Are non-Segwit nodes, full nodes?

[Wind_FURY]

No poll. I believe this would be a very good topic wherein everyone can discuss, learn, and express their own opinions.

Are non-Segwit nodes, full nodes? Why or why not?

[Herbert2020]

they still are considered full node but i believe they shouldn't be. because one of the things a full node should do is to store full transaction/block data and be able to give it to you when you request it. and as far as i know non-segwit nodes do not store witness part of transactions which means although they still function but they can't give you "full" data so maybe we shouldn't consider them full nodes. they also don't relay SegWit transactions.
but as i said that is only one of the things full nodes do, they do a lot of other things which the non-segwit nodes are still capable of.

[DooMAD]

I'd argue they are absolutely full nodes.  Every user relaying and verifying transactions is helping the network.  Each user has a choice if they want to utilise SegWit or not.  The best part about SegWit is that it was implemented as an opt-in softfork.  So, having given users that choice, I don't think it's healthy for the community for us to start defining some nodes as a kind of secondary underclass just because they've opted not to support SegWit.  We've had enough in-fighting already.

[franky1]

legacy only nodes do not:
1. receive unconfirmed segwit transactions
2. relay unconfirmed segwit transactions
3. verify unconfirmed segwit transactions
4. verify confirmed segwit transactions in blocks
5. receive true full length blockdata
6. store true full length blockdata
7. relay to other nodes true full length data

if a legacy node sent a stripped legacy block to a segwit node(for syncing purposes). the block would get rejected as its stripped and not got the data segwit nodes need

so legacy nodes are not part of the:
1. full validation
2. full relay
3. full sync
4. full archival

if you did not know this. then please research:
stripped blocks aka filtered block data
downstream nodes aka filtered/bridged nodes
if you think all blockdata is sent and everything gets the same thing. research
--witness
which is a parameter addition to the blockdata RPC call that asks for full data. which legacy nodes do not get to call

the bitcoin devs were very clear about this

legacy only nodes are treated as downstream(separate layer) nodes that receive filtered data
https://bitcoincore.org/en/2016/10/27/segwit-upgrade-guide/#not-upgrading-1

In this configuration, you set your current Bitcoin Core node (which we’ll call the “older node”) to connect exclusively to a node running Bitcoin Core 0.13.1 or later (which we’ll call the “newer node”). The newer node is connected to the Bitcoin P2P network as usual. Because the newer node knows about the segwit changes to the consensus rules, it won’t relay invalid blocks or transactions to the older node—but it will relay everything else.

When using this configuration, please note that the older node, if it uses Bitcoin Core defaults, will not see transactions using segwit features until those transactions are included in a block.

If you don’t upgrade, you may experience one difference: if someone who has upgraded to segwit pays you, your wallet may not show you the payment until after it has been included in a block. This is a safety feature that prevents your wallet from seeing transactions it doesn’t completely understand until they’ve been confirmed by a miner.

they even draw pictures for the non-technical minded to understand
legacy only nodes do not fully validate blocks. as the devs have said they rly on segwit nodes to validate fully and then for the segwit nodes to send out data the segwit nodes believe to be true. and so the legacy nodes then simply accept them as true without validating signatures of segwit transactions in blocks. because they are not supplied with signatures that they can understand to verify.

this was explained many many times.
heres one example, explaining it even before segwit was activated that the compatibility was not actually as compatible as proposed.. i even draw pictures..(yea i tried to help ELI-5 it for those that are not technical)
https://bitcointalk.org/index.php?topic=1760149.msg17580661#msg17580661


(p.s before triggering a rage reply. take time to notice the source of the information(the link is from the devs themselves). understand the information. and reply without insult and only with content about the topic. do not meander it into personal attacks at me to avoid learning the truth from the devs themselves)
ooh and do not reply just to reply about this grey paragraph.

[franky1]

So IMO, non-SegWit node is full nodes, but can't run all non-critical tasks.

its not a full node. because as you say it doesnt do all the critical tasks
but its what some call a 'heavy node' or 'old node' or a few others (who dont understand the meaning) a 'compatible node'
or as the majority prefer. to save confusion.. 'legacy nodes'

as its not a lite node that is just has a wallet.dat and a transaction creator ability, but not a full node either

[squatter]

No poll. I believe this would be a very good topic wherein everyone can discuss, learn, and express their own opinions.

Are non-Segwit nodes, full nodes? Why or why not?

Yes. They validate transactions and blocks against the consensus rules, as always. They process Segwit transactions because they're fully compatible with the protocol. They just don't directly participate in them.

