Yes. Exactly! I am trying prove what I think I see in the StealJS marketing paragraphs; that Steal is a client side loader that will allow rapid user interface prototyping in a development environment that leads to a simple transition to production builds.
An artists sketch starts with a light pencil drawing that helps determine shape and proportion. It can be quickly changed without too much rework. As you have suggested I could within a matter of seconds copy and paste the starter Bootstrap template into the html, add some custom components as proof of concept, and have a rough sketch. Right away however the issues of controlling asynchronous module availability and scope will need to be addressed to continue working with the sketch in the browser. My vision is that the sketch templates can essentially be serverless mostly functional examples. For example using rawgit dot com to serve an index.html of the template would be enough to demonstrate the look and feel. Obviously not a production solution.
Separation of Concerns is the key word I think I am after here.
It seems like the browsers do not care where the resources come from as long as the CORS requirements are adhered to. Basically load, wait and then run. The casual browser user interface observer has no need of all the development tooling details, such as linting, version control, precompilers, build tools, documentation etc. They just want to see and test the look and feel of the re-usable components. Some of the online IDE’s such as Plunkr, Codepen, and others like that sort-of achieve that ease of sharing, but dowloading a zip of the files leads to a complete rework to really start the project toward production. Starting with that kind of online IDE with production in mind is counterproductive (to be kind).
For example a developer wants to show off three pretty nearly functional user interface alternatives internally within the work group to three other reviewers, the example links powered by the StealJS loader would be provided and each developer/reviewer could click on the link within a couple of minutes each times three. Total demo time six minutes. At two dollars a minute that would cost $12 for the review.
Consider the powerful developer alternative npm install. Yes, you have to be a pretty well aqcuainted power developer to do this. You cannot take a quick look until you have achieved power developer status. All there? Great! (Not a developer, sorry you cannot participate…) Depending on your setup it could take a while to get everything downloaded. Okay now you have the full install on your laptop, type the magic build and run commands, copy the provided URL, open that up in a browser. Each super smart power developer experienced no installation errors and they are now able to proceed with the review. Let say a minimum of seven minutes to install and review each times three developers totaling twenty one minutes. At two dollars a minute that would cost $42 for the review. That only cost three and a half times as much as the above example to get a quick glimpse of the template. Now multiply that cost times hundreds or thousands of interested developers on an open source project. Now remember to delete all those files just downloaded, it was a temporary sketch after all.
Hopefully you see the cost saving perspective from these two narratives and why I am excited about the prospects of applying StealJS to create quick starting point templates.
I have assumed that to switch from the development template to production project the steal configuration object used client side can simply be copy and pasted into the server side package.json along with copying the config.json. Very slick (if it works)!
If I am delusional, please just tell me to stop thinking about doing it this way, I will bark up a different tree