I love that feature from RStudio. What’s missing with shinyapps.io is an easy and inexpensive way to host it on your domain!
I was a heavy Shiny user and I created one of the first popular Shiny extensions in shinysky. This is my 2 cents, we definitely a low-level layer that doesn’t assume any framework or UI components. Shiny was based on jQuery and Bootstrap, but that made extending shiny much harder. On the other hand, we need high level contructs that wraps some library to get started. So if shiny was broken up into a low-level Shiny that you can use any framework with and a ShinyBootstrapjQuery then it would be fine. Julia shouldn’t make the same mistakes.
Thanks for the feedback. It’s something that I spent a lot of time thinking about lately.
Architecture wise it seems to me that the best approach would be web components. It’s a mature technology that now has been embraced by all browser vendors - and it doesn’t lock the user into any framework. A web component is a complex UI element which encapsulates all the functionality and which looks like a regular HTML element: a tag with attributes. This would allow users to a) use any of the existing components; b) build and distribute their own; c) write framework agnostic code which can be coupled with any frontend framework.
Then there’s, of course, the issue of coupling these, auto-magically, with the Julia backend, for 2-way communication, managing state, etc.
Re shinyapps.io - yes, 300/mo for your own domain is a bit expensive. I guess their reasoning is that you only need that once your app is ready to make money.
Web components sound good but a bit vague to me. In the case of Polymer, is your idea now still version 2, or rather version 3, or even LitElement and Material Web Components, or something different than Polymer?
At the moment the project lives mostly in my head. I’m working on design and architecture plus early experiments. The plan is to get a working prototype by June to present at JuliaCon.
@Diego_Javier_Zea Nothing to show yet, I’m afraid. ATM the focus is on getting Genie to v1 (which should happen in a few weeks) to have a stable foundation. Once that’s done, the focus will be on applications. There are a few projects in the pipeline, including “Shiny for Julia”, GenieCloud (for deploying web apps), and a static website generator.
I’m New to Julia, Compiler Design is well planned and Advanced. last one month i have search Package in Hybrid-App for high-performance numerical analysis and computational science. We Need,
Best Interaction of Input from HTML
Interact Widget & WebIO it’s fantastic, but missing filepicker and not fully customizable
Knockout used for well Data-binder
Easy to link with Core
Julia
Best Visualization
Interact with PlotlyJS is really better performance. but we need more output display other than graphs
Support High-Level MVC
Genie.jl is a high level Web-framework but missing Interact like works.
When we Combine these all goodness. Really We will develop --“Full stack Research/Academic/Industrial App Development with Julia”
