I think setting up a public HAF server is problematic, because of the issues around providing write access to schemas in your HAF database (at the very least, you'd need to figure out how to set disk storage quotas for users of the database).
I think the simplest thing to do is setup a regular public testnet, then have each dev group create their own HAF server that is getting data from the public testnet. This would eliminate all the resource allocation issues. Storage requirements for these HAF servers should be small, because the testnet won't have a lot of blocks (relatively speaking).
It might be interesting to provide a docker image for such a HAF server that is already configured to connect up to a public testnet. But probably all that's really needed is a proper config.ini file, then devs could use the config.ini file + the install_haf_server.sh script to setup their server.
Just to clarify the above: each dev group would run a hived server connected to the testnet that would then feed their local HAF database.
Good points. I think that's a good way forward. I'll write a script to automate the steps and a new config.ini to start off.