Its a ton heavier, because you need a full LibreOffice install effectively, but it does a lot more than Apache POI, so it depends on whether your use-case warrants the extra features.
Companies like Collabora (for general linux stuff) and Igalia (for browser type stuff) exist to do things like this.
They employee a bunch of people who work on open-source stuff, and you can pay that company to get stuff done on various open-source projects.
The stumbling block is that the cost is almost always much much more than you would be willing to pay, because, surprise surprise, our line of work is very very expensive.
AMD ran into significant compatibility issues with AM4, with some motherboards not being able to supply the amount of power needed by the newer CPUs and PCI-E Gen 4 support being removed in the final BIOS release due to reliability issues. A lot of motherboards also didn't have enough flash space to contain the firmware needed for the whole CPU range, so the update to support the newer gen had to remove support for the oldest gen.
Turns out it's really hard to guarantee compatibility with several years of third-party products which weren't designed with your new fancy features in mind.
Newer doesn't usually mean more power, though. You can just have a power limit, it's fine. And I don't expect PCIe to get faster with a CPU upgrade anyway.
The flash space was pretty forseeable and they dealt with it, more of an excuse than anything.
I'd bet customer perception is also a factor; there's a risk the old boards (not even made by Intel) die and cause problems, and Intel wouldn't want to deal with press/comments like "This CPU stopped working after 2 months" or "I installed this new CPU, and within 2 months it killed my motherboard that had been working fine for 7 years".
They've released CPU upgrades with the same socket before, I'm sure they have the sales data to know how that performs vs new socket.
Laptops have outsold desktops for well over a decade, and their CPUs pretty much non-upgradable. I can't easily find a nice chart to reference, but intuition tells me the desktop industry is similarly trending towards complete "system" sales vs individual parts. In other words: Most people don't upgrade their CPU, they upgrade by replacing the entire system. If true, this also means the socket would be almost entirely irrelevant to sales performance.
Erlang is being sneaky because an Erlang "process" is not the same thing as what the rest of call us a "process".
As soon as Erlang needs to talk across real processes is runs into all the same problems.
Now, admittedly, Erlang has a nice supervisor toolkit to help with this, but its still a real problem, even for Erlang.
Okay but then you're talking about designing a completely new kind of kernel which has a completely different concept of what a "process" is, which is quite far away from "we should've just used SmallTalk".
Besides, no matter how you design the system, there will be a categorical difference between communication which is expected to be fallible and communication which is expected to be infallible. The "communication" between a caller and a callee is expected to be infallible, for example. Communication between threads in a process (in the UNIX and Windows model) is expected to be infallible.
Communication between an app and a system service is expected to be fallible; at the very least, the system service has to expect that the app might misbehave in any number of ways, even if you argue that the app should expect the system service to be perfect.
OOOh, very very nice. I might have to copy this into the H2 database engine. We already use variable-size encoding for numbers, and this would make a nice addition to reducing the size of row data containing text.
> Sloppy apps ignore this, but the devs are quickly notified.
Anything that increases friction for developers is bad. The API for HiDPI should be so seamless that the only thing developer does is to provide higher resolution bitmap icons.
Imagine what web app developers need to do when their users switch from a regular display to a HiDPI display. They do nothing; the browser knows how to scale on its down. And that should be the bar we are aiming for.
reply