@hongminhee wow, this is very cool, and inspiring.

I'll be watching and may have to borrow this idea.

You really think of everything

@dansup@mastodon.social That's very kind of you, thank you!

We're excited about it too. And “borrowing” is highly encouraged! We'll be happy if our idea can be useful to other projects.

@hongminhee Sounds ambitious but worthwhile!

You might have some overlap with https://github.com/Letterbook/Sandcastles by @jenniferplusplus, @feditest, or https://github.com/berkmancenter/activitypub-fuzzer/ by @darius. Hopefully there are some facets that are already solved or at least lessons to incorporate.

@julian@fietkau.social Wow, thank you so much for these links! This is exactly the kind of prior art and community effort we need to be aware of.

I'll definitely be studying Sandcastles, ​@feditest@mastodon.social, and ActivityPub Fuzzer. It's super helpful to see what's already been done. We're all for incorporating existing lessons (or even tools) rather than reinventing everything.

Really appreciate you pointing these out!

@julian @hongminhee @jenniferplusplus @feditest I'd love for any of these projects to use the data I've collected at the Observation btw

https://observatory.cyber.harvard.edu/snapshots/

@darius@friend.camp @julian@fietkau.social This is fantastic, thank you! A dataset of “in-the-wild” objects is the perfect complement to the testing plan.

I can see this being hugely useful for creating our test cases, ensuring we handle all the variations and quirks that different servers actually produce.

Bookmarked! We really appreciate you sharing this.

@hongminhee funfedi.dev already runs some compatibility tests on Hollo

https://funfedi.dev/support_tables/

@hongminhee By the way, according to this test Hollo can't process an activity if @context is not set:

https://funfedi.dev/support_tables/context/

I would appreciate if you fix that.