exposure-notifications-server VS upspin

Not a small one, but a real-life one, the exposure-notifications-server is usually what I like to recommend, keep in mind that you probably need a bit of Go knowledge and context before looking at it so perhaps it isn't the best introductory project to look at; however I recommend you to bookmark it for future reference.

Here, I can even link you to an article of how exposure notifications work: https://www.google.com/covid19/exposurenotifications/

Google's Exposure Notification Reference Serve is a real life example of what you're asking for.

Which is just a massive misunderstanding of how the contact tracing API works, even if you do believe that it was "surreptitiously installed ... on users [sic] phones".

https://www.google.com/covid19/exposurenotifications/ (“The Exposure Notifications System was introduced in an update to Google Play Services. This update did not automatically turn on Exposure Notifications or enable any Exposure Notifications functionality. Exposure Notifications only works if you decide to opt-in.”) Android 6.0 and above.

Source: https://www.google.com/covid19/exposurenotifications/

Super intriguing. Thanks for sharing!

It reminds me a bit of an early Go project called Upspin [1]. And also a bit of Solid [2]. Did you get any inspiration from them?

What excites me about your project is that you're addressing the elephant in the room when it comes to data sovereignty (~nobody wants to self-host a personal database but their personal devices aren't publicly accessible) in an elegant way.

By storing the data on my personal device and (presumably?) paying for a managed relay (and maybe an encrypted backup), I can keep my data in my physical possession, but I won't have to host anything on my own. Is that the idea?

https://upspin.io/

There are a few Go projects meant to be learned from:

- https://github.com/pion/opus for to learn audio

- https://github.com/benbjohnson/wtf for overall production quality

- https://github.com/upspin/upspin difficult to explain, personally I'm not a fan of the errors

You could also take a look at some real-world open-source projects. I like upspin for its idiomatic approach.

For example, Rob Pike's upspin places all its validations in the separate package. Do you agree with that approach? Which yet proven options there are?

Just a few projects that could perhaps interest you in terms of design of your own solution :

Upspin: https://upspin.io/

The early error wrapping work which emerged out of the Upspin project, that eventually made its way into the errors package, included stack traces in the wrap error. This would provide exactly what it appears you seek.