I have worked with domino for more than 15 years. Let's be honest. Domino is a dead platform for new users and a dying platform for existing users. Domino had been a pioneer in technology a long time ago and has been left behind since then. Due to the owner's hesitation and bad decisions, it has disappointed developers and users a million times. It's really sad to see the last hope is diminishing after HCL committed to continue domino's life. In my humble opinion, most ideas here would have no impact or even accelerate the final death of domino, so would be HCL's most efforts. Below are the points.
Enhance the database dramatically. If that couldn't be done, create a new one.
Domino = notes db + client/server + RAD + infrastructures e.g. directory, mail. Database is the base of the ecosystem. As one of the first document database, notes db can hardly compete with other decent document databases like mongodb.
1.1 Small data (versus big data).
1.2 Inflexible data models, e.g. Can not embed documents like Json, field size limits.
1.3 Can not scale up.
1.4 Partly due to the default behavior of the client, e.g. ctrl + s to save, save conflicts can not be properly managed.
1.5 Convenient but inefficient security model. Access to an application and to a database are not separated. When an user open a view, no matter in the client or in code, domino has to check each document to see if the user has access to it. It's ridiculously slow. The normal practice is to control the application access with custom code and query the database with an uniform account. In that way, a single database query can find all the documents satisfying the user's access level. The built-in ACL mechanism gives no room to build a real role based access system either.
2. Drop the client development using forms and views.
IBM has always been hesitated about whether to drop it. Notes client/server applications are precursors of web applications rather than the peers of other client/serve applications. The user interface and codes are downloaded from the server and run in the client. But html + css + js surpass notes counterparts greatly in various aspects. Focus on web develop in the future or domino would fail in both fields.
3. Embrace the request-based web frameworks.
There are 2 types of web frameworks: request-based and component-based. The former are thriving (e.g. struts, spring, flask, django) and latter depressed (e.g. jsf, asp.net without mvc) because it's inflexible and can hardly follow the rapid progress of web technology. Request based frameworks involve writing user interface with templates. But contrary to many people's minds, with all the technological flexibility and automatic tools, the development speed is not lower than using the existing components, especially taking into account of testing and maintaining.
Xpages is just a jsf implementation with a notes db backend. Check the java world, how many people are using jsf and how many are using spring? Something like node.js + notes db is a correct choice. With proper API and toolsets, pure coding is also RAD. So why people would use notes db as a backend, see the first point.