Introducing the Graphcool Framework
blog.graph.cool
blog.graph.cool
Also on Product Hunt today: https://www.producthunt.com/posts/graphcool-framework
Question: How do we sync schema changes between dev environment and production? As of right now, this just means we can see the well-designed Graphcool backend and don't have to worry about getting locked into a cloud offering, but it needs schema migrations to use in dev environments.
We couldn't be more excited to take this big step forward by open-sourcing the Graphcool Framework. This means you can now:
* Run Graphcool locally (also works offline)
* Deploy a Graphcool cluster to your own servers/cloud
* Get involved in the development of the framework on GitHub
We can't wait to see what you will be building and look forward to your feedback! (Here is a 5 min demo video: https://www.youtube.com/watch?v=xmri5pNR9-Y)
I'm not particularly experienced with Docker, however from what I've read online the consensus seems to be avoid containerizing your database [0]. Was this considered when making the decision to include the DB within your container?
[0] https://myopsblog.wordpress.com/2017/02/06/why-databases-is-...
Go ahead and read and decide for yourself though: https://news.ycombinator.com/item?id=13582757
Out of curiosity, why did you decide to build your own graph database, rather than use Postgres or some other mature solution?
Edit: Oops, I didn't realize it uses MySQL under the hood. This section of the readme made me think you had a custom db implementation: https://github.com/graphcool/framework#graphql-database
With that said, GraphQL is a misnomer anyway. It's not really about graphs. It allows you to have a query syntax that lets you easily query related and nested objects but it doesn't have the concepts of directed edges and nodes like you have in a graph unless you explicitly code it.
[1] unless they added a feature that I was not aware of and couldn't find in the 30 seconds I spent googling
With this route, we just have to wait for the framework to be developed and evolved by the awesome community (which supported the initial graph.cool before)
If I only control the server, and my customers control the client, I hate GraphQL. It's too powerful.
I wonder how the migration between both of these services would be
I like prototyping apps with Vue and Firebase, but considering anyone with the firebase key could mess with the database, it doesn't work for real apps.
It seems to be harder to retrofit a new format onto an existing framework, than to write a new framework altogether. Once the new framework is out, it will be replaced by another shiny framework that offers a brand new feature that is actually a rehash of something old.
I'm optimistic that there will be something next that will replace framework churn, but what?