The discovery solution needs to have the pointers. But the actual content doesn't. I'm advocating keeping discovery on dedicated discovery sites (which would of course be full of links), rather than mixing it in amongst the content.
The environment sets the context of the content. But your content may appear different when viewed from a different aspect (under a different context.)
A piece of content barely stands on it's own, that's the charm of the web. It's an endless task trying to atomise it.
A piece of content has to stand on its own (at least on a physical level; obviously works are embedded in their cultures), because reading is still a fundamentally linear activity (and watching a video more so). Displaying a composite piece of content derived from multiple sources is a rainbow people have been chasing since before the web (see xanadu), but it still looks as far away as ever.
I don't know why I said barely - should have said 'doesn't solely stand on it's own.' But of course ideally it would be nice if it had meaning all by itself.
How can a discovery solution not provide pointers to the thing you're discovering?