Cryptocurrency: Blockchains As Global Code Repositories?

grarpamp grarpamp at gmail.com
Thu Jan 27 17:12:32 PST 2022


On 1/27/22, Undiscussed Horrific Abuse, One Victim & Survivor of
<gmkarl at gmail.com> wrote:
> what kind of chain do you imagine backing it?
> on-chain

Any kind of structure for which msgdata (patches) can be sent
to addresses (repos), and that maintainer address can then
sign and send a msgdata tx (branch update, release) out
to itself that contains a list of accepted patches (selected
from earlier txn's that were sent to that address).

BSV, ETH... anything that could store a repo's worth
of ongoing patch tx's since a repo (address) was init
(generated and published as being a project)... could be
used for this right now today, if the tx cost were free.

Even if not free, crowdfund to that address, hand the
funds out to developers to use to pay tx cost, fund
devs addresses directly, etc.

Code branch/fork side/subtoken chains, whatever...
in the end, like onchain, they're just bitfields and part
of global available cpu/disk/network.

> would the data be ipfs

Content-addressible requires you to be passing
hashes around with some other tool, it's an
unnecessary layer when a chain can just hold
all the patches themselves directly as tx's.
Perhaps then use the txhash as index referencible content hash,
patches would want to include them to make it easier and
automatic to figure out what they were based on and apply to.
Decentralized probably involves more work there figuring out and
deciding/marking what the tip of HEAD looks like than does git.

Maybe you're not going to be using some git to keep rereading
even a fully indexed chain directly (inefficient), but are
maybe just scraping patch sequences that were sent
to an address out of the chain, through maintainer/committer
acceptance filters, into a simple local repo, then you can
use a hacked git to emit your commits as patches sending
them back out to the repo address on the chain.

It's not meant to be featureful, but it could turn out that way
because you can define the API, bitfields, etc.

However it may solve how to do a globally distributed repo,
detached realtime redundant cryptographic uncorruptible
consensus etc. None of today's "dvcs (Git)" or even
"databases (Oracle)" seem to offer that.

Closest historical simplistic analog would be
stitching together split files posted to NNTP.

This one has its own ticker symbol: [G]CODE
People can invest in it, work remotely, earn dividends, etc.


More information about the cypherpunks mailing list