That's why the "legacy nodes don't fully verify" argument is weak. If you run a legacy node, you don't directly participate in Segwit transactions -- you aren't accepting payments where you don't verify signatures.

[franky1]

(facepalm)
not fully verifying..
dont directly participate

full nodes doesnt mean just verify transactions aimed at you. it means fully validate all data to ensure NETWORK security.
come on. imagine if all nodes didnt fully verify.
the whole point of full nodes is that the whole block should be verified.

this debate has been discussed. if full nodes should only verify transactions that only relate to a payment recipient then the payment recipient should only request blocks containing transactions that are addressed to the recipiient.
short answer to that debate was that full nodes should validate everything. not just its own personal needs but validate the networks data.

That's why the "legacy nodes don't fully verify" argument is weak. If you run a legacy node, you don't directly participate in Segwit transactions -- you aren't accepting payments where you don't verify signatures.

legacy nodes are on a sublayer and also not part of the full network relay
sending out a stripped block to a full node will get rejected. you can only send a stripped block to a fellow downstream (legacy) node.

         0
       /   \
      o     o
   /    \ /    \
0  -o- o -o-  0
   \   /  \    /
     o     o
       \   /
         0

the actual topology of the network if you were to lay it out in a straightish line would be (left is top(upstream) of network right is bottom(downstream))

                                                              
pools->fibre ring network->segwit nodes  - > segwit lite wallets
                                                             \-> legacy(old) nodes -> legacy lite wallets

blue and red are segwit nodes part of the p2p network(pool and fullnode). then on the outside are the green legacy nodes. they receive stripped block data but they dont relay full data. hense they end up as leachers rather than seeders (using torrent terms)
the red and blue work at the centre and full data sync of p2p using the '--witness' parameter of the rpc call for block data..
which legacy nodes are not privileged to. so legacy sit downstream (outside edge) of the network

(let me guess instead of a open consensus debate. certain group will now say that by not upgrading to their code your a network leacher..(did i accidentality give them the next REKT buzzword for next future REKT campaign) as their excuse to force people to adopt new code their not comfortable with activating.. (although it would still end up getting activated via their tricks anyway but i still see future REKT campaigns)

[Zin-Zang]

No poll. I believe this would be a very good topic wherein everyone can discuss, learn, and express their own opinions.

Are non-Segwit nodes, full nodes? Why or why not?

Yes. They validate transactions and blocks against the consensus rules, as always. They process Segwit transactions because they're fully compatible with the protocol. They just don't directly participate in them.

That's why the "legacy nodes don't fully verify" argument is weak. If you run a legacy node, you don't directly participate in Segwit transactions -- you aren't accepting payments where you don't verify signatures.

 
LOL, you are one of those confused people.

A True Bitcoin Full Node has a complete copy of the blockchain, and relays all Transactions included per the last code update.

A non-segwit node is basically a Node that never bothered to upgrade , it used to be a full node , but it lost that status with the segwit update.
At best now called a legacy node, since it contains a copy of the legacy bitcoin and only relay original pre-segwit transactions data.

Nodes that are not Full Nodes are Pruned Nodes / Lite Nodes / non-segwit nodes.
Now you have many foolish ones in here that claim they are.
So anyone claiming one is Ask them Directly why the Bitcoin Network Can Not Survive if all of the Nodes are only Pruned / Lite / non-mining segwit.
*Bitcoin-Segwit Network Can only survive if their are Full Nodes in Place, and those Full Nodes allow the Pruned / Lite / Non-mining segwit nodes to pretend to be valuable members of the network, those other nodes do not allow the full transactions versions or relay the complete chain , therefore they are inconsequential in the bitcoin network.
Just a PR Tactic to make people running Pruned / Lite / non-segwit nodes believe their support is vital ,
when in fact they are entirely irrelevant to the bitcoin network.      

Disclaimer:
If all of the Nodes revert back to the Legacy non-mining nodes , then you basically have a fork that removes segwit from bitcoin and then and only then would those legacy nodes revert back to being true full nodes. But only with the death of segwit, which by the way will kill Lightning network.

FYI:
The Contrast is quite amazing, on 1 hand Bitcoin Network Energy Waste is hailed as a great thing, but on the other hand pruned/lite/non-mining segwit nodes are used because god forbid, everyone be required to run a full node which is a necessity of a running network, let's pretend these sad pathetic imitations actually served a vital function, when the truth is anyone running one of those nodes is just too cheap to maintain a Full Node.
Maybe it is because the core devs are too cheap to devise a way to pay those running a full node they want to trick them into wasting their money running those useless nodes that won't even work without Full Nodes.  

update:
Wind Fury can't handle the truth.  

