WTF Bites
-
And don't get me started on what pile of manure is Azure.
I fucking hate Azure and everything that goes along with it. I especially hate the push to Azure AD. Fuck right off. For almost all of our clients it is a shitty fit with ongoing costs and they can go fuck their hats.
The push to Azure and AWS
cloud computing also annoys me. I have yet to do a cost analysis on it where it is not significantly more expensive. I am certain that there are use cases where it makes sense, but none of them are ones which we work with.
-
@Bulb yes, yes, and yes, to all the above.
-
The only reason to have Windows is that deployment is easier and customers want Windows anyway. Oh, and to run Office, Teams and the rest of those abominations.
And basically every single LOB application ever produced.
Deployment and management of large numbers of devices is significantly easier with Windows but the goddamn bugs, bugs and more bugs with every single update and shit built in to the OS just not working most of the time (like the search function outlined above) are driving me up the wall.
I am probably more annoyed than usual because since we moved in June I never even bothered to setup my Windows workstation in my home office. I connect to a Windows VM when I absolutely need to, which is not that often. But my iMac and Macbooks are a pleasure to work with and I have no reason to setup my Windows machine so far. But, the thought of trying to deploy hundreds of OSX or Linux machines for a client makes my eye twitch.
-
@Polygeekery said in WTF Bites:
The push to Azure and AWS
cloud computing also annoys me. I have yet to do a cost analysis on it where it is not significantly more expensive. I am certain that there are use cases where it makes sense, but none of them are ones which we work with.Our customer has made us to a couple of studies on moving to the cloud and it's always been more expensive. We basically run our own private cloud for our stuff (except for Oracle, which we keep running on physical machines because Oracle).
-
And doesn't close, either. Sometimes the shit is so slow, I open the start menu, click Firefox, it doesn't react immediately and doesn't close, so I click it 3 more times. Then Firefox opens 4 times. This definitely used to be not this retarded back in the day...
x ∞
-
I don't think that deployment is easier, but central administration is, by large margin. Central administration is what companies want and Microsoft provides while Linux vendors are way behind with it (Canonical has something, but it's quite limited compared to the Windows tools).
Deployment and central administration go hand in hand. They are essentially one and the same under Windows.
If we setup a Windows machine and join it to AD then it pulls in the GP for whatever AD group we place it in and installs applications, printers and all of that for us automatically and usually just works. When a user signs in it pulls all of their GPs and does the same. I have forgotten all of the specifics of how we do it but we even have USB keys at client locations which they put into machines before first power on and it automatically joins them to the domain and wireless network and all of that.
You can do deployment without central administration, but you are going to be doing a hell of a lot of very repetitive work. As far as I am concerned they are essentially the same thing in any sane system that is designed to help system administrators.
-
@boomzilla said in WTF Bites:
Our customer has made us to a couple of studies on moving to the cloud and it's always been more expensive.
The biggest annoyance is when I mention the offsite backup/disaster recovery business. I always get asked if we run it on AWS or the like. It would be literally a couple of orders of magnitude more expensive to do so unless we made severe compromises in the service that we deliver. Sure, Glacier is cheap storage but I have to assume they named it what they did because of how occasionally chunks fall off of it and disappear.
We could move to AWS or something and still remain profitable at our current price levels if we wanted to not have to deal with managing our own machines but one of our big selling points is that for local clients we can offer very fast recovery times. We can deliver their data and bootable VHDs of their machines in not much more time than it takes to move the data to external hard drives. Internet bandwidth can't touch the speed of hard drives delivered via courier service, or myself if we are doing the recovery.
-
@Polygeekery said in WTF Bites:
I have yet to do a cost analysis on it where it is not significantly more expensive.
They only really start to work out cheaper when you also add in the amortized cost of the building itself, and then only if that's a dedicated building in the middle of a city (= high land prices) that you're comparing to, and all provided you can even do what you want with the service that is offered (if there's weird custom hardware involved you can just forget it). Not needing to build and operate a whole damn datacentre is quite a saving, and allows us to use what land we own for other things.
Can also be worth it if you want to have a geographically distributed service and yet don't already have service staff presence around the world, since that will let you save the cost of needing to start up new sites in completely new jurisdictions with all the fun with labour laws everywhere. (This isn't a concern we care about at all at WTF-U, but it would be for some other orgs when growing.)
You need to be pretty big to give a damn about any of these things.
-
Can also be worth it if you want to have a geographically distributed service and yet don't already have service staff presence around the world,
This and reliability are big things for us. We've gone from having regular downtime due to issues at the co-lo datacenter to rarely, if ever having any downtime and being able to be much more resilient and available worldwide without needing to maintain our own alternate datacenters and fail-overs.
-
@Benjamin-Hall said in WTF Bites:
Can also be worth it if you want to have a geographically distributed service and yet don't already have service staff presence around the world,
This and reliability are big things for us. We've gone from having regular downtime due to issues at the co-lo datacenter to rarely, if ever having any downtime and being able to be much more resilient and available worldwide without needing to maintain our own alternate datacenters and fail-overs.
Yeah, our customer has sites ~1,500 miles apart, so they have that covered already.
-
Not needing to build and operate a whole damn datacentre is quite a saving, and allows us to use what land we own for other things.
Colocation is a thing. Someone else builds a datacenter and you lease part of the space from them. We are in an awesome space and it is surprisingly affordable, at least in our area it is. YMMV, obviously.
Can also be worth it if you want to have a geographically distributed service and yet don't already have service staff presence around the world, since that will let you save the cost of needing to start up new sites in completely new jurisdictions with all the fun with labour laws everywhere.
Then we have another much smaller colocation lease in another area for clients that want or require geographic diversity. For that site we can usually get by with a remote hands service call if we need a hard drive swapped or something.
-
@Benjamin-Hall said in WTF Bites:
having regular downtime due to issues at the co-lo datacenter
It must have been a totally shit colocation facility. We have very literally never had that issue in......I don't know.....close to a decade? Literally the only downtime we have ever had that was their "fault" has been scheduled downtime for upgrades, etc. I would wager that it hasn't amounted to an hour total in the past five years.
-
@Polygeekery
At least in my company's case, we're realizing a good amount of at least short-term savings from the cloud allowing us to deploy resources "skinny" and be able to scale them later if necessary, rather than our old pattern of getting big iron with plenty of (read as: way too much) room for future expansion.Granted, we might have been able to save even more by doing the same evaluation of actual use case for right-sizing our hardware and accepting that future growth would mean deploying additional hardware, but it was definitely easier to latch onto the "cloud migration, ooooo" shiny and leverage that for the right-sizing. Plus now we can slice up the costs per department easier (not that we're back billing yet, but still makes it a lot easier to have conversations about who really needs what when the department sees the bottom line that's "their" fault).
-
@Polygeekery
At least in my company's case, we're realizing a good amount of at least short-term savings from the cloud allowing us to deploy resources "skinny" and be able to scale them later if necessary, rather than our old pattern of getting big iron with plenty of (read as: way too much) room for future expansion.Granted, we might have been able to save even more by doing the same evaluation of actual use case for right-sizing our hardware and accepting that future growth would mean deploying additional hardware, but it was definitely easier to latch onto the "cloud migration, ooooo" shiny and leverage that for the right-sizing. Plus now we can slice up the costs per department easier (not that we're back billing yet, but still makes it a lot easier to have conversations about who really needs what when the department sees the bottom line that's "their" fault).
It seems like virtualization could help with most or all of that. Split your stuff over VMs in a logical segregation pattern and then as you need more resources you can add servers and migrate VMs to rebalance your load over the cluster.
But to each their own. The use case where I see it always works for is if you just don't want to dick with managing your own stuff.
One amusing side effect of not using the
cloud is when there is big AWS or whatever outage and people ask me if we were affected by it. For all the resilience that AWS and Azure provide in services they themselves are also a single point of failure and they're always one controversial statement or policy away from getting everyone's shit DDOS'd.
-
Not needing to build and operate a whole damn datacentre is quite a saving, and allows us to use what land we own for other things.
Can also be worth it if you want to have a geographically distributed serviceWe already built the damn things, and we built them in different parts of the country.
I think I've mentioned it before and I'm not going into any real detail here but for some reason the grownups brought in some "CLouD aLL ThE ThInGS!!!" obsession that for most of what ends up there doesn't make any sense.It's not even faster to spin new servers up on EC2/Azure than on our own VMWare clusters because basically all of the lead time to spin up a server is internal process and not actually provisioning the servers, and there are additional steps in the process for connectivity when the server is being spun up in EC2/Azure.
-
@Polygeekery said in WTF Bites:
they're always one controversial statement or policy away from getting everyone's shit DDOS'd.
Eh, you're a pretty milquetoast kind of guy. I wouldn't worry about that.
-
-
Stockholm syndrome, I think.
-
@boomzilla said in WTF Bites:
@Polygeekery said in WTF Bites:
they're always one controversial statement or policy away from getting everyone's shit DDOS'd.
Eh, you're a pretty milquetoast kind of guy. I wouldn't worry about that.
I know you were being sarcastic, but I am a low profile target. I am not so much worried about what I say or do. I'm worried about what others do. Like Amazon or MS or even some other asshat that I might be unfortunate enough to be on the same set of resources with.
One of our clients was nearly dead in the water a couple of years ago because Amazon did....something, and the LOB they use was down for many hours. Or maybe they just had a bunch of their equipment shit the bed at once and I am confusing that client outage with another client outage?
-
@Polygeekery said in WTF Bites:
Or maybe they just had a bunch of their equipment shit the bed at once
Wasn't it that BGP configuration thing?
-
@Tsaukpaetra said in WTF Bites:
@Polygeekery said in WTF Bites:
Or maybe they just had a bunch of their equipment shit the bed at once
Wasn't it that BGP configuration thing?
I think that was Farcebook, but
-
@Polygeekery said in WTF Bites:
I don't think that deployment is easier, but central administration is, by large margin. Central administration is what companies want and Microsoft provides while Linux vendors are way behind with it (Canonical has something, but it's quite limited compared to the Windows tools).
Deployment and central administration go hand in hand. They are essentially one and the same under Windows.
If we setup a Windows machine and join it to AD then it pulls in the GP for whatever AD group we place it in and installs applications, printers and all of that for us automatically and usually just works. When a user signs in it pulls all of their GPs and does the same. I have forgotten all of the specifics of how we do it but we even have USB keys at client locations which they put into machines before first power on and it automatically joins them to the domain and wireless network and all of that.
You can do deployment without central administration, but you are going to be doing a hell of a lot of very repetitive work. As far as I am concerned they are essentially the same thing in any sane system that is designed to help system administrators.
With Linux you can also have a pre-configured install image that will include setup for networks and add it to AD and whatever management system, but then you need the management system, something like puppet or chef – or Ubuntu Landscape – to manage the user- and group-dependent configuration and that is nowhere near as streamlined as the Windows group policies.
-
@Polygeekery said in WTF Bites:
And don't get me started on what pile of manure is Azure.
I fucking hate Azure and everything that goes along with it. I especially hate the push to Azure AD. Fuck right off. For almost all of our clients it is a shitty fit with ongoing costs and they can go fuck their hats.
Azure AD is the part I have least problem with. That part is kinda stable and authentication works mostly fine for us. But the other parts tend to get a lot of pointless
innovation and then either break because they don't get properly tested, or just don't work the way one would expect because proper analysis and design are so last millennium.The push to Azure and AWS
cloud computing also annoys me. I have yet to do a cost analysis on it where it is not significantly more expensive. I am certain that there are use cases where it makes sense, but none of them are ones which we work with.When you include the effort of the staff building and setting it all up, it's not that big, but of course it's more expensive. There is some economy of scale, but there is also a lot of extra effort going into building all the support infrastructure. The main benefit is that you can get the resources set up quickly. And you can scale them on demand, though it's a lot harder in practice than the marketing makes it sound.
-
@HardwareGeek said in WTF Bites:
@boomzilla said in WTF Bites:
Oracle, which we keep running
Oh yes, Oracle... I remember. Long long long ago
, 
decided to offer our product to a customer who insisted on Oracle. Our application worked with Microsoft Access or Microsoft SQLServer only then. So I had to go thru all SQL statements and clean them up - virtually all were created by string concatenation, get that to parameterized queries. And then there were some SQL dialect differences which required another load of work. Etc. Anyway, with the cleaner database access, the work was worth it, even when then did not win the contract with that customer.
Years later, a different customer who had actually hardly any reasonable amount of data (Access would already look like overkill for their purposes), insisted on Oracle. Well, they got it, as requested.
-
-
Today's
is VmWare.I asked for some space on our VmWare vSphere, and got it. So I started pushing the .iso images I want to start from, and I got an error. A generic error saying just “The operation failed”. But looking into the devtools (when using the web client), the important PUT request simply returned 404.
It did not work for me in Edge, it did not work for me in govc and it did not work for me in terraform vsphere (both saying
404 Not Found, but I couldn't get the actual request that returned it from either). But it did work in Chrome for the admin.In edge there was reference to a ‘KB’ article saying I should add certificates for the hypervisor (the vcenter,
vctr.dev.company.com, has certificate signed by the corporate root, but the hypervisors themselves,hyp2.dev.company.com, have self-signed ones), but that made no difference. Also, I had certificate verification simply turned off in the command-line tools.Well, it turned out that:
- When trying to upload a file, the vSphere first asks some endpoint for URL where to PUT it, and then PUTs it there.
- When that endpoint is called from Chrome, it responds with URL to the hypervisor itself (
https://hyp-2.dev.company.com/folder/path/to/the-file?dcPath=DC-A&dsName=STORE), but when called from anywhere else, the same endpoint returns URL to the vCenter front (https://vctr.dev.company.com/folder/path/to/the-file?dcPath=DC-A&dsName=STORE). - The former worked fine—and is why you need those certificates—but the later returned 404.
- Under admin account the later also worked though!
- So the admin filddled with the permissions and adding “low level datastore operations” fixed the problem!
So
1
VmWare vCenter does something different based on user agent (because the query for upload endpoint was otherwise the same between the browsers).
2 Permissions are checked if one way gets chosen, but not if the other does.
3 The way that checks the permission returns 404 not found instead of, perhaps, something about permissions (that would be 403 forbidden).
4 The permission has a name that says nothing. Or would you guess that “low-level datastore operation” is actually upload files? The fact there is a delete file permission, but no create or upload file permission might be a hint, but not very strong one.
-
@Polygeekery said in WTF Bites:
Windows just becomes more and more shit every single day. How bad could it get before they actually started bleeding market share?
Losing market share to ..... who?
If they did, I cannot even think of a viable alternative.
And that's the problem. No matter how shitty Windows becomes, there is no viable alternative. If there was, you would already be using it.
-
@Polygeekery said in WTF Bites:
I have yet to do a cost analysis on it where it is not significantly more expensive.
They only really start to work out cheaper when you also add in the amortized cost of the building itself, and then only if that's a dedicated building in the middle of a city (= high land prices) that you're comparing to, and all provided you can even do what you want with the service that is offered (if there's weird custom hardware involved you can just forget it). Not needing to build and operate a whole damn datacentre is quite a saving, and allows us to use what land we own for other things.
Can also be worth it if you want to have a geographically distributed service and yet don't already have service staff presence around the world, since that will let you save the cost of needing to start up new sites in completely new jurisdictions with all the fun with labour laws everywhere. (This isn't a concern we care about at all at WTF-U, but it would be for some other orgs when growing.)
You need to be pretty big to give a damn about any of these things.
Their ideal scenario is something very bursty, like if you have a Mother's Day card business so you don't want to build a whole data centre for it, keep it powered, employ a full staff for security, maintenance, custodial work, etc year-round. Most of the time you could have it running on a couple small instances, but when it got busy you could spin up a gajillion VMs for a day
-
With Linux you can also have a pre-configured install image that will include setup for networks and add it to AD and whatever management system
You can do that under Windows also. But why would you? That is a shitty way to do things. Under Windows the image and the deployment configuration are entirely independent. You can drop one file into your install media and be done with it.
-
@Gern_Blaanston said in WTF Bites:
@Polygeekery said in WTF Bites:
Windows just becomes more and more shit every single day. How bad could it get before they actually started bleeding market share?
Losing market share to ..... who?
If they did, I cannot even think of a viable alternative.
And that's the problem. No matter how shitty Windows becomes, there is no viable alternative. If there was, you would already be using it.
I'm oddly optimistic about steam os. Not that I have faith in Valve or Linux in general but because gaming is the only real use case I have for windows besides the surface I bought. If Valve can lure gamers away they might make a dent.
I suspect stuff like Samsung Dex has a better chance though. At the very least we might start getting better usbc dongles.
-
@Polygeekery You just drop one file into the Linux install image as well. It needs the information on how to connect to the systems that will give it the rest of the configuration, which is like the Windows one containing the information on how to connect to the domain. It's the systems that provide the further configuration that are much harder to set up (and are more fragile).
-
I suspect stuff like Samsung Dex has a better chance though. At the very least we might start getting better usbc dongles.
I hadn't heard about that yet so I go to check it out.
Bottom of the page:
There's still a lot of work to be done in regards to vendor lock-in, DRM, stupid ass policies bullshit.
-
@Gern_Blaanston said in WTF Bites:
@Polygeekery said in WTF Bites:
Windows just becomes more and more shit every single day. How bad could it get before they actually started bleeding market share?
Losing market share to ..... who?
If they did, I cannot even think of a viable alternative.
And that's the problem. No matter how shitty Windows becomes, there is no viable alternative. If there was, you would already be using it.
I'm oddly optimistic about steam os. Not that I have faith in Valve or Linux in general but because gaming is the only real use case I have for windows besides the surface I bought. If Valve can lure gamers away they might make a dent.
I suspect stuff like Samsung Dex has a better chance though. At the very least we might start getting better usbc dongles.
There's convergence on pretty much all the mobile platforms, or it's being worked on.
When Linux mobile is ready for use, I'll go for that as cell phone os, but that is years away by how it's looking right now.
-
When Linux mobile is ready for use, I'll go for that as cell phone os, but that is years away by how it's looking right now.
Year of the Linux Mobile presumably a little while after Year of the Linux Desktop!
-
@Polygeekery said in WTF Bites:
I have yet to do a cost analysis on it where it is not significantly more expensive.
They only really start to work out cheaper when you also add in the amortized cost of the building itself, and then only if that's a dedicated building in the middle of a city (= high land prices) that you're comparing to, and all provided you can even do what you want with the service that is offered (if there's weird custom hardware involved you can just forget it). Not needing to build and operate a whole damn datacentre is quite a saving, and allows us to use what land we own for other things.
Can also be worth it if you want to have a geographically distributed service and yet don't already have service staff presence around the world, since that will let you save the cost of needing to start up new sites in completely new jurisdictions with all the fun with labour laws everywhere. (This isn't a concern we care about at all at WTF-U, but it would be for some other orgs when growing.)
You need to be pretty big to give a damn about any of these things.
Their ideal scenario is something very bursty, like if you have a Mother's Day card business so you don't want to build a whole data centre for it, keep it powered, employ a full staff for security, maintenance, custodial work, etc year-round. Most of the time you could have it running on a couple small instances, but when it got busy you could spin up a gajillion VMs for a day
I know that my company has a standing contract with AWS so that developers can fairly easily set up servers to...do stuff. We don't use it because our customer finances hardware that we have in a company run datacenter, but I've worked with other people in the company who use AWS for stuff like that.
Not exactly "bursty" like you describe, but always changing for sure and not always easily predictable.
-
@loopback0 said in WTF Bites:
When Linux mobile is ready for use, I'll go for that as cell phone os, but that is years away by how it's looking right now.
Year of the Linux Mobile presumably a little while after Year of the Linux Desktop!
In that case, it'd already be. Instead it's quite a mess.
-
@Polygeekery said in WTF Bites:
I suspect stuff like Samsung Dex has a better chance though. At the very least we might start getting better usbc dongles.
I hadn't heard about that yet so I go to check it out.
Bottom of the page:
There's still a lot of work to be done in regards to vendor lock-in, DRM, stupid ass policies bullshit.
Yeah, it's a shitshow at the moment but everyone has a phone and most carry a bag. A trendy looking dongle and an okayish android on the desktop and most people would be set. Better chance than the year of the Linux desktop.
Surprised there isn't something like this for iphone to be honest.
-
Why is there a climate change banner on a tech video about computers?
-
@loopback0 said in WTF Bites:
Year of the Linux Mobile presumably a little while after Year of the Linux Desktop!
Which will come right after fusion power plants here on earth.
-
Surprised there isn't something like this for iphone to be honest.
I could have sworn that there was one, but I am either mistaken or that is another product that they used to have and then quietly killed and memory holed.
-
@Polygeekery said in WTF Bites:
@loopback0 said in WTF Bites:
Year of the Linux Mobile presumably a little while after Year of the Linux Desktop!
Which will come right after fusion power plants here on earth.
Legends, scattered scraps from the Pnakotic Manuscripts mostly, say that the An-droid platform is a brain-damaged bondaged-up Linux platform, but who listens to the mad ravings of wind and ghosts.
-
@Polygeekery said in WTF Bites:
@loopback0 said in WTF Bites:
Year of the Linux Mobile presumably a little while after Year of the Linux Desktop!
Which will come right after fusion power plants here on earth.
Legends, scattered scraps from the Pnakotic Manuscripts mostly, say that the An-droid platform is a brain-damaged bondaged-up Linux platform, but who listens to the mad ravings of wind and ghosts.
Fair. But currently the market is being fractionalized between vanilla Android (which I believe is only on Pixels?), Samsung's shitshow of their own ecosystem that is Android-ish, and all of the rest that do not allow you full control of your device.
