Are we talking a link that would automatically cause the save-as menu to pop up, or a direct image link? The latter would probably be more helpful to more people, as it stands, and would still allow for easier downloading.
I figure there aren’t quick download links on the thumbs page since the ads are displayed along with the pictures. If there were, it would probably take all of a day for someone to cook up a batch downloader that would just fetch everything uploaded that day(filtered per your settings, even) without you having to visit the site proper. Just imagine if you suddenly went from people cherry-picking to downloading everything. The server HDDs would probably start bleeding while glowing images of eyes opened and closed all over the cabinet casing.
@Background Pony
Your concern is appreciated but the APIs available for interaction with the site are sufficient that someone who wanted to already could make an automated downloader that fetched whatever set of images you wanted without having to browse the site at all. The number of people who would actually do this would probably be pretty small and wouldn’t actually put much load on the server because serving static images and json is really light on the load - producing and serving html and generating/optimising image thumbnails is where the grunt work is. Bandwidth use might be a bit higher but that’s not a bottleneck for us at this time. In fact, people who solely used such a tool to interact with the site might actually put less load on the servers than otherwise, depending on how often it checked for new content. Clover may make statements that directly contradict information in this post because he knows more about this stuff than I do and I am not infallible so don’t accept anything I say about server matters as the gospel truth
Bandwidth is cheap for us. We’re on 10GigE pipes and plenty of monthly transit. So we’re fine with people downloading things. The main issue is actually congestion on that particular chunk of the UI - we can’t fit anything else there in the current design without removing something else. We’ll think about this in our next iteration, though.