[franky1]

Disclaimer:
If all of the Nodes revert back to the Legacy non-mining nodes , then you basically have a fork that removes segwit from bitcoin and then and only then would those legacy nodes revert back to being true full nodes. But only with the death of segwit, which by the way will kill Lightning network.

not exactly. (pools would need to stop collating segwit transactions)
i say not exactly because you excused mining pools from your downgrade. meaning you want pools to allow segwit

a better example:

if there was a segwit bug. requiring segwit deactivation.. the network would end up having to:
orphan off all the blocks back to mid august 2017 (very worse case(research anyonecanspend debate))
or
continue at current blockheight. but revert to legacy block additions.. but then people with funds in segwit addresses wont be able to spend their segwit funds, if segwit got deactivated.
thus funds end up lost to the true recipient because legacy nodes wont 'playball' with them due to no signature.

yea i know what you are thinking. in such a bug the reversion would mean old nodes see segwit UTXO as "anyonecanspend" but not anyone will... trust me. the pools will grab the funds not average joe legacy node

but yes .. other side effect.. bye bye to btc access to the SEPARATE NETWORK that many coins can use called LN

[Wind_FURY]

Stop it. You were already replied to, https://bitcointalk.org/index.php?topic=5061961.msg47803087#msg47803087

Continue your debate there.

Disclaimer:
If all of the Nodes revert back to the Legacy non-mining nodes , then you basically have a fork that removes segwit from bitcoin and then and only then would those legacy nodes revert back to being true full nodes. But only with the death of segwit, which by the way will kill Lightning network.

not exactly. (pools would need to stop collating segwit transactions)
i say not exactly because you excused mining pools from your downgrade. meaning you want pools to allow segwit

a better example:

if there was a segwit bug. requiring segwit deactivation.. the network would end up having to:
orphan off all the blocks back to mid august 2017 (very worse case(research anyonecanspend debate))
or
continue at current blockheight. but revert to legacy block additions.. but then people with funds in segwit addresses wont be able to spend their segwit funds, if segwit got deactivated.
thus funds end up lost to the true recipient because legacy nodes wont 'playball' with them due to no signature.

What would be a more sane path to take though? Patch the bug or Segwit deactivation? Would a deactivation have consensus?

yea i know what you are thinking. in such a bug the reversion would mean old nodes see segwit UTXO as "anyonecanspend" but not anyone will... trust me. the pools will grab the funds not average joe legacy node

but yes .. other side effect.. bye bye to btc access to the SEPARATE NETWORK that many coins can use called LN

In such an outcome, which one of the split would the community follow, the Core developers support, and the market value?

[Pursuer]

From other discussion thread, that depends on your definition of full-node

sometimes definitions are not opinion based! there is a clear definition and it can not be changed. and that definition for a full node according to bitcoin.org is this:
"A full node is a program that fully validates transactions and blocks."
an old node can not "fully" validate transaction and blocks so it should not be called a full node. although since we have no other word for these types of nodes we have no choice but call them full nodes. we may call them "old full nodes" or something like "limited full node".

[DooMAD]

sometimes definitions are not opinion based! there is a clear definition and it can not be changed. and that definition for a full node according to bitcoin.org is this:
"A full node is a program that fully validates transactions and blocks."
an old node can not "fully" validate transaction and blocks so it should not be called a full node. although since we have no other word for these types of nodes we have no choice but call them full nodes. we may call them "old full nodes" or something like "limited full node".

But is that not simply a case of updating the technology and not updating the terminology along with it?  If someone invents a flying car, do we suddenly have to stop calling the conventional ones "cars" because the flying cars can utilise extra space that the conventional ones can't?  

[Pursuer]

sometimes definitions are not opinion based! there is a clear definition and it can not be changed. and that definition for a full node according to bitcoin.org is this:
"A full node is a program that fully validates transactions and blocks."
an old node can not "fully" validate transaction and blocks so it should not be called a full node. although since we have no other word for these types of nodes we have no choice but call them full nodes. we may call them "old full nodes" or something like "limited full node".

But is that not simply a case of updating the technology and not updating the terminology along with it?  If someone invents a flying car, do we suddenly have to stop calling the conventional ones "cars" because the flying cars can utilise extra space that the conventional ones can't?   

although this is not a good example but if we also remove roads so that the conventional cars could no longer be used in transportation then you can't call them cars any more. they become obsolete.

[NeuroticFish]

although this is not a good example but if we also remove roads so that the conventional cars could no longer be used in transportation then you can't call them cars any more. they become obsolete.

