I don’t think that’s an unpopular opinion. And I don’t think it needs to be either or (easy for me to say I don’t do coding). But I feel like within a thread a search is redundant as is the hamburger. Those feel like things I could reasonably be expected to navigate to find. But my avatar and my notifications are things I just always want to see in Discourse. On our community you could reasonably accommodate the title + the avatar.
Fully accepting that’s my take and my (our) take only.
This is now both user configurable and site default configurable as long as you install this component.
Alright I didn’t suggest the idea originally but let me be blunt.
The old header is more important than the topic header, to me and to every user I have spoken with on my site.
However, I see no reason anybody’s opinion of which is better should stand in the way of making it an optional feature, that can be toggled on and off, whether by the site admin or by the individual user, or at least having the scroll direction be swappable.
There is literally no reason to insist that this behavior is undesirable if there are obviously people who desire it (and are coming to Discourse Meta specifically because they desire it).
As such, I cannot thank @Sam enough for the optional component to fix this feature. I will be talking with our site admin, and hopefully this should be up and running soon. It’s super cool that Discourse is a platform that allows tweaks like these to be used when changes are made that users dislike, and I’m really happy that I’ll soon be able to turn this off. If you ever feel like expanding that component, I’d still love to try a version out with the behavior flipped, but for now this is more than enough to make me happy with my options!
So, thanks again to the Discourse team for keeping the platform updated, cool, and user-friendly. 
I’m curious, would you have chosen Discourse had the new behavior always been the default?
One very good reason not to do this is familiarity - Discourse is evolving as a product and it’s already quite jarring to visit forums which clearly haven’t updated in more than 12 months, many of those quality-of-life improvements which keep many of us highly engaged are lacking. Assume for a moment that people are visiting more than your community, how fragmented do we want their experience to become?
That decision is really for the product architects, we can’t speak as to the overall vision and may never get there if all we do is bicker over how much better life would be if we could stay at cheese station C.
A big part about using open source software is being a passenger on that ride. Look at how many people pushed back on Gutenberg prior to WordPress 5.0, but slowly the community is adapting to the concept that familiar != best.
It would be a huge shame if the team ever held back on innovations to their software over fear of backlash from the trolls under the bridge.
This is a terrible, terrible way to view engaged users of a product. Folks who disagree with an implementation detail are not trolls.
IMHO, If the developers choose to implement a feature (or not), they will use a myriad of factors to do so:
Etc.
The more options there are. the greater the possibility of negative interactions between them, or for newer options to break older ones. Just this most recent beta a long-standing interaction between unrelated settings was corrected that affected our community adversely, and another was discovered as a regression.
IMHO, no one in this topic (or indeed, on most of meta) are “trolls under the bridge”. They are passionate users or maintainers of communities advocating for their particular vision of what the interface to their communities should look/work/behave like. And I have little doubt that the development team would rather like their product to be widely used and enjoyed, and as such almost certainly welcome such feedback, without having to adopt an adversarial stance against those who post here.
I didn’t choose it, my site admin did.
If I’m being honest, the main applications my particular community uses Discourse for make Discourse an sub-optimal choice, for other reasons on top of the new header. However, the site also acts as a feedback hub for an indie game, so Discourse was chosen to facilitate that part, and we have learned to deal with it. Although that does sound like I’m much more negative towards Discourse than I am, it’s the truth - it’s not a very good platform for long megathreads and games of Forum Mafia.
I still like and appreciate the platform and developers a lot from my interactions with them, and I really do think it’s good for many applications, if we’re talking about my specific one it’s already a bad fit. The new header behavior is merely another glaring roadblock, as it makes the searches and notification-hopping we do extremely often much more tedious than it previously was.
Again, I promise I like Discourse a lot - I’m just being honest about how it applies to the usage case of my particular community. And there’s no reason to think that it couldn’t be improved in those regards, through updates or plugins. That’s another huge advantage to Discourse right there.
So no, I wouldn’t have chosen Discourse to begin with. I do, however, expect that this header would have been one of my top reasons for electing not to do so (along with the abscence of pages - those are really important for FM, unless your timeline tools are significantly increased). Am I unhappy with the choice to use Discourse? Not really, but the header thing definitely made me more unhappy with it than I was before. Still, there’s a workaround now, so everybody wins! 
Then make it an end-user toggle, exactly like has been done. That’s really all I’m asking for, although you have to admit that this problem also fails to exist if the global default has flipped its functionality, another suggestion I support.
But when a change has been tested for, what, half a month now, and it’s still nothing but a hindrance to normal, everyday forum use and navigation, when do we get the right to say it’s a change we dislike? Not every change is good, and it’s rarely a good change to add complexity and extra steps where there used to be simplicity. This is one of those cases.
I didn’t choose it, my site admin did.
If I’m being honest, the main applications my particular community uses Discourse for make Discourse an sub-optimal choice, for other reasons on top of the new header. However, the site also acts as a feedback hub for an indie game, so Discourse was chosen to facilitate that part, and we have learned to deal with it. Although that does sound like I’m much more negative towards Discourse than I am, it’s the truth - it’s not a very good platform for long megathreads and games of Forum Mafia.
I still like and appreciate the platform and developers a lot from my interactions with them, and I really do think it’s good for many applications, if we’re talking about my specific one it’s already a bad fit. The new header behavior is merely another glaring roadblock, as it makes the searches and notification-hopping we do extremely often much more tedious than it previously was.
Again, I promise I like Discourse a lot - I’m just being honest about how it applies to the usage case of my particular community. And there’s no reason to think that it couldn’t be improved in those regards, through updates or plugins. That’s another huge advantage to Discourse right there.
