I'm really excited to see the Web Annotation specs stabilize, and see implementations happening all over the place. I really hope it succeeds in becoming a technical recommendation of the W3C. It's not so much that I have particular hopes for strong tool interoperability as much as the re-usability that a common vocabulary brings to projects and modules. I have hope that some of the ES6 modules I've written for the HTML environment will actually find use in EPUB readers and such.
As far as user experience and products go, I am actually really excited to see the Clipboard APIs advancing in the browser, and I hope Chrome and IE implement custom MIME types for the clipboard soon. I think cross-frame communication is one of the hardest parts of browser extensions and third party site add-ons, but for interactive reading and writing we have this really wonderful bus (cross-application, even) that has been underutilized for decades.
I'm also inspired by full-text search like what worldbrain.io is trying to do.
A mash-up of these ideas is something I'd like to see. Rather than ambitiously indexing all that a user sees, keep track only of highlighted or copied snippets and offer to auto-complete or paste them with reference in editors. We're all familiar with auto-completed @ mentions and # hashtags, but why not quotes? The storage and privacy concerns are greatly reduced this way and it becomes tenable even in a mobile environment, I think.
Anyway, that's enough rambling. Part of what excites me about Coral is that it's tackling community problems really directly and those problems were often derailing when I was trying to focus on reading/writing workflow. Specialization, composability, and mash-ups with the right level of platform support have much more promise to me.