything that says otherwise.

Google (rightly) used rel="author" to figure out who authored a specific article. This use of the author link relationship is the one documented in both the HTML standard (§4.6.6.2) and on the microformats wiki (/rel-author).

This very page uses it. And there are several tools that will extract this information. The IndieWeb even has an “authorship discovery algorithm” worked out that also incorporates it.

So What Is rel="me" About?

The me link relationship can be seen as saying: the URL I am pointing to is about the same person as the page I am on. Schwartz’ Twitter profile (@rustybrick) links to his RustyBrick profile page using this link relationship, making it clear that those two pages are about the same person. And that is what rel="me" is for. The microformats wiki (/rel-me) calls this “profile equivalence”.

When we know that different web pages are all expressing parts of the same person, we can start to draw them out as a graph. The RustyBrick profile page also includes links to Schwartz’ YouTube, LinkedIn, and Facebook. If those links all had rel="me" on them, the graph would show us that Twitter user @rustybrick is the same person as Facebook user @barryschwartz. Without a specified link relationship there is no way to know this as the profile page may just have been linking to a random Facebook page.

We can even put this graph to use. Lets highlight 2 use-cases:

  1. Kevin Marks has played with Distributed Verification. The idea here is that we can be pretty sure who owns rustybrick.com and that a profile page on there really means something. But anyone can walk up to Twitter and register themselves as “Barry Schwartz”. We see fake Twitter profiles all the time. But only the Schwartz in control of rustybrick.com can put a link with rel="me" on there.

    If the Twitter account I am looking at links to the RustyBrick profile page, and the RustyBrick profile page links back to the same Twitter account, each using rel="me", I know the Schwartz I am looking at on Twitter is the same person! By using the same link relationship at two ends we have verification.

  2. Building upon the distributed verification we can let people identify (“authenticate”) themselves as well. That is the idea behind RelMeAuth.

    There is way more useful information about Schwartz on the RustyBrick profile page than on his Twitter account. But if we want to know that profile belongs to the person we are talking to, they need some sort of proof of ownership. Because Twitter can be used to login to external parties, we can use that as a way of showing ownership. The person capable of logging in to @rustybrick on Twitter can be said to also own the RustyBrick profile page.

There might not be a solid use-case for Google to work with this link relationship right now. Especially not for surfacing pages in reply to a search query. But the rest of the web can make great use of it. This brings me to the last thing I want to say:

This Markup Is Not For The Sake Of Google

I was very happy to see that Search Engine Roundtable is still including rel="author"! Because of course they should include it even if Google isn’t using it for Rich Snippets (or whatever).

Sir Tim, best known for tweeting during the 2012 Summer Olympic Games opening (and inventing the web), wrote a note on Linked Data where he highlights the following question:

If someone has the URI of that thing, what relationships to what other objects is it useful to know about?

If you are reading this page, it might be useful to know who its author is.

The fact that “Google’s algorithm” isn’t “considering” rel="me" does not change the answer to the above question. It also does not affect how we can use it.