What's the point of recompiling using 9 year old source, vs. using a 9 year old driver?
So let's go through all the issues with what you are saying.
1) Doesn't have to be "9 year old source".
As addressed in the Kernel doc I am posting for the 3rd time, if the subsystem changes, the driver should ideally change - and adapt to the improved subsystem, instead of keep using your 2007-vintage driver. Note that this applies the same to Linux, to Windows and to macos.
2) And I say "should" because DKMS is there to help backporting to older kernels. But of course, for that there must be something to be backported.
3) Note, if the driver had been sent upstream, the updating would be done automatically by the kernel maintainers. It's the sane thing to do,
and the kernel maintainers ask companies to do it, FFS. The case of a company publishing sources for the drives but failing to send them upstream is most probably myopic laziness, which of course means problems for the users. Contrast: non-myopic, sane laziness would be to just let the kernel maintainers do the work for them,
for free!
4) So we have the case of a lazy, or no-longer-existing company, which hasn't updated drivers in 9 years. This means that probably the moment that something changes in Windows, you're hosed. But in the RocketRaid case you yourself posted, there are instructions to recompile your drive from the original kernel 2.x release, to 3.x, to 4.x - meaning, CURRENT. So, in the same way, you can expect that the hardware will keep working for as long as there is some interest in it - not only the manufacturer. And if bad comes to worse, you yourself could do it.
5) Particularly, look at the changes mentioned in the RocketRaid case. The changes themselves are rather small - meaning, probably doable just by checking what function renames happened between kernel versions. BUT the effect can be huge; look at how one of the changes is from hpt_queuecommand() to hpt_queuecommand_lck(). That sounds like an important change, which will probably be significant for performance and/or reliability. And hey, now your driver is adapted to it! Instead, if it was the Windows case and such a new Kernel API appeared, the Kernel would anyway keep the old ABI compatibility and you would stay using the outdated functionality. But hey, at least it's convenient! Right? No matter if this means 2 conflicting APIs, which typically means that the newer one is crippled by having to coexist with the old one.
Cruft for everyone, even people who avoid 9-year-old hardware.
Those improvements occurred, and they're opaque to the driver. You don't need to recompile to get the improvements.
And how do you know, in a closed-source kernel?
Well, at least we do know that the driver itself keeps working as in 2007
.
Here are the Ubuntu instructions for the RR62x controller -
https://help.ubuntu.com/community/RocketRaid .
By the fifth or sixth screen it should be obvious why many software applications have never been ported to Linux.
Do you realize that that page shows you how to adapt
2 different drivers (not a mere "software application") for
3 different major versions of the kernel? Something that is plain
impossible in Windows?
... are you comparing that to a driver that didn't get updated even by its creators in 9 years?
As already said - pseudo-convenience over reading and having the most up-to-date system. Seems like you made your choice
.
[doublepost=1479809191][/doublepost]
Are you sure?? Maybe you're just not noticing it, because I've experienced this issue with ALL of my Win 10 computers. Let me explain a bit more in-depth: It's how the scrolling FEELS that's inconsistent, not how it looks. Edge has silky smooth scrolling, very similar to macOS. Chrome's is okay, but it's not as smooth. The right side of Explorer (not IE) is really jaggedy when you scroll, and the left side of Explorer is really SLOW (& kind of jaggedy too). The inconsistency is INCREDIBLY annoying. If all the scrolling just felt like Edge's, I'd be super happy. Edge's scrolling is fantastic. (Maybe because it's designed for touch input?)
It's getting funnier! Turns out that Edge scrolls sideways well (if we accept "jerkily in a mushy way" as "well") - it's just that other applications, like Microsoft Help Viewer, or Microsoft Explorer, just don't accept the sideways scrolling from the trackpad.
And meanwhile, Microsoft Visual Studio Express 2015 scrolls vertically kinda acceptably in the editor pane, but the Output pane scrolls vertically in a jumpy way, and horizontally in a heavy, pixel-by-pixel way.
And it's all Microsoft applications in a Microsoft OS!!
BUT WE ALL LOVE DIVERSITY, RIGHT?
