In this case, I doubt someone thought "hey, let's mess with Linux users and serve them a bunch of garbage, it'll encourage them to move to Windows".
Also, the percentage of desktops running Linux is still tiny relative to Windows. The last time I saw any numbers (about two or three years back), it was ~90% Windows, ~7.5% OS X, ~1.5% Linux, so there is little incentive for Microsoft to care about Linux desktops.
Not for this one reason, but for 'webpages just work faster on Windows', sure.
If the only change necessary to improve speed is to update a string in the UA, I would say that's pretty hard to do accidentally. In theory, it could be an artifact of some internal QoS systems that prioritize requests from Windows boxes without the explicit intent of damaging the experience of non-Windows users, but that seems like a stretch. It'd be interesting to know if Mac UAs experienced similar slowdowns or not.
It's much more likely that Microsoft is interested in doing what they can to make sure that people feel "things just work better on Windows", and little limitations or breakages like this are the perfect way to do that. Windows is Microsoft's bread and butter and they haven't forgotten that.
Not knowing the code behind it, but knowing the history of the OneDrive for Business client, I could see a few ways this could happen. The OD4B client has gone through a bunch of iterations before being merged with the regular (public) OneDrive client, so along the way, there could have been optimizations added on that were only available to the new merged client. Due to constraints, this could have been done using a simple UA check. These optimizations could have been merged down to the OSX/Linux clients (not sure if they're using a merged one there, or still the separate ones) but since different teams work on the clients (most likely different teams on each client) and server, the Linux client team never told the server team that the it also supports these optimizations.
Easily could see the above scenario happening, it's not completely crazy when talking about big companies that teams don't have the best insight into eachother's work.
Disclaimer: Work on Microsoft. Not on Windows, not on OneDrive, completely different part of the company.
I'm sure a company like MS, who've had traumatic run-ins with the DOJ in the past, has internalized this phony management persona even more than most. And I do believe low-level developers who work on the individual clients are probably 100% earnest. They usually are. It's the people in the political structure who set their priorities and schedules and use them as pawns that you have to worry about.
Even if your postulation is exactly accurate, MS still bears responsibility for crippling the Linux client, no matter which specific employee the communication "fault" falls on.
I have no inside knowledge of Microsoft, this is just generic commentary and extrapolation on corporate politics in general. It is possible that I'm incorrect and this is truly an honest mistake. If this gets any attention, management will certainly claim that, and we'll never know the reality of whether it was unintentional or whether "The Linux client supports those speedups" was intentionally swallowed by someone along the line (or whether the patch was delayed for additional review because it was "accidentally bundled" with a bigger branch, or whether the priority has just been "innocently" pushed down, or whether the team is "just understaffed", or...).
The proof is in the pudding with corporate politics. Being in management is being a professional politician. That's always true. They're always going to try to tell people what they want to hear and cloud up the picture around things that that person doesn't want to hear. If Microsoft is continually engaging in a series of accidents that hurt non-Windows clients, that pattern is sufficient proof of management's intent, regardless of what they claim.
There is no reason with the modern setups that entirely different UI pages need to be sent based on user agent detection. Every best practice I have read says to detect for features, and fall back to something more general when a feature isn't present. Clearly that is not happening here, and that little bit of technical incompetence is enough room to slide in a few hundred mb/s of transfers for certain targets.
Microsoft has a long history of pretending to play and throwing curveballs. All the way back to Dr DOS. Their run in with the DOJ was just one time they got caught. They are even doing stuff now with CPU detection and not providing windows updates to people with and older but still "supported" OS and a newer CPU. Just because they won't let it go, they are still patent trolling android handset manufacturers.
Why do we think microsoft has changed for the better?
My guess is that they whitelist javascript or other things based on specific browser strings.
Onedrive for Business is a real shitshow product. IMO, the people supporting it scramble to make it perform at all.
That would only be slightly less worse than the sabotaging theory.
At the same time, this company has printed money for 30 years, not just beat the competition but killed the competition in many cases, and even made that an internal mantra. They have a pile of very smart people, money, etc.. there are also browser compatibility test services. They could have a VM grid with every interesting browser ever and it's not a big cost to them. A 15 person startup? I get they you might skip some browsers and focus on a key supported set. The tooling for that VM grid is also something I think a lot of the community would eat up of MS made available, there are lots of ways to build good karma and improve community relations.
Intentional or not, it's just bad optics, if nothing else. A slate with the windows 10 linux subsystem looked like an interesting environment to play around with and maybe give a shot. This sort of thing makes it look a lot less interesting. No joke, I was thinking of getting my kids one.