jashkenas
commented
12 years ago Looks like you didn't get any takers for comments.
After looking over the code a bit more, I do think it's best to leave Stickit as a nice plugin for now -- there's still more agnosticism we'd like to have in Backbone core. The basic concerns are the same as before:
- Two way binding makes sense for basic form elements, but a lot of modern UI elements are more sophisticated JavaScript widgets.
- Form serialization to JSON isn't hard.
- Some folks want inputs to be two-way bound on
keyup. - Some folks want inputs to be two-way bound on
change. - Some folks want inputs to be two-way bound when you click
savefor the form.
delambo
I just extracted and opened up a project, Stickit, which is a highly configurable model binding plugin for Backbone. My goal for Stickit was/is to conform to Backbone's style and to fit naturally within Backbone's core functionality. You can read more at the following github:
http://nytimes.github.com/backbone.stickit/
I respect the Backbone community's (and jeremy's) worldview which prizes minimalism and uniformity: if a high threshold of users won't use a feature or if a feature feels out of place, then they won't consider it. Model binding is a popular feature, but it has been hard to envision where it would fit in with Backbone's hands-off approach to rendering. Where other libraries/frameworks and plugins have failed by moving model binding configuration into the DOM/presentation layer, Stickit was designed to keep configuration in the place that it should be - in the JavaScript and in the context of the view controller which encapsulates the element and model.
I was thinking about drawing up a pull for Backbone which included a model binding implementation like Stickit, but I would like to get opinions and ideas from the community first.
Thoughts?