'Linux'.. WOW, I'm almost toltally SOLD now!!!

[PKTKS]

wayland

Wayland ... I think I have to make my X11-terminal project compliant with Wayland.
At the moment it's a nano-X compliant.

What is nano-X? It's a directive used in xorg to compile it with the minimal features.

It will be a challenge

Well..  I am really unsure about how things will be in a couple of years..

But..  following the money .. they are diverting developers (paying them)
to steer the whole "desktop" base of all *NIX (read Xorg drivers) to that
"thing" ..

Putting some pieces together it looks like demolishing X Windows...
which is impossible to overtake with these single display license model..
(remember X Windows is fully network aware..)

It looks like all points to X Windows implosion by removing developers (paying them)
to steer the whole "desktop thing" to run only under that premises..

Without support from NV and AMD it is impossible to have proper displays
working .. easy to demolish XWindows (call it old, bloated, etc.. etc.. remove
people from that and let it just  fade away. like nobody wants it works on it..)

So MS will CLAIM PROPERTY OF YOUR/OUR HARDWARE and decide
what can be run, see, listen..  and obvious..

Only the shell companies (partners) will be allowed to run under WSL..

SO.. conclusion:  If investing time, resources, hardware...

Remember that  the whole thing has being has been(is being) taken by assault..
subverted and perhaps (if no one cares) we will  be forced to pay MS licenses to
have a working proprietary video driver... (implosion of Xorg)..

And wayland + systemd  shit working under WSL..

MS will again claim property of all hardware by just shitting  Xorg dead ended

speculation based on facts.

they will doom  CYGWIN
they will doom WINE
they WILL DOOM Xorg

Paul

[DiTBho]

I am investing a part of my spare time in haiku-os. The project is stalled at the moment, but there is a beta CD for x86 systems, and who knows? If anyone puts a lot of money into this project ...

[PKTKS]

With these BIGGGG  pockets of RH and Canonical and IBM
diverting competent developers (read bribing) from key
driver devel?

NV is giving a shit to anything besides their closed stuff
under MS and their own pocket in ARM...

The options shrunk  to less than a hand

Paul

[madires]

Proprietary drivers for graphics cards have another common issue: licensed technology (patents). So the manufacturer can't "open source" the driver because of licensing agreements. It's a wicked situation. For the last decade AMD's Radeon is the least evil and the X11 drivers work generally well. Intel's integrated graphics are also supported in case you're happy with basic 3D performance. But always check if a specific chip is supported by X11 before buying. Otherwise you'll have to use the generic VESA driver.

[PKTKS]

I can not speculate over "closed" stuff never seen.. actually

But given the AMD status on opening some code base..
i guess it must be more FUD than really an issue.

They can open the ting but probably trading a bit
chunk of their extortion algo. deprecating chips and
boards every 15 days

True as well we can run VESA (still INT10H ?) but since
some versions of Xorg drivers I have far better results
running FB when the "official" chip driver has critical issues

FB works marvelous and better than VESA.. although
acceleration is crappy and some apps will just argue about OGL.

BOTTOM LINE:  Wo official and OPEN vendor support there is
no way "THE DESKTOP" metaphor will be smooth on *NIX.

SO  far SO good it works - as you said - by carefully chosen hardware.

AND WE DID EVEN TOUCH THE BIG ISSUE OF SECURE BOOT
(aka paying keys to MS to boot with that jerk asshole shit)

Another post..
Paul

[DiTBho]

Proprietary drivers for graphics cards have another common issue: licensed technology (patents). So the manufacturer can't "open source" the driver because of licensing agreements. It's a wicked situation.

Perhaps open hardware video cards? I am surprised a couple of similar projects just died five years ago

Do you happen to know a PCI dual head VGA video card with open source driver for X11?
I don't need MESA, I need a fast 2D hardware acceleration and Xinerama support.

Just to consider an alternative to my wild hack to the G450, which works, but it's not as fast as it should be 

[madires]

I'd suggest to check out some older Radeons.

[PKTKS]

I'd suggest to check out some older Radeons.

hmm even older RADEONS still require the gallium compiled stuff
by LLVM - you will need things like DRI and OGL support

You must go BEFORE that to avoid such problems..

