Genome Biology has a (non-free, for some reason) article/rant by Gregory Patsko about supplementary data attached to papers:
I always thought that the most important thing in any scientific paper was supposed to be the data and how they were obtained. Everything else is window-dressing, because it's filtered through the lens of subjectivity. The background, the discussion - these are somebody's opinions. If the experiments have been done carefully and analyzed thoroughly, the data are the only facts in the paper, the only thing that can be trusted. They're what I want to read and understand. The people who obtained the data have the right to tell me what they think it all means, and I often find their opinions useful, but I also have the right to decide for myself. Yes, I can still do that if I dig out the supplementary material, but I shouldn't have to dig. If our priorities are straight, the methods and the data should be the centerpiece. And in the modern era, there's no reason not to put them there.Genome Biology | Let's get our priorities straight
In the example page I've been setting up—based on OJS—as much of the data as possible is embedded directly in the article, so that when the page is saved, the media should be saved with it. The 'supplementary data' will be linked to directly from the page as well, ideally with a useful rel attribute so that it can be automatically fetched along with the rest of the page. This should make things much easier, rather than having to go to a separate page and click links to see each microscopy movie, or open and scroll around tiny pop-up windows for every table or figure.