Contrary to popular belief blockchains can make external web calls

[jellul]

It is widely accepted that blockchain systems (including Bitcoin) cannot make external calls to web services if required. This is widely believed because each node needs to execute computation that is deterministic in nature - this requirement however became an assumption that this meant that every node in a network needs to execute the exact same computation.  This is false.  Every node needs to reach a deterministic state, regardless of how they reached that deterministic state.

We believe this to be a breakthrough and in our recent paper have shown that Blockchain systems (and potentially Bitcoin) could be implemented to make external calls to web services themselves (without requiring oracle inputs or external systems).  The paper can be downloaded here (it's being submitted to be published):
https://www.researchgate.net/publication/351497108_Towards_External_Calls_for_Blockchain_and_Distributed_Ledger_Technology

Get in touch if you would like to know more!

Joshua Ellul

[1714022823]

1714022823

[1714022823]

1714022823

[1714022823]

1714022823

[1714022823]

1714022823

[1714022823]

1714022823

[jackg]

Doesn't this affect the decentralised nature if you're making calls to Web 2.0 interfaces?

Admittedly the fees are high at the moment but the hope is that all information can be stored in a decentralised blockchain (where necessary). If private data needs to be put on there it can also fairly easily be obfuscated (I think legal companies have done this in the past to use each other's computation without giving away personal info).

[jellul]

No affect on decentralisation.  Instead of waiting for oracle input, this mechanism allows the blockchain to go get the input itself.

[jellul]

Also, here's a link to a more digestible brief of the work: https://joshua-ellul.medium.com/contrary-to-widespread-misconception-blockchain-dlts-and-smart-contracts-can-make-calls-to-94d864415ca7

[dkbit98]

We believe this to be a breakthrough and in our recent paper have shown that Blockchain systems (and potentially Bitcoin) could be implemented to make external calls to web services themselves (without requiring oracle inputs or external systems).

I don't understand how this could have any real use case for Bitcoin blockchain so maybe you can elaborate more on this and give me some examples, but it can certainly destroy some centralized altcoin projects.
Maybe some scripts could be used for this but I think that fees would be very high and you would still need to trust some trusted third party, and who says that someone is really trusted?

[gmaxwell]

Signed party provides a signature (or other verifiable data) is a model that has been used for years in Bitcoin and elsewhere, it's not novel.

No one ever suggested that you couldn't do that.  It just isn't terribly valuable because your security still rests on the third party-- and if they were really that trustworthy why not eschew the blockchain and just have them process the transaction? There can be reasons, but it's those reasons that justify using such approaches...

[jellul]

I don't understand how this could have any real use case for Bitcoin blockchain so maybe you can elaborate more on this and give me some examples, but it can certainly destroy some centralized altcoin projects.
Yeah, it would be quite wacky to see this in bitcoin. But one option could be that it could automatically reach out to your systems or a trusted custodian automatically to say release a payment.

Maybe some scripts could be used for this but I think that fees would be very high and you would still need to trust some trusted third party, and who says that someone is really trusted?
Trusted third parties are dependent on the use-case. If you need one (or a group), this is just a way to reach out to them.  If not, then it's not needed.

Signed party provides a signature (or other verifiable data) is a model that has been used for years in Bitcoin and elsewhere, it's not novel.

No one ever suggested that you couldn't do that.  It just isn't terribly valuable because your security still rests on the third party-- and if they were really that trustworthy why not eschew the blockchain and just have them process the transaction? There can be reasons, but it's those reasons that justify using such approaches...
We are not proposing that that is novel.  What is novel is showing that if needed bitcoin or any other blockchain could make a call to an external web service if it required. It's a common misconception that blockchain systems cannot actively make calls to external web services.