The problem is that you could share a link and it'd be seamlessly downloaded by a receiver of the link....you could even use them to handle images in web pages for example.
Now? The person who is receiving the link has to go to a web page, then click download. It's an unnecessary extra step and a downgrade in the user experience.
Even worse, it's now harder to keep track of what you have shared and what you don't.
From DB's perspective this is most likely to force additional eyeballs onto their product by forcing them to hit a webpage before download, but it's a lousy user experience for all involved. To the receiver of the link DB just looks like an old fashioned file-sharing site a la megaupload and not something special.
Both are direct links, and neither tell you much about the site that you're downloading the pictures from. The indirect link actually tells people that Dropbox exists as a service. Whether that's better for the end user or not isn't relevant to what I said.
Dropbox is much different from Megaupload. On Dropbox you get a direct link and you know whose directory you download from. On Megaupload the uploader is anonymous (to the downloader) and you do not get a direct link.
Now? The person who is receiving the link has to go to a web page, then click download. It's an unnecessary extra step and a downgrade in the user experience.
Even worse, it's now harder to keep track of what you have shared and what you don't.
From DB's perspective this is most likely to force additional eyeballs onto their product by forcing them to hit a webpage before download, but it's a lousy user experience for all involved. To the receiver of the link DB just looks like an old fashioned file-sharing site a la megaupload and not something special.