Hacker News new | comments | show | ask | jobs | submit login

Posted on SO:

Meteor has data validation already, though not user authentication.

The model is that the server chooses what data to expose for reading, and the server also performs the real writes against the database. The client only simulates the mutations. You can define your own mutations ("methods"), and have them simulated on the client or not. If access checks pass on the client, they still might fail on the server. The server can run additional or different code, and you can hide this code from the client by putting it in the "server" directory. If the method results in different side-effects on the server than on the client, the differences are synced down to the client -- exactly what you'd want.

However, in new projects, to make things easier for first-time Meteor developers, we auto-publish all server data. To turn this off, type "meteor remove autopublish". That is, auto-publishing is the default for new projects but not the default for meteor overall.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact