Props for citing real numbers! I hope other people reading this thread are looking at your comment and understanding that this is how you make reality based comments. One tidbit I will add: that's more than they have ever spent on development historically, including after adjusting for inflation. IIRC it's about quadruple what they spent back when browsers were desktop only when they had their highest market share.
I think you're probably about as dead wrong as it's possible to be on this front. First they ship millions of new LoC to Firefox on a monthly basis so the engineering efforts are open for all the world to see.
Secondly, if more than half(!?!) was spent on, say, Pocket, or Fakespot, then you would see a rise and fall in spending coinciding with the onramp and closure of those programs over their lifetimes. But in reality we have seen a steady upward march in spending, and so the interpretation that passes the sanity check is that they fold these into their existing budget with the existing development capacity they have which is variously assigned to different projects, including(!!) Firefox, where again, their annual code output is monumental and rivals Google.
Again I have to note the blizzard of contradictory accusations throughout this thread. According to one commenter the problem is they are biting off more than they can chew and need to scale back all of the excessive Firefox development they are doing (and I recall previous commenters speculating that 30+ million LoC was not evidence of their hard work but "bloat" that was excessive and that they probably could cut a lot of it out without losing functionality). But for you, the obvious problem is they're wasting all that capacity on side projects and not putting enough effort in the browser.
I run Immich for more than two years and there was an upgrade to 1.33 I think around spring 2024 that required special instructions on editing docker compose file because they changed the vector database. I think there was also a database migration same year when - if you did not update the version regularly - would need to run two step upgrade. They provided plenty of documentation always. A while ago sync was quite wonky but they improved that a lot lately.
Huh? What are you maintaining? The PostgreSQL db and extensions are provided in the container image. You do not have to use your own external PostgreSQL.
Of course, you may have reasons to do that. But then you also own the maintenance.
I have never had to maintain any PG extensions. Whatever they put in the image, I just run. And so far it has just worked. Upgrades are frequent and nothing has broken on upgrade - yet at least
Hmm.. I guess if this explains why my new work Dell Latitude becomes extremely laggy and unstable when doing Teams meetings with multiple video streams. My 5+ year older Dell Latitude did not have this problem.
Screensharing on a browser and trying to change tabs is a constant frustration; you have to wait for that overlay to get into the moveable state. Sometimes I have to switch between chats several times to get it to acknowledge that I've seen that latest message. When you add a code block at the end of your message, it's a toss-up whether you can type outside of it afterwards with the right or down arrow-key. If you can't and you need to, you just have to start over AFAICT. Making the codeblocks in the first place is usually a hassle even though parsing markdown is a solved problem. It has a lot of redundancy with Outlook and I often need to clear notifications from both. Search is unreliable. Sometimes I open an active conversation and it decides to scroll me back a month.
That covers what I'll encounter in a typical week. There are one-offs as well. It's not the worst software I work with, but talking to my team should really be zero friction.
I'm using teams at work and it's a laggy buggy mess, even with fairly beefy machines (eg. 64GB RAM, nvme ssd, workstation gpu). By this I mean when you click on a button or hover over something on the UI there is frequently more than 5 seconds for it to respond (eg. stuff like hovering over a button, it should show a hover state, but that won't appear unless you park the mouse over it for several seconds).
We have 5-6 different "endpoint protection" and security related pieces of software running on our machines at all times. We also have enterprise SSO via SAML2 which is constantly logging us out, saying we aren't logged in, re-prompting over and over to enrol the machine into some management policy which then hangs the program if you click yes, and makes you re-authenticate (eg redo login and MFA) if you click no.
It frequently just hangs when you click join on a call. Sometimes when you are talking it stops responding but other people can still hear and see you, which is annoying because if you un-mute or take over the screen in a large company meeting, but then get stuck with mic on or presenting, everyone can awkwardly keep watching you while you can't stop doing either of those for 45-60 seconds.
Many of these problems are probably just due to the machines being hampered by huge amounts of instrumentation/monitoring/interception, but teams is much worse than other electron apps. For example, Slack and vscode do not exhibit these problems on the same machine.
I'm at the absolute opposite end: Teams was Good Enough when it launched, but declined ever since: you can no longer fullscreen screen share, fat empty margins everywhere in the UI and it nags you about addons and AI stuff.
Apart from it being slow, a memory hog and having a shitty web-based UI that feels out of place regardless of what OS you’re using. It’s missing basic features.
One example is the inability to share only part of your screen. This is essential if you’re working on a large, ultra-wide monitor. There’s been a feature request for this on Microsofts feedback site for years.
Also, how embarrassing is it that the biggest software company in the world is not able to make a decent native app and has to resort to this html-app nonsense.
perception and expectations play a big role. different users have varying sensitivity to latency and people's standards for software performance can differ based on what they're accustomed to.
So are you saying I have low standards or high tolerance in regards to software quality?
Let me say this again, Teams works perfectly on my machine, it's largely indistinguishable performance wise compared to Slack on the same machine, if anything it runs a bit better.
I'm not going to try claim Teams isn't in fact terrible for other people, my point was that as I suffer no problems with it, it makes me wonder what's different on my machine compared to these other people's, and as I don't have any security software except Windows Defender, I speculate it's maybe their corporate managed security software is a factor.
FYI the big companies provide their own NTP servers and pools. You can use them if you’d like.
They also don’t use the reference implementation (which is maintained by the group this donation is for). Your distros and software probably doesn’t use it either.
The commenter above who thinks shutting down the NTP Foundation will hurt FAANG because they “leech” off of NTP Foundation is completely uninformed.
This seems incredibly unlikely. They all want clients to have accurate time, because it underpins things like sessions and TLS certs. It would also almost certainly be trivial to proxy back to regular NTP.
Even if the NTP pool somehow died, all it takes to make your own Stratum 1 NTP service is a GPS chip. An old phone probably makes a great small-scale NTP server, or an ESP32 with a GPS chip attached. 20 years ago it would have required exotic parts, but they're mundane, cheap and omnipresent these days.
Not only that, but the owners of big companies are actively lobbying to pay even less taxes. They are ideologically opposed to supporting public benefit projects.
