UASF is about NODES
and yes them nodes users need to upgrade yet again by august (that makes 4 downloads just to try different things related to segwit activation(facepalm), before even getting to the real segwit keypair wallet download which is the real 'feature' gesture(wont be available until weeks after activation))
A UASF forgoes the need for miner signalling because economic nodes are given more time to upgrade to the new rules and begin enforcing in the future.
UASF (bip 148) is CODE so it IS about nodes. not some website that people socially say yes to.
but right now because UASF is not ready, the only way to show support is by 'social flagging' (old nodes displaying a comment in their useragent)
(sybil attack risk) and lame websites where 'consultants' that dont even need to run nodes can sign many times as different subsidiary of their consultant business (for instance: bitcoin embassy, satoshiportal, bylls are one business with multiple subsidiaries pretending to sign independently to gain support count 'socially')
although the CODE is not ready. blockstream/core/uasf are telling people to change their useragent comments to flag for it if they want it
Successful User Activated Soft Forks require a strong consensus from the economy to be successful. BIP148 also is subject to changes as it is reviewed, so some minor details may change before it is ready. Until there is sufficient consensus, it is not advised to use any binaries that implement BIP148. However, the BIP148 reference can be applied for testing and review purposes. If you want to signal support, change your user agent comment to be “UASF-SegWit-BIP148”, along with communicating with other Bitcoin users that you support BIP148.
again proof that it is about code and nodes (software) (that is not ready yet)
Can BIP148 be cancelled?
Yes. In the event that the economic majority does not support BIP148, users should remove software that enforces BIP148.
and now for the technicals.
UASF requires 85% hashpower by august to be sustained where by if still at 85% near the end they will start orphaning off nay-sayers/abstainer pools blocks to then (trick) the block count into thinking there are less nay-sayers/abstainers(this 85% may change before the CODE for UASF is publicly available)
remember you cannot reject a block in the network by just having a signed message on some polling website. so it IS about nodes that have code that does the rejecting (not available yet).
Soft forks rely on the economic incentives of the majority of miners and economic actors to reject invalid blocks based on the new ruleset. Since the new BIP148 rules are a stricter set than the old rules, any chain split means the chain with the old rules would be in danger of being wiped out. If the majority of miners enforce the new ruleset, all blocks produced that are invalid in the new ruleset will become orphaned.
and yes, by rejecting blocks. its not about pools doing funky things like throwing random data.. its simple (word twisted) that if your pool is one of the 15% nay sayer where the block actually contains valid transactions and the blockreward follows the current 12.5btc amount.. but is not wearing the segwit sponsored version bit.. its biasedly orphaned and treated as a naughty boy and thrown aside.
this will cause chain splits. again not based on bad data but based on rejecting blocks purely because they are not kissing segwit ass.
(thats the funny part 'segwit is backward compatible' but try making a normal old block and get orphaned)
TL:DR;
UASF
actually means hard consensus(if you untwist the blockstream shep stroke to sleep 'soft' wordplay games) with 5-15% block rejection/split of network if it gets to a sustained (currently proposed 85%) minimum in august-november
the real funny part is.
blockstream believed that a november 2016 rule change could have been active within a month via only pools
blockstream believe that august rule change could be active by november via a node+pools
so if we actually, back in late 2015 done a proper node and pool (hard)consensus of a 1 merkle upgrade for the entire network with a 4mb baseblock(no weight) where that 4mb was dynamic after that.. we would already have had a united community where everyone gets what they want.
+ real blocksize growth even for native keypair users (because 1mb baseblock no longer existing)
+ those that want to use segwit keypair can
+ all nodes on same peer network fully validating old and new keypairs equally (no TIER upstream filter/stripping block nodes)
+ and hopefully other common sense changes like a lowered txsigops limit of say 2000, that does NOT increase even if the blocksize increased to truly mitigate quadratics.
- (though native key users can still malleate(which is still possible even in soft segwit))
all without wasting 2 years and 4-5 downloads in that timespan
even funnier. even if segwit still gets rejected for the half gesture it is.. they wont just accept it as a no, and then ask the community what would work best to get a nited community.. they will just try again and delay things for another year with segwit in its current form but with yet another deadline
Can BIP148 be cancelled?
Yes. In the event that the economic majority does not support BIP148, users should remove software that enforces BIP148. A flag day activation for SegWit would be the next logical steps and require coordination of the community, most likely towards the end of 2018.
so if the community say no this year(for good reason) dont expect any new/alternative community uniting re-code from blockstream devs before 2019. they will just re-push the half baked segwit as-is today with a new deadline and new buzzword play to stroke ego's to sleep