Just the old nodes are not completely useless/obsolete, so maybe we should rename them into legacy nodes or something like that, since the "new" full nodes will remain the only full nodes.
Legacy nodes' name should not contain the word full, they are not full nodes because, as said, they "can not fully validate transaction and blocks".

[DooMAD]

My stance remains there are different varieties of full node and that we just need to update our vocabulary a little.  We've always been flexible with this.  I don't see why that should suddenly change now.

A full node can either be selfish or networked.
A full node can either be pruned or archival.
A full node can either place limits on their bandwidth or not.
A full node can either be SegWit-enabled or not.

They'll all reject a block as invalid if the coinbase reward contained 100BTC.
They'll all reject a block as invalid if it was 100mb in size.
They'll all reject a block as invalid if it contained a double spend.
They'll all reject a block as invalid if the block header isn't formatted correctly.

Each help in their own way.  Whatever settings they might have, they're all still enforcing the current consensus rules.

Legacy nodes' name should not contain the word full, they are not full nodes because, as said, they "can not fully validate transaction and blocks".

They still fully validate what they can understand.  And again, these nodes still enforce current consensus rules.

Plus, as I alluded to in my first post, if you start to define some nodes as a secondary underclass, you're only playing into the hands of those who seek to create tensions and divisions in the community (and look who showed up shortly after   ).  

[franky1]

validate full data=no
validate full rule set=no
archive full data=no
mempool full transaction types=no
relay full transaction types=no
relay full blockdata=no

is it a full anything=no

logic check.
if a node only has X rules it specifically knows and checks X rules.. does it mean its a full node.
wait.
a) what if X was 1 rule, 10 rules, 12, 15
wait
b) what if x was the NETWORKS FULL list of rules

[franky1]

In such an outcome, which one of the split would the community follow, the Core developers support, and the market value?

mainly people would follow whatever the majority of merchants use. and the merchants will follow whatever one will cause them least losses and most functionality without most delay to get back to business as usual.

until an event occurs there are many decisions.
but having diverse nodes of different codebases/teams would atlease give options to remain running instead of spending hours worrying/or having only one choice.
 
EG take the 2013 bug. 0.8 introduced a new database format to store block data in a different way(levelDB instead of berkely). but they done it as a inflight upgrade that didnt require network consensus.

nodes of 0.7 were treated as equal as 0.8 but beneath the social buzzwords. rules did change. although the sentiment was that it was all "compatible"
however
0.7 nodes couldnt store data of a certain amount of 'locks'. so they stalled out at blockhight of 225430 while v0.8 continued on. thus causing a orphan drama event of differing lists of blockheight.

the network decided to downgraded to 0.7 orphaning off blocks accepted to 0.8. simply to get everyone back up and running and business as usual. and then regrew a new block height using rules about the data being below the database 'lock' limit..  until it overtook the blockheight of the chain that was 0.8
. then later released version 0.8 again and promoted that everyone should upgrade
only once majority moved to 0.8 as a mandatory event. and then progress where those running 0.7 were just told to upgrade or stay stalled out as they were just laggers/second class(not to be treated as important to the majority)

so in short the 2013 was a downgrade.. then patch then upgrade then ignore those that remained downgraded
....
the 2018 bug. if a DDoS would occur taking out all core nodes of 0.14+ would have been a patch to all versions of 0.14+ because not everyone just runs new code(0.16 -0.17) "on trust" without independent review.

so not everyone would instantly upgrade to 0.17 and some would fear downgrading to version 0.10-0.12 as this would cause bigger issues around the whole mandatory ban/segwit drama stuff(as a second event ontop a DDoS event) so patches would have been released to version 0.14+ much quicker. where a quick review that only one code function/rule was changed.

so in short in a 2018 DDoS event it would have been a patch and then upgrade later

but if segwit itself had a bug. then depending on the bug and its impact. it could go as
downgrade, patch, then upgrade
or
patch then upgrade..  

until you know the possible cause, historic damage(orphans) and future impact(functionality). asking specifics are unknown

[squatter]

(facepalm)
not fully verifying..
dont directly participate

full nodes doesnt mean just verify transactions aimed at you. it means fully validate all data to ensure NETWORK security.
come on. imagine if all nodes didnt fully verify.
the whole point of full nodes is that the whole block should be verified.

The point of running a full node is to validate payments received. That's what "being your own bank" means -- validating payments you receive against the protocol instead of trusting a centralized authority.

If a legacy node forwards valid transactions it doesn't fully understand, how does that jeopardize network security? Can you explain how Segwit jeopardizes network security and why we haven't seen the "ANYONECANSPEND attack" come to fruition?

short answer to that debate was that full nodes should validate everything. not just its own personal needs but validate the networks data.

