IMO, it's because OWA Light was a last resort, and the wrong one. For a Linux user that was looking for lightweight access, using a 3rd party email client over IMAP was the way to go. HOWEVER, Microsoft effectively killed that already.

One can technically access O365 mail from 3rd party clients, even making correct use of XOAUTH2 from ancient clients like Mutt and Pine (now "Alpine"). But that requires the domain admin to specifically allow those clients because the auth flow includes a client identifier - and you can't just fake it like the HTTP User-Agent. That effectively bars use of all 3rd party clients (Ex. I can't even use Thunderbird to get my work email). In addition, they've recently changed some things related to the "Authorize" and "Device" flows (https://support.mozilla.org/en-US/kb/microsoft-oauth-authentication-and-thunderbird-202). Lastly, the "Authorize" flow requires that a company exists and is registered to the client (FWIW, this is currently preventing Alpine from supporting the Authorize flow, as it is a one developer open source thing).

Long story short: The exchange ship already set sail. The real argument/debate (besides not using exchange to begin with) is in ensuring other IMAP clients can still use the service. Who cares about OWA Light?

Dude, you can't even read. Your ad hominem is backed up by nothing yet you want to question others thinking? Weak sauce.

How many widgets did you sell before you stopped believing it was something no one wanted?

FWIW, that's covered by "A":
A) More and more Windows users are changing their user agent string to something unknown.

A privacy blocker is but one way of doing so.

About 20 million units sold. And that doesn't count the other brands of VR goggles.

Do you know how many companies make screwdrivers? And they've been around for AGES. Pet peeve of mine is when people bash a person/company/product because it only managed to obtain 1/10th the market share that Google had - most small businesses would be massively overwhelmed by that! Support small businesses! (and yes, I know Meta is not a small business, but that furthers this point)

For one, people don't have to ask for a thing for it to be a good idea to introduce to the world, let alone a very small subset of the world. Secondly, one persons failure is very very often someone else's success (ex. mp3 players and the ipod, or smartphones/feature phones and iphone, or electric cars (first was in 1832), etc etc..). Long term, baring a complete societal collapse, I think they'll be everywhere.

The stats from TFS don't add up to anywhere near 100%. Turns out, the latest numbers have an "Unknown" bucket that's 21.45%, and its rising path almost matches the declining path of Windows in their graph. So either:

A) More and more Windows users are changing their user agent string to something unknown.
B) More and more Windows users are moving to another OS and using a browser with a user agent string that is unknown.
C) New user agent strings have shipped but the statcounter site hasn't updated its user agent detection
D) Combo of the above

My bet is "C" with some growth of Mac* and Linux usage. When the OS stat counter can't identify over 1/5th of browsers, I don't know if I'd put a whole lot of weight into it. That "Unknown" bucket should definitely be broken out more.

But this CLEARLY does not NEED to be turned on, so why default it to on?!?

If they're defaulting some settings to off that really should default to on, that's also an issue! That doesn't excuse this behavior.

Just want to throw this out there for anyone that has such an account but doesn't want Microsoft to see their data...

You can use rclone (https://rclone.org/crypt/) with an encryption layer over One Drive. This works on Windows, Linux, Chromebooks, Macs ... probably lots more. It can encrypt file names and file contents. Use that free space!

I'd posit that variety is the solution that fixed this already. In the Linux ecosystem, you can pick a distro that behaves as you'd like, or even linuxfromscratch it and make it exactly as you'd like. That also highlights the issue here - Microsoft can't or won't support all the administration levels that all the people want.

IMHO, this particular item is something that should have been rolled out differently to Home, Pro, Enterprise, and Education editions. For commenters above that claim OS's aren't for us power users anymore but for morons, fine - push this backup feature to "Home" and make it opt-out. For users on "Pro", push it out but make it opt-in, maybe with a news item pop. For "Education", hide it from the user entirely and have the opt-out/opt-in setting controlled via the group admin. For "Enterprise", don't even push it, but offer it as an add on that they can add and deploy via group policy or what have you.