-
@Polygeekery said in WTF Bites:
Samsung's shitshow of their own ecosystem that is Android-ish
It's Enlightened.
-
@Polygeekery said in WTF Bites:
Surprised there isn't something like this for iphone to be honest.
I could have sworn that there was one, but I am either mistaken or that is another product that they used to have and then quietly killed and memory holed.
You can plug an iPhone or older iPad into a display with the Lightning to HDMI dongle and even use a mouse and keyboard but it's just iOS on a bigger display rather than a desktop experience.
ipadOS 16 has a more desktop-like experience if you plug an M1 iPad into an external display which is hardly surprising considering at this point iPads are basically Macbooks without keyboards.
-
@Polygeekery said in WTF Bites:
@Polygeekery said in WTF Bites:
@loopback0 said in WTF Bites:
Year of the Linux Mobile presumably a little while after Year of the Linux Desktop!
Which will come right after fusion power plants here on earth.
Legends, scattered scraps from the Pnakotic Manuscripts mostly, say that the An-droid platform is a brain-damaged bondaged-up Linux platform, but who listens to the mad ravings of wind and ghosts.
Fair. But currently the market is being fractionalized between vanilla Android (which I believe is only on Pixels?), Samsung's shitshow of their own ecosystem that is Android-ish, and all of the rest that do not allow you full control of your device.
Google is working on convergence on Android, it's just not ready yet, so you have to go out of your way to test it.
-
@Polygeekery said in WTF Bites:
@Polygeekery said in WTF Bites:
@loopback0 said in WTF Bites:
Year of the Linux Mobile presumably a little while after Year of the Linux Desktop!
Which will come right after fusion power plants here on earth.
Legends, scattered scraps from the Pnakotic Manuscripts mostly, say that the An-droid platform is a brain-damaged bondaged-up Linux platform, but who listens to the mad ravings of wind and ghosts.
Fair. But currently the market is being fractionalized between vanilla Android (which I believe is only on Pixels?),
Pixels have Pixel Android, with different programs and features in those programs (some controlled server-side) than are available in AOSP or for other phones using the Google Play Store. Pixels probably have less "bloat" than many others, though.
An AOSP ROM + the Play Store would be the closest to vanilla that is still useful given Google shoving so much functionality into Google Play Services and similar.
-
@Polygeekery said in WTF Bites:
@Polygeekery said in WTF Bites:
@loopback0 said in WTF Bites:
Year of the Linux Mobile presumably a little while after Year of the Linux Desktop!
Which will come right after fusion power plants here on earth.
Legends, scattered scraps from the Pnakotic Manuscripts mostly, say that the An-droid platform is a brain-damaged bondaged-up Linux platform, but who listens to the mad ravings of wind and ghosts.
Fair. But currently the market is being fractionalized between vanilla Android (which I believe is only on Pixels?), Samsung's shitshow of their own ecosystem that is Android-ish, and all of the rest that do not allow you full control of your device.
Google is working on convergence on Android, it's just not ready yet, so you have to go out of your way to test it.
It's convergence to a walled garden though.
-
@Polygeekery said in WTF Bites:
@Polygeekery said in WTF Bites:
@loopback0 said in WTF Bites:
Year of the Linux Mobile presumably a little while after Year of the Linux Desktop!
Which will come right after fusion power plants here on earth.
Legends, scattered scraps from the Pnakotic Manuscripts mostly, say that the An-droid platform is a brain-damaged bondaged-up Linux platform, but who listens to the mad ravings of wind and ghosts.
Fair. But currently the market is being fractionalized between vanilla Android (which I believe is only on Pixels?), Samsung's shitshow of their own ecosystem that is Android-ish, and all of the rest that do not allow you full control of your device.
Google is working on convergence on Android, it's just not ready yet, so you have to go out of your way to test it.
It's convergence to a walled garden though.
That's the way everyone does it these days. Which is why I'm looking towards mobile linux. The desktop on phone is actually working there. The problem is that the phone part of the phone is utter shit.
-
This post is deleted!
-
The problem is that the phone part of the phone is utter shit.
I suspect it's because the phone-related hardware has bajillion of versions, each with its own quirks that the manufacturers are reluctant to document, because they are happy with users using their version of the operating system with the apps they were paid to preinstall.
That makes it harder to port the open system to more phones, which in turn limits the number of people who have all of the skill, the desire to work on it, and suitable hardware at the same time.