Why? Explain how legacy nodes are endangered, and why the network is now insecure.

[NeuroticFish]

if you start to define some nodes as a secondary underclass, you're only playing into the hands of those who seek to create tensions and divisions in the community (and look who showed up shortly after   ).  

This is a valid point I cannot argue with, although everybody knows that the older software may not perform as good as the newest and here it's somehow the same thing.
And although my previous statement can be read that I'd prefer everybody would upgrade, I'm not. I think that some legacy (not deprecated!) nodes would be good to have, just because you never know...

[franky1]

(facepalm)
not fully verifying..
dont directly participate

full nodes doesnt mean just verify transactions aimed at you. it means fully validate all data to ensure NETWORK security.
come on. imagine if all nodes didnt fully verify.
the whole point of full nodes is that the whole block should be verified.

The point of running a full node is to validate payments received. That's what "being your own bank" means -- validating payments you receive against the protocol instead of trusting a centralized authority.

If a legacy node forwards valid transactions it doesn't fully understand, how does that jeopardize network security? Can you explain how Segwit jeopardizes network security and why we haven't seen the "ANYONECANSPEND attack" come to fruition?

short answer to that debate was that full nodes should validate everything. not just its own personal needs but validate the networks data.

Why? Explain how legacy nodes are endangered, and why the network is now insecure.

to those that do want to learn about full nodes. research
NODE_NETWORK, NODE_BLOOM, NODE_WITNESS, NODE_NETWORK_LIMITED (1037)

1. litenodes are more intended to just validate that your own personal transactions are verified in a block(there are many variants that can give you variant levels of data that are as simple as just blockdata related to your own list of addresses(bloom).. just the latest 288 blocks(nodenetwork limited) all the way up to full blocks of data(nodenetwork)).. but a full nodes purpose is to have full data and have that full data fully verified(all of the above).

2. "instead of trusting a centralized authority" yep. exactly.. so legacy nodes are trusting the segwit node validated ALL  transactions... (ofcourse its not your transaction,, but a full node is about network. not just personal)
(emphasis "full node" does not mean "personal only my transaction node" it means "full node")

3. full nodes are to validate the FULL DATA for network security. and to have full data to allow the other network participants to have a source to get data from.

4. and no. dont try to dilute the meaning of full for 'personal use' as thats just weakening the amount of nodes available to properly sync to.. imagine it. nodes pruning out transactions that are not important to that person.. means the next peer cant get them

imagine if some nodes had a stripped block of legacy only.
imagine if some nodes had a pruned block of only data that was still UTXO
and only 1 node was actually fully archival.

now imagine a fresh user wanted to run a actual full node. guess where they will get thier data from.
it wont be from the stripped nodes.  it wont be from the pruned nodes.

thats why its important to identify whats actually a full archival node.
same goes for validation
no one wants to be messed around with receiving invalid data, even if their actual full node is going to full validate anyway. but atleast knowing theres many true sources of believed full valid data. saves time and hassle of not having to drop connection and wait for data

in short. imagine if out of 10,000 nodes.. not all are actually full archival, fully validing nodes.
but a mish mash.
thats not supporting the network with a 10,000 strong network of full 100% 10k copies of the blockchain. the reality is far less than 10k.

as for the anyonecan spend stuff.. thats where many argued from 2016+ and the end result was dont release the wallet for making a segwit tx until weeks after nodes activated.
(the arguement prior to activation was that its fine, all compatible and people could make segwit transactions inflight without consensus)

but end result was consensus was needed. hense the november 2016 consensus
consensus was not reached. hense the mandatory
then after the mandatory. the wallet facility to actually make transactions was allowed weeks later.

..
now if there was a bug and segwit was deactivated the transactions revert back to anyonecanspend. and because pools are the ones collating the transactions. guess whos gonna pick up those funds

..
as for the if legacy dont forward transactions it dont understand how does that jeopodise the network.
(funny)
if the network doesnt actually have 10k full nodes that do forward.. but only a few... imagine that,

trying to say there are 10k nodes that do a to z. when in fact theres one a few thousand that do a-z shows how people dont understand that the numbers are fudged

...
ok imagine it this way
you go into a shop. wanting a buy a few things. you prechecked online and seen 10 in stock.. you go to the store and told actually theres only 3 in stock as 7 are either older models, asian 'replica's missing vital things or they are ex display

would you treat them all the same or question the quality of the 7
you probably would likely ask why were they not advertised as ex demo or exrepair or replica models to save your journey and so that your fully informed before getting involved in a decision process.