Last year I set up a Cardano testnet as part of the Plutus Pioneers program. I'm now using my Daedalus installation to experiment with the cardano-cli, which I am finding much easier, with the added benefit of knowing that my transactions actually work in the real world. The transaction cost so far is trivial. Is there any reason not to use the mainnet for R&D, learning, and experimentation? I can't imagine I'm stressing the network, and I'm putting only my own ADA at risk (little 2-5 ADA transactions). I can imagine I could clutter up the blockchain with pointless metadata.
Closely related, is there a way to use Daedalus to talk to the testnet?
I just found a very good answer for why not to use the mainnet for learning and experimentation. Fortunately I did not learn this thru experience but by reading about other's difficulties.
If ADA is sent to a script, and no datum/hash is sent, or if the order of parameters is incorrect, the ADA will be locked forever in the script (assuming the script does not also have a fail-safe endpoint built in to return funds after a time if no other endpoint is triggered).
Perhaps this problem has already been fixed as it was discussed a year ago, but I'm sure the risk remains. Fortunately I am not at the point yet of experimenting with scripts.