I think that new bookmark service should be build around context(as mentioned by @codegeek) of URL, its lifecycle and its content. It should leverage the infor

I think that new bookmark service should be build around context(as mentioned by @codegeek) of URL, its lifecycle and its content. It should leverage the information on the page and not only the user input. I'm convinced that the most important information in bookmark like services are the URLs saved by users and the content of the pages they want to bookmark. It's not just about building dummy database with links, references and "static" features on top of it. I think that really good bookmark service should be self-taught based on your usage of the system, your likes and your information from social networks.