(Some reasons) Why segregated-witness is bad -
https://www.reddit.com/r/btc/comments/5zn4oi/some_reasons_why_segregatedwitness_is_bad/1) segregated witness is a big change from the whitepaper, it changes fundamental concepts and outlines of Bitcoin.
2) it adds a huge more 'technical debt' which will make actually fixing the blocksize problem much, much harder than it is right now.
3) it does not actually fix anything regarding scaling, it merely pushes the issue off for another 6 months-2years, but if btc survives we will be back here discussing these same things again, but with more difficultly as segregated witness makes solving the blocksize issue much more difficult on a technical level
4) there is a massive, absolutely huge amount of misinformation and lies that are being used to try and persuade people to use segregated witness, I don't usually like things that have to be lied about to sound good.
5) segregated witness does not double the block size, far from it. In fact bitcoinCore gives an estimate of 1.6 or 1.7 times the current size (which was too small over a year ago if we want to avoid full blocks like Satoshi said we should). They estimate that we could get up to 2MB with a working LN (not yet developed)
6) it increases the amount of data that is sent at a higher rate than it increases capacity, increasing waste
7) segregated witness as a soft fork includes far more technical debt than as a hard fork, which makes all the above problems worse
8 ) there is no reason why a hardfork would be bad if the supermajority if the hashpower was needed to be using it before it activates, such a well prepared fork is just an upgrade
9) the things that segregated witness does do (tx malleability) can be done with FT or something better than segregated witness, or segregated witness as a hard fork (to lessen the amount of technical debt that will interfere with a real scalability fix)
.