They validate everything according to the consensus rules. To reiterate Satoshi (emphasis mine):
The receiver of a payment does a template match on the script. Currently, receivers only accept two templates: direct payment and bitcoin address. Future versions can add templates for more transaction types and nodes running that version or higher will be able to receive them. All versions of nodes in the network can verify and process any new transactions into blocks, even though they may not know how to read them.
are you even reading your own words..
old nodes do not validate signatures of segwit.
the quote you made about satoshi is where he is saying he envisions a time in the future where instead of downloading a whole new client. users can download a 'patch' (template) that makes old nodes recognise new transaction types.
but here is the thing.. core didnt insert that template feature into their code to allow users to patch core without needing a full upgrade. so when segwit activates, old nodes dont see a segwit transaction as segwit and have a template to be able to validate signatures of segwit.. what old nodes see is an "anyone can spend" which is the trick where its telling old nodes
there is no signature, but blindly pass it on anyway.
i dare you to show me where a users of version 0.3-0.12 can simply patch their old version to not treat segwit as a "anyone can spend" and instead actually sees and validates signatures.
as for your other waffle, about the test net stuff.. what may work in a secondary environment may not work in bitcoin.
hard fork implementations have actual code running on bitcoin, relaying data to other bitcoin implementations. segwit however has not had this finalised code included in a BITCOIN node untill a few weeks ago. and before that the code on the testnets was completely different.
you mentioned segwit has been around for months.. so show the finalised code 3 weeks ago. then show the same code, from any other time you have pretended that segwit was part of bitcoin.. go on
show me code from 2 months ago where they match up.
show me code from 4-6 months ago where they match up.
screw it.. ill do it for you.. blockweight was not even a variable in bitcoin or testnet or segnet's consensus.h until only a few weeks ago
as for the stuff about bip109. read it. maxwell was making blind assumptions,
but later admitted he didnt know what code the nodes were running or why. meaning he had no clue,(he even says it was just a hypoethis)
What are you testing? What system are you attempting to validate by making Bitcoin Classic unable to use testnet and making BIP 109 perpetually unusable there?
Were you testing if people would notice that a BIP 109 signaling miner that was not enforcing the BIP 109 rules would go unnoticed until is significantly broke a network? If so-- I think you've confirmed that hypothesis.
and knowing he lost the argument because he didnt know how or why, he then started to talk about rogers investment amounts as a subtle way to drop the subject without having to actually say he was wrong by making assumptions.
anyway you have meandered these posts off topic to avoid the question.
using the
CONSENSUS mechanism where there is no intentional split(no blacklisting) and because the remaining 5% gets rejected/orphaned to keep the chain going in a single direction the majority is agreeing to.
again emphasis using the same consensus mechanism that takes care of bitcoin every single day.
again not meandering down your false doomsday of intentional splits
where is the problem?
remember people running full nodes want to remain full nodes, they dont want to be blinded. so ofcourse the minority that didnt upgrade will upgrade, to avoid seeing orphans and avoid other issues, and remain full nodes.
full nodes want to be full nodes for good practical reasons of decentralization, security to protect bitcoins immutability, etc etc
if they wanted to be a blind node they might aswell run multibit
there would be no logical reason for a full node to choose to run as a blind node where they are not validating all the data they receive
there would be no logical reason for a full node to choose to run as a blind node where they are wasting bandwidth on orphans
there would be no logical reason for a full node to choose to run as a blind node where they they cant trust 0-1 confirms due to orphans
again reply in context of a consensus.. not your dooms day intentional split controversial rhetoric