So no, I wouldn’t have chosen Discourse to begin with. I do, however, expect that this header would have been one of my top reasons for electing not to do so (along with the abscence of pages - those are really important for FM, unless your timeline tools are significantly increased). Am I unhappy with the choice to use Discourse? Not really, but the header thing definitely made me more unhappy with it than I was before. Still, there’s a workaround now, so everybody wins!
Then make it an end-user toggle, exactly like has been done. That’s really all I’m asking for, although you have to admit that this problem also fails to exist if the global default has flipped its functionality, another suggestion I support.
But when a change has been tested for, what, half a month now, and it’s still nothing but a hindrance to normal, everyday forum use and navigation, when do we get the right to say it’s a change we dislike? Not every change is good, and it’s rarely a good change to add complexity and extra steps where there used to be simplicity. This is one of those cases, in my eyes, and I’m clearly not the only one who thinks so.
I get this is a metaphor for people preventing you from moving forward, but I’m in agreement with orenwolf here. Most of us dissenting with this change seem to be perfectly okay with trying out various new ways to find the best of both worlds (especially just flipping the direction to activate each header). That doesn’t look like arguing for no change for the sake of no change to me, it looks like acknowledging that there is reasoning for a change but disagreeing with the implementation.
I certainly do not think of it that way. I fully support:
A site level customization to revert the behavior for site owners who prefer the old way, and @sam already built it.
The idea that when a notification comes in the header should switch back to the global one, in a just-in-time manner.
That said I do feel quite strongly that the new default is a much better behavior for 95% of the audiences out there … and that’s the entire purpose of defaults.
I have full confidence in the thick skin and outspokenness of decision makers like @codinghorror and others to doubt this would ever happen.
And if this comment is meant as a backhanded remark toward the dissenting faction in this thread, then I don’t even know what to tell you. Things may have become ever so slightly above room temperature at times in this thread (and I admit my prose skews toward very blunt in my online writings), but certainly nothing remotely resembling trolling.
You just ramped up the temperature to 1000 celsius 
for certain people using the word “thread” to describe a “topic” is a trigger word.
Anyway… I appreciate all the feedback here, and love that people are popping in with opinions and ideas.
I am also so impressed by theme creator, it makes building theme components for Discourse so easy… you don’t even need to install Discourse to write themes… but I digress…
My main objection earlier on was around how repetitive this got. Anyway, we will get the switch on notify idea in some time in the nearish future, just need to figure out how to make it safe so stuff does not get to flickery.
I’ve been trying to get used to this change for a couple weeks now, but I still don’t like it. Personally, I’ve never had a need to see the topic title in the header, but I used the notifications all the time. At the same time, I rarely used the search or hamburger menu, so I agree with the people I quoted above; I’d like to see the avatar/notifications always, and just hide the search and hamburger menu. I understand that now I can use the theme to use the old version, but I still wanted to give this feedback to (hopefully) continue improving the new version.
Just to give a little background, one of my primary use cases is when I have a whole bunch of notifications (say in the morning). I want to see what all of them are, and since most of them are likes I just want to take a quick look at what was liked and then go to the next notification. The old workflow was “click, glance, click, glance, repeat”. Now it’s “click, glance, scroll, click, glance, scroll, repeat”. Doing this for 10+ notifications is a pain. While this is an important use case, my other workflows also benefited from notifications being readily available, so I really do think keeping the avatar/notifications visible at all times is important, even if the rest is hidden and replaced by the topic title.
This change is really bad for me. I am constantly trying to scroll up a bit in Chrome on Android but it keeps being re-hidden for some reason - maybe because my finger brushes down or something but my experience is that it seems to depend on the post size so a big post prevents it from returning. Very frustrating and our community are not happy
On top of that you seem to have made another change: previously if I hit my user icon the menu drop down appeared and then you hit the user icon again to hide it. But now the user icon disappears and I can’t work out how to get back to where I was without reloading the site.
(There was a link at the bottom that said something like ‘hide this’ but when I clicked it it just hid the link and it’s not there any more…)
These new menus simply slide back. Swipe to slide them away again.
The advantage of the new approach is that the menus have more real estate now.
(But like you I still prefer the ‘classic’ show/hide buttons. Looking forward to that Theme Component 
)
You can use this theme component, it will restore the previous header. It was previously mentionnent, but maybe you missed it:
Ah thanks. This is what I refer to as the Apple effect: making changes without actually providing any method to know those changes. They do this all the time. One day I couldn’t push up with all fingers on my iPad to see all running apps; sometime later I discovered I had to double-click the button now. Then there was that time I updated OSX and could no longer swipe the desktop right on the mousepad to get calendar and calculator apps.
Ah, thanks a lot. Okay I got that installed. Annoying you have to select a theme for it to apply, though. It doesn’t seem to latch onto default, but I’d just created a new theme with the default light scheme.
its wery good lookong of new ux mobile version …
It seems to be velocity sensitive (not sure if this is what you’re seeing). Unless I scroll quickly and deliberately, it doesn’t seem to pop.
When on Chrome on Android, I noticed that you do need to scroll up a fair bit. O am suspecting it is because of the URL bar which reappears at the same time. Not sure whether this could be improved. In the PWA mode it is much better.
I just thought I’d come back to this and add some further thoughts. I’ve been doing some decent level reading and browsing here the last couple of weeks and have left this setting be.
I completely get this now, but specifically in the context of this forum where I tend to be reading threads for much longer than on the main forum I’m using where I do a lot of bouncing around between topics. In that sense the choice is great to have so it’s good work all round.
It is definitely velocity sensitive. Scrolling verrrrryyyy slowly up does not trigger it.
I’m completely happy with this. My short term memory (*) is a bit shot at the moment, and this really helps me.