You're treating it as if the "OS" can't be provided in more than one way, though they do that all the time for other features.

HAHAHAHHAAHHAH AHAHAH WTF ARE YOU TALKING ABOUT!?!?!

The correlation of Windows Device ID to site URLs visited at a specific time seems to be the key. Visited URL's (the website) do NOT get the Device ID in the traffic. That ID + IP + time would need logged somewhere for this.

AFAICT, Microsoft retains a log that includes the users IP address (apparent and real?), Device ID, and time. This may be normal windows check ins/telemetry stuff. I don't get the impression they are recording all URL's the machine had visited, but I don't know they're not.

The websites have logs of users apparent IP, the URL they visited, and time.

The IP + time can be cross referenced to then correlate traffic with the Windows Device ID.

Going back to your far fetched claim... something would need to be logging a GUID from your Linux install and your current IP address (and your apparent IP address) and the time. That's not happening. Furthermore, why the hell would a filesystem GUID be used for that purpose?!?

PS: Apparent IP refers to the IP that other hosts on the internet see as your source IP when you make outbound connections, as opposed to the IP address that is bound to your network adapter. IE: think NAT.

Hey, we're in agreement on the practicality of these space data centers! What I keep bringing up is "why" then, cause these are all quite obvious issues.

In my doomsday scenario, the rest of the internet is gone or in shambles, or at least the big data centers thereof. And no, not everyone would get a starlink terminal. In fact, I think they'd cut off many of those users, reserving access to their data center in the sky for only the rich and/or powerful.

You don't need to exit the network if what you want is on the network.

FWIW, it's very easy in vim.

* CTRL-v
* use movement keys to highlight whatever block/column you want
* "y" to yank that into a buffer (or "x" to cut it)
* move to the line you want to start pasting on and to the column position where you want it to start
* "p" to paste in that block

The yank, cut, and paste commands should be second hand if/when you're using vim for any length of time. It's like knowing CTRL-c/CTRL-v/CTRL-x in Windows.

CTRL-v (visual block) is the magic button there. TBH, I've never used it - I usually zipper the lines together with a macro - but I should use it more. I use "v", which is normal visual mode, quite a lot though.

Anyway... sounds like it's just as easy, if not more-so, in vim. And you can get gvim if you want a GUI for it.

Just adding a CLI option: scrot

Example:
scrot -F ~/Pictures/Screenshot_%Y-%m-%d_%H-%M-%S.png -s -e "xclip -selection clipboard -t image/png '\$f'"

Running that puts the mouse into select mode. Click and drag a rectangle and it saves that area to the clipboard and to the file. You can then paste it anywhere, or open that file if you want to edit it. You could easily change the "xclip" part to run whatever image editor you'd like on the result (Eg. for the markup you mentioned).

Probably not what you want, but that's very simple in vim. I'm assuming you'd want it to do some nice formatting of the data so you can edit it?

vim somefile.json
:%!jq .
... make any edits you want to make ...
:%!jq -c .
:wq

Explaination:
* ":" - puts the editor into ex mode
* "%" - grabs the buffer contents (the currently open file)
* "!" - pipes the data to an external command
* "jq ." - is the "jq" program applying json formatting
* "jq -c ." - runs jq's compact on the json to put it all back on one line
* ":wq" - enter ex mode, write, quit. IE: save the file and exit.

If you just want it to open, just open it. If syntax highlighting is taking too long, turn it off.

I think you missed my point. Millions of seniors dying a bit early DOES impact the statistics in the handful of years following it; They *would* have died and contributed to those numbers later.
To put another way, I was not implying that the COVID related deaths had a noticeable impact on the birth rate.

I don't disagree with the rest of what you said, but part of this is off the mark. The "C-levels ... do not get" part. They get it. It's a risk/reward thing. People introduced loads of bugs as well - that's what the LLMs are exploiting! IME, they're aware, but they believe it can be solved by more use of it.

Whether or not the issues will cost them enough to regret this is another question. With so many doing it, it's like the old mantra, "Nobody gets fired for buying IBM," but in this case they can just cash out and move on.