Hacker News new | past | comments | ask | show | jobs | submit login

Neat concept, but it wouldn't work for creating modern GUIs, where half the controls are invisible until moused over (desktop), or hidden behind hamburger menus and swipe gestures (mobile).

Not only that, all those magical solutions are doing is solving the easy part of development, which is laying down the widget-based layout. The actual hard parts (different behavior based on context such as viewport size/orientation, actual UI business logic) are an afterthought.

Impressive work, yet misguided in many ways in my opinion.

It is a cool concept. However, what it doesn't seem to cover, and what would fix the issue you are speaking about, is coding for a set of images. Given a set of wireframes, with expected post effect and interactivity hints interleaved, I would expect this generator to build the code that covers all the specs, balancing the spec needs and being able to hold the stack in memory until all the known elements are accounted for before writing the final code.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact