Questions about SegWit and Compatibility with Old Bitcoin Nodes

Hey folks, First time posting here. I've been looking into SegWit and the whole Ordinals prunability issue. Got a couple questions. So, I heard Witness data can be pruned. But what happens when a new node tries to verify pruned transactions? Wouldn't it just reject the block since it doesn't follow SegWit rules?

1. The block header still has enough info to check if the block is valid, even without witness data. 2. Older nodes see those transactions as spendable but not invalid. They stick to the longest valid chain rule, so they won't propagate invalid blocks, keeping the network secure.

But how does it work in practice? Like, if I can't verify transactions without witness data, how can I trust that everything's legit? I get that buried blocks mean multiple verifications have happened, but if witness data is missing, I can’t check signatures, right?

SegWit is probably one of the most poorly explained updates ever. For a Full Node to do its job, it needs complete transaction data. You can’t just strip witness data and still validate transactions. Full Nodes keep the witness info intact.

Right, but pre-SegWit nodes will see SegWit transactions as invalid. The way the transaction flag is set up messes with the input count for older nodes. They interpret those transactions differently and miss some fields.

Non-standard? Sure. Invalid? Nope. It gets converted into a non-SegWit transaction. Check out this example link. Old nodes can still reference this transaction ID, and it stays the same.

Thanks, guys. Things are clearer now. I get that stripping witness data makes validation impossible for SegWit transactions. If you mess with that data, it breaks the whole consensus. I'm diving into Bitcoin to learn about Layer 2 services like BRC20 and Stamps, and I saw some comments about stripping transaction info.

Stripping happens mainly with SegWit transactions. A node removes those extra fields so the transaction hash matches what an old node would see. There's also WTXID, which is the hash including SegWit fields, and a separate Merkle tree for miners.

The transaction data and witness data separation allows for validation even with pruned data. As long as the existing data meets the requirements, it can still be valid. Consensus rules check things like block structure and signatures but not witness data.

In Bitcoin, the blockchain is immutable. Anything in it can't change without making it invalid. And pre-SegWit nodes don’t relay those transactions by default since they’re non-standard. It’s about input count, not just the amount being transferred.

I think you’re right about pruning the Witness structure, but it doesn’t mess with block verification. New nodes can still check a block's validity using txids and the Merkle tree. SegWit doesn’t get rid of witness data, it just separates it.

SPV clients don’t fully verify blocks. They just check if transactions exist in the block header using the Merkle tree. Calling it separate is misleading; witness data is part of the transaction.

Thanks for clearing that up! "Segregated" sounds better than just "separate". Most of what I said was from what I absorbed when I was new. I think I read about 'getwitnesstx' in some older Bitcoin Core version. Anyway, I’ll look for those sources.