Inbox has the right idea, in that the protocol and API set it has devised are open source (GNU Affero GPL licensed), and the project is designed to appeal most directly to developers of email applications building on mobile platforms. A similar project both in its approach and its design is JMAP, a protocol proposed by FastMail. JMAP uses JSON to encompass and package all the possible requests and responses used for email: sending and receiving, calendaring, contacts, and so on.
Building a better inbox
Given how tough it is to rip and replace email at the protocol level, small wonder many have concentrated instead on fixing the client experience. After all, most of the headaches users experience with email revolve less around the protocols and more around managing email so that it doesn't turn into a job unto itself.
The current plans in this vein go well beyond a more elegant-looking client or one better suited to mobile use. Instead, they use statistical analysis to automatically classify and act on email -- to figure out with as little user intervention as possible which emails can be dumped, which can be circled back to later, and which need to be replied to right now. "The goal of email 2.0," said Dave Bagget, CEO of Inky Mail, "should be to make email clients more like personal assistants than mere tools for sending, receiving, and organizing email."
Google has been hard at work on this approach. Google has carried Gmail's autocategorization system to a further extreme with its new mail product called Inbox. With Inbox, mail is batched together automatically in "bundles," according to certain criteria. Emails that Inbox believes are most important, such as updates on product purchases or travel arrangements, are emphasized, as are to-do items and reminders.
IBM, too, has a similar project in the works, one meant to extend on its existing user base for Lotus Notes. Verse ditches Lotus Notes' heavy native client for a lightweight Web-based option, focuses on people and conversations rather than on individual email messages, and uses IBM's Watson machine learning service to help classify messages. What's more, IBM has pitched its early-access program for Verse at users, presumably to see whether it carries over from there into the enterprise.
The problem with algorithm-driven inboxes: They need to be at least as good as -- if not better than -- human-powered curation. They also won't gain much uptake with business users if they don't provide functionality taken for granted in those circles. (As of this writing, Google's Inbox is still missing a few such items, including signatures, shortcuts, and advanced filters.) Any such inbox needs to provide users with a fallback to an uncurated view of their messages, or people will begin to feel like their email isn't really their email anymore.
Sign up for Computerworld eNewsletters.