I'm pretty sure even with 2MB blocks we'll continue to see these unconfirmed transactions, we're getting big every day so expecting users vanish magically is not a reasonable thing to do,
Even with 1MB blocks very soon normal users will not be able to run full nodes and normal users really don't need to run a full node unless you are invested enough money to need validating every thing yourself, then you wouldn't mind to spend less than $2500 to get hard drive a decent mother board, ram, a decent CPU-GPU and at least 100MBps internet access then start running a full node to secure your plus $100K investment under your own terms.
Now we could all agree Bitcoin was never meant for our time and we should wait another 10 years for technology to improve so normal users could easily run and maintain a full node.
While BU wants a simple block increase, SW is going with a smart plan to reduce the size of transactions and same time increase the block size.
Those who refuse to upgrade to SW compatible versions better to stay behind, those miners continue to mine spam and fill blocks with native keys should get blacklisted and their blocks rejected, we're not playing games this is serious, if SW has a fix for spam attacks and allows miners to filter them then any body not signalling it should be blacklisted and their blocks rejected.
If miners keep doing what they're doing then where are the honest miners wanting Bitcoin to thrive and not stuck like now?
in short your admitting segwit as is doesnt solve the issues.. thats step one which i congratulate you for admitting and seing..
but..
simply blocking pools/nodes/tx is literally killing the 46m outputs. literally creating an altcoin. literally causing a tier network and literally causing MORE network issues than a hard consensus of real scaling would have solved, which 'going soft' was trying(but fails) to avoid.
segwit DOES NOT!!!! guarantee 2x. its all dependant on keypair utility,
segwit DOES NOT!!!! prevent native malicious key issues
segwit DOES NOT!!!! guarantee bringing fee's down
the solution is simple
more transactions per block by having a single merkle block thats more than 1mb. not the tier cludge
reduce the chance of one spammer having 20% control of a block with just 1tx (the cludgy maths core will not sort)
new code that limits bloaters
bring in a NEW fee priority formulae which actually works to punish frequent spammers/bloaters and reward infrequent/lean tx makers. instead of punishing everyone nearly equally.(if they spend 6 times an hour or just once a month all punished the same)
core need to really stop trying to push the cludgy temporary gesture right up until 2019 and instead make a plan B of a hard consensus peer network that actually does a proper job without the cludge.
its not about core abandoning the core bus and jumping on the BU band wagon. its about core getting their thumbs out their ass and actually doing things the community want
and dont even try to pretend utopia via mentioning litecoin.. although "activated" the litecoin pools are not actually using segwit. their sending their block rewards to native keys(L not 3). yep the pools are not risking using segwit keypairs. so all the gestures and hope cant even be judged yet. because litecoin is not even truly utilising segwit.
get your heads out your asses thinking its segwit or no way and realise there is another route. where you can have a proper single block for everyone. and yes that still includes the voluntarily opt-in segwit/ln features. but actually does things the community want like limiting sigops and adding a new fee priority formulae and other things can than only occur in a hard consensus upgrade.
this combined would unite the community and get everyone having their cake and being able to eat it