If I need 80 GiB+ on disk (and considerable CPU effort and RAM) to be able to retry one or at most a hand full of 16 KiB transactions, that does not seem to be the best software engineering solution. As said, light wallet apps could and should just cache transactions (don’t know to what extent they already do, there was no real congestion after the SundaeSwap incident that I noticed).
(By the way, I would love a stand-alone light wallet app instead of those browser-based thingies.)
Using that node inside Daedalus for other purposes is a very valid use case. I have said that above.
One of my earlier posts (Start Daedalus with tmux to use cardano-cli) was even about how to use the cardano-cli
coming with Daedalus. (Since the shell inside that nix-chroot
is very limited, I’m now using an additional cardano-cli
outside instead and just point $CARDANO_NODE_SOCKET_PATH
to the Daedalus node.)
Using it to fill a cardano-db-sync
is also a very valid use case.
I’ll probably switch to a standalone node without the Daedalus UI, which I’m not really using anyway, at some point.
All I’m saying is that just using it as a wallet app seems at least unnecessary to me. And the impression that all these resources somehow benefit others, benefit the network as a whole is just wrong.
For a large percentage of these users, it is important to know that it is not soo much safer to use a full node wallet, that they don’t need to invest all these resources, that other wallet apps than Daedalus (and Yoroi) are totally viable. I can’t count the “How am I supposed to use Cardano? Daedalus now barely starts!” posts in all the channels.
That one I can totally subscribe and I think I’m doing quite a lot for that. For “informed”, it is important to explain the workings of Cardano over and over with different approaches. For “welcomed”, it is important to try to find the right level for each and every user. And I’m really trying that.
I just don’t think that for “involved”, it is really valuable to run a behemoth of a full node wallet app if you don’t really know and need what is in there.
… and thus, we might have to agree to disagree on that one.
(BTW: What do you mean with “full node Nami”? Running own cardano-submit-api
and pointing Nami to it? Then “full node Eternl” should also be possible and we can maybe meet there.)