[dasmoth]

I don’t know for sure, but I certainly hope we’ll see some fresh thinking about user interface design and construction. The past couple of decades seem to have been substantially about recapitualating what came before in the web browser, and while webification has it’s good sides (easier deployment), the actual interfaces for data-entry type tasks still seem as clunky as ever.

AR is potentially an interesting sub-field, but doesn’t seem to be the answer for everything (e.g. those form-like data entry tools...)

[TheOtherHobbes]

I think UI progress is unlikely without good AI, and good AI has to be much better than human to be passable.

(If you're not convinced, try watching how often you have to ask your fellow humans what they meant by a communication and/or a request for information. It's probably more often than you expect - but you give fellow humans a pass because you're used to it, and so are they.)

Either that, or personal data has to stored in a central server so it can be accessed on demand by web apps - which would eliminate a lot of web forms, but would have uncomfortable political and social implications.

There's still room to improve form-based pages, because there's still far too little research into best practice. But forms are an efficient way to collect information, so it's hard to imagine a secure and private UI paradigm that would eliminate them altogether.

[dasmoth]

Either that, or personal data has to stored in a central server so it can be accessed on demand by web apps - which would eliminate a lot of web forms, but would have uncomfortable political and social implications.

That doesn't necessarily require centralisation. Web browsers have some form-filling capabilities now, and that data can stay under end-user control. Perhaps there's scope for building on something like this (although the growth of, _e.g._ "social login" doesn't leave me too optimistic. That perhaps does count as an example of UI innovation, although one which hasn't particularly registered with me since I tend to avoid it).

There's still room to improve form-based pages, because there's still far too little research into best practice. But forms are an efficient way to collect information, so it's hard to imagine a secure and private UI paradigm that would eliminate them altogether.

Agreed. I don't see easy wins, but trying to make forms as good as they can be seems a very worthwhile area of endeavour. I suspect part of this might be trying not to go too far in terms of baking "business rule" type stuff into forms, which has a tendency to leave people in impossible states (thinking, for instance, of academic grant systems which can end up with some very strong assumptions about career paths built in)