I’ve been playing with ipfs and applications that use ipfs. I’ve noticed every application use ipfs daemon in a different way and doesn’t necessary interact with each other.
For example, IPFS-GUI starts ipfs daemon but doesn’t recognize if I launch daemon from command line. Arbore launches a daemon from an installed go-ipfs. Akasha embeds his own daemon (which conflict with already installed go-ipfs). While Chrome extension always detects if a daemon is running.
It’s not easy and confusing to test multiple apps… Should there be a consensus on own applications handle ipfs daemon ? (detect an already running daemon, check ports, propose to change path for a second daemon…)
I’m also interested in this question. I am currently developing a project with a cli client and would like to use or embed ipfs daemon. I’m thinking what is the best way to it and incentivize the users to keep the daemon running.
I am thinking about this too, IMO there are applications that should use their own store/daemon. I’ll ignore them here right now.
Currently I believe that a normal app should detect if a daemon is already running and let the user decide to opt-in to the existing daemon/store or launch a separate new one.
Akasha too: it comes with its own ipfs installation, and it doesn’t recognize the user’s default ipfs installation, let alone an already running ipfs daemon. You need to stop your running daemon, then launch Akasha, which then launches its own ipfs daemon, which then however writes to the default ipfs datastore path (afaik).