Claiming content

We love our Flattr button, but we also know the button itself is one of our biggest problems, because you need to add it in some way. Wouldn’t it be wonderful if Flattr could figure out that you own a piece of content and automatically let others flattr it, without you doing anything? Yes, awesome, but how to do it? We have banged our heads around this and here is the first part of our proposal.

While we could do our own solution that works for us and none else such a solution wouldn’t scale (and we hate closed things). All services needing to identify ownership of a webpage should be able to work on the same data – but which data should that be?

There has been a lot of work done in the discovery area, spearheaded by Yahoo’s Eran Hammer-Lahav, resulting in a mechanism based on links and relations on top of the Web Linking specification. In its registry we can find an “author” relation defined by HTML5 which “refers to the context’s author”. Assuming that authors are also owners we can resolve any such link from a webpage to find which Flattr user to assign a thing to.

How to add such a link? Not through HTML anchor-tags – those doesn’t represent the page in itself. Links would preferably be added as part of a site’s host-meta file, specified in another of the discovery specifications – Web Host Metadata. If the site can’t add a host-meta file then links could instead be added as HTTP headers or HTML link-tags in a webpage’s HTML head section.

What should a author link point to? Something that “refers to the context’s author” and which Flattr can confirm belongs to a Flattr user. It could be something we know, like the URI to a Flattr profile or the e-mail used to register a Flattr account, or to a webpage which has a confirmed path of XFN me-relations between it and a Flattr profile. The latter is something that eg. Google indexes and exposes in their Social Graph API and which they also have tools to inspect.

Since it’s probably not always the authors who should receive the money from flatters this single relation doesn’t completely solve the claiming of content. The owner of a web page should be able to say that someone else but the author should receive money. For this some new relation types would need to be defined as a next step.

We haven’t yet implemented any of this – not even the me-relations. This is just a proposal on which we want feedback. We think it’s important to be open and especially so when trying to work with open standards which are pointless if noone else agrees on them. So – what do you think? Is this how content should be claimed?

7 thoughts on “Claiming content

  1. o dont really understand what all this mean can you explain a littel more in a easy why but if its cool im in :P

  2. I see what you mean with the relationship between website/page not always being the right one for Flattr to use; there are indeed a number of cases where that wouldn’t be the right thing, mostly related to user-generated content.

    Wouldn’t it be fine then to have a “flattr:user” property on specific content elements, to be referred to with a #id-suffixed URI? Just search for the at the top if there is nothing in the specific element. Of course this would make it a bit more complex to “canonicalise” an URL.

  3. Nice idea indeed.
    Concerning standardisation, are you aware of other platforms/service providers using semantic tags to define the owner of some page/element?

  4. @mattias: I will probably do more detailed posts in the coming months – will perhaps easier to understand then. They will still be rather technical though.

    @Diego: Some solution for what you’re referring to would be nice – but let’s first solve how to claim full URL:s :)

    @nicoulas: I’m surprised that no other services are doing this. The closest I can think of is the authors defined in RSS/Atom, but that isn’t very usable for us. Some systems, like WordPress, uses the microformat variant of Atom – hAtom – to mark up content – we might be able to use that as an additional way to discover authorship, it’s not as explicit though.

    @Andreas: Using RDFa in what way? I think one should be able to discover an author without relying on parsing HTML. That way you can easily discover the author of all kinds of content. RDFa and Microformats could perhaps be good as additional discovery mechanisms though, but we’re of course open to suggestions if you have anything specific regarding RDFa in mind :)

Comments are closed.