Odd "best practise"

[kholbekj]

"With a JAMstack project, anyone should be able to do a git clone , install any needed dependencies with a standard procedure (like npm install), and be ready to run the full project locally. No databases to clone, no complex installs. This reduces contributor friction, and also simplifies staging and testing workflows."

What?

Well you'd have to install whatever runs the API, including any DB? This doesn't sound like a consequence of the "JAM stack", but as a side effect of connecting your local environment to an external API, hence not having the whole project.

[verythorough]

This doesn't sound like a consequence of the "JAM stack", but as a side effect of connecting your local environment to an external API, hence not having the whole project.

The idea here is that when you use a JAMstack approach, you no longer have to have the whole project running locally. This could be because you're connecting locally to an external API, or it could be that you're working with files that are stored directly in the repo, rather than in a DB.

Contrast this with a frontend developer making changes to a WordPress theme or Rails app. In both of those cases, even after installing and running a local LAMP or Rails server, the developer can't see their changes on actual content without either creating a dummy DB with sample content, or cloning the actual DB. This step is unnecessary with a JAMstack approach.

Of course, the frontend is only half the story. If a JAMstack project connects to a backend API, then a backend developer contributing to that side of things would of course need to do a local install of whatever powers that API. However, the backend code is in a completely different project from the frontend, and the install again is simpler.

The key is that the component parts are decoupled, so contributors to one part don't need to install all the parts for local testing and development.

[ghost]

I use a BaaS approach with Serverless and check my API in with my front-end code, FYI

[verythorough]

Oops, I didn't mean to close this! I do a lot of work in Intercom, and in that, my default action is to "close and comment." Sometimes muscle memory kicks in over here! 😬

[verythorough]

@jhabdas In your approach, I would say that AWS Lambda and the Serverless framework would be the separate "backend" projects in that case.

[ghost]

I suppose that depends on how one defines separate. Here's an example Serverless service checked in alongside a Jekyll repo. Tomato tomato. 😜