It sounds like you object to the new design on paper, but not in practice.
1) I require legacy ports. The USB-C ports "sounded" like an issue, but in PRACTICE turned out to be MORE flexible and with LESS to fail.
2) Like a majority of users, I adapted to the keyboard and type faster on it with greater accuracy. Apple did their homework and did many tests, so this was a no-brainer upgrade. It's a drag if you spent time trying to adapt and couldn't, but most did. Apple sells to "most."
3) What workflow do you have that's impacted by the new design? I'm a pro photographer and writer, and I connect to a ton of gear. I'm curious where the new machine fell short.
4) Terraced battery will come. But after a week and my machine settled down, it BEAT my 2015 MBP for battery life and isn't far off from my air, while also delivering superior results via the screen.
Anyway, if the old tech works for you, great. But the tech is moving AWAY from you. It won't do a U-Turn just for a few stragglers hung up on 5 year old tech.
R.
With me and my colleagues combined experience we can comfortably object the new design just on paper. We won't be far too wrong in practice.
We have perhaps 5 boxes full of previous gen Apple branded dongles, even beige ones before they went all white, I know full well how to adopt myself when standards around me changed. What I am trying to say is not really that hard to grasp, I just want some means of smooth transition instead of being thrown off a train. We owned Macs such as Mac Mini 2012 / MBP 2011 that had both TB1 and FW800, I don't know how much potential TB1 bandwidth I was losing by having FW sticking around, but I could stay my FW devices plugged in until the rest of the industries picked up on USB3/TB. By the time FW was phased out, those machines were also due their service time, at that point the retina MBP with TB2+USB3 without FW made perfect sense.
Whether or not the butterfly 2 is an improvement is subjective, but what I can say for sure is that a KB is bloody basic part of a computer, if someone wants to roll out a replacement with certain benefits, fine, but the prerequisite is to make sure it absolutely works first, for most people. I am not gonna pull some number out of my arse, but at least myself, or every other Mac users around me who have tried the new switches, none of us had a particularly good impression about it where most of us noted it would take substantial amount of time to get used to before typing on it error free. And we all did fine on the previous gen of Apple KB.
I work in music recording, the workflows primarily deal with audio of course, but some times overlap to visual media in print, photography, and video, so it is a wide spectrum of needs I have. Digital audio interfaces need to run with minimal latency, any extra layer of interconnection would create dropouts or delays. Dongles / adaptor / cables what have you. And USB-C or TB3 native audio interfaces don't exist yet, and when they do they will be expense that does not enhance our quality or workflow, as audio doesn't need a lot of bandwidth until you tread into like 48 channels or very high frequency/bit depth, in which case there are dedicated hardware and protocol like MADI and workstation / reck class computers for the task, laptops have no business in that chain. In other words, a 2015 rMBP with direct USB3, or a 2011 unibody with FW800 actually are more useful in for example a live stage scenario where your mixer or guitar box runs with USB2/3 or FW (yes FireWire, they are still everywhere).
I honestly don't know what you are talking about with "in PRACTICE turned out to be MORE flexible and with LESS to fail." I understand TB3 is a flexible standard for how much it encapsulates so yes it can be more flexible than ever, but how is there less to fail. Any extra stop gap in a signal chain is a point of failure, and any adaptor has 2 extra ends of such points. As long as the world that the MBP needs to interact to are not mostly USB-C, it entails more points of failure than many previous gen laptops with "legacy ports".
SD card slot I can deal without, since some of our better cameras are on CF/CFast or XQD which requires card readers anyway, and being USB-C only doesn't make the situation better or worse. Though again there isn't a XQD USB-C reader yet and likely won't come any time soon. The irony is that even USB3 port in some higher end DSLRs is still a fairly recent adoption. And while we are at it, there is though one positive of going all USB-C that I can acknowledge, for video related transfer, be it during shooting or post or delivery, having higher bandwidth is always a plus, and the 15" MBP indeed is a beast at that especially for a laptop.
HDMI, or any direct video output, is what I consider essential on a laptop. The chance of occasions that the display you need to plug into supports HDMI is like >80%, it is the current standard and won't go away any time soon. Not just the monitors on our desks, I can be in a room trying to playback our music videos on some client's HDTV or projector, or I can be in our guest room trying to show some rough edits on our studio couches. While it is true that the USB-C / TB3 specs already encapsulate DP and HDMI so technically it *is* a direct video output port, but the fact is the rest of the world hasn't moved on as quick as Apple wishes to, and the user will just be trapped in between. I am glad that at least with macOS and iOS, the AirPlay display mirroring gives a wireless solution for our in-house usage, in the whole building we got like 10 of Apple TV 3rd gen specifically for mirroring purposes. But again this only happened because we got full control of our premise wiring, once step outside the world is HDMI, or the occasional church or office that stick with VGA...
I am smelling something similar with the tbMBP situation, it reminds me when the trash can Mac Pro rolled out and some folks here unironically recommened 32 daisy chained eGPU with the TB2 ports. In our studio, we still have a cMP 3,1 8-core running 10.6.8 and FCP7 cutting HD videos, it only recently got retired due too much demand to go 4K. We would have moved up an nMP or even an iMac for this purpose but we couldn't, because the primary storage is that 14-bay Xserve RAID that runs to the cMP via dual optical fibre channel, which is a PCI card. Thunderbolt fibre channel solutions do exist, but why should we invest in a solution where the problem should not even be there?
The 2016/2017 MBP is a beast, but it is a specialized beast. Without an alternative, if the compromised features are essential to a user, then it is a useless machine.