Mach64 or Rage3d boards...  or the excellent line of..
[SiS] 661/741/760/761 PCI/AGP  a.k.a. = SiS 671/662 bridges

Paul

[PKTKS]

AND WE DID EVEN TOUCH THE BIG ISSUE OF SECURE BOOT
(aka paying keys to MS to boot with that jerk asshole shit)

MS is being very kind on this case. They allowed user generated keys (MOKs) for Linux, and that's remarkable.

Being old MS, they would have asked an extortionate amount of money per version per distribution, and only on good day they may potentially approve that.

As far as I am aware of that they  allowed *ONLY* their shell partners
do have such keys.. e.g. Canonical RedHat and *may be* some other
player in direct interest for their strategy..

No surprises on that "technology" emerged from them.

Secure... for whom ?

Paul

[Monkeh]

I'd suggest to check out some older Radeons.

hmm even older RADEONS still require the gallium compiled stuff
by LLVM - you will need things like DRI and OGL support

You must go BEFORE that to avoid such problems..

Mach64 or Rage3d boards...  or the excellent line of..
[SiS] 661/741/760/761 PCI/AGP  a.k.a. = SiS 671/662 bridges

Paul

Yes, let us remain in the 1990s. Good plan.

Seriously, stop with your baseless conspiracy rants. It's getting old.

[PKTKS]

No rants..  no conspiracy.. business as usual.

Alas very same it was on the 90s.

For the boards I am a collector.

And i have these suggested (SiS  Mach64 and a Rage3D AIW)
plus some very nice S3 Diamond Stealth..  some Rage AGPs..

And some of  these crappy totally closed shit RADEONS and NVs GPUs.
Which took me very hard time to proper setup using these LLVM
compilers and closed source instruction sets... no fun no joy
and overpriced...

SiS as suggested has one of the *BEST*  clear crystal images
with an awesome X driver suuport. It works dual headed or
multiheaded without a single issue and a nice control panel..

Rage3D also works fine (including DVB) as long patched with
proper midia (they are xorg based drivers)

They still work.. last time I checked.

I have fun collecting them..

Paul 

PS.  Unfortunately my old Trident boards are not booting ..
 i have half dozen Tridents but just one ISA capable MOBO..
and the boot kernel is really not fitting in that FPM memory

[coromonadalix]

Anyone tried Q4OS  ?

[mikerj]

But how is this  a linux problem?more like  lazy hardware manufacturers who cant be arsed to provide drivers,If you bought  new hardware  for  a windows machine that didn't run would you be shouting at microsoft?

Ultimately it's the user's problem if they can't use their device properly, whatever the underlying reason is. Blaming manufacturers for for not providing drivers does not fix the issue or make Linux look more attractive to people used to having stuff just work (or at least mostly work) under mainstream operating systems.

[PKTKS]

But how is this  a linux problem?more like  lazy hardware manufacturers who cant be arsed to provide drivers,If you bought  new hardware  for  a windows machine that didn't run would you be shouting at microsoft?

Ultimately it's the user's problem if they can't use their device properly, whatever the underlying reason is. Blaming manufacturers for for not providing drivers does not fix the issue or make Linux look more attractive to people used to having stuff just work (or at least mostly work) under mainstream operating systems.

I don't think so.. (it is the user solely problem).

Actually we have paid for a product...
and having a DYSFUNCTIONAL product is not a good point.

Whatever reason may be it boils down to the fact that due
to some very basic "bad practices" we are collecting such SHIT
products across DECADES..

Whoever thinks this is any sort of new thing have not lived
the bad days of 90s across 2000s  where the fast pace of
deprecating motherboards and cards caused a nasty prejudice
for *ALOT* of people while filling MS and half dozen shell partners.

Due to ignorance most manufactures do not have a clue how
to actually "support"  their devices on the market by providing
users with the proper tools to adapt the product to the real world..

This ignorance driven by a simple choice of following what is
mostly a lie that  MS is some sort of "prevalent desktop" or computer..
will then free them to support the wider and better real computer world.

Result is that all folks old enough probably have a *BIG PILE* of
DYSFUNCTIONAL hardware due to unsupported or deliberate crippled software.

from 2000s on to 2010s things got better just because a large
user and developer base of *NIX folks play a role..

THE O.P.  HAS VERY GOOD REASONS ASKING FOR DRIVER DIRECTIONS

but the bottom line is that under a VM it will never ever works as it should..
And proper decisions should be made when chosen the OS and the products.


Paul

 

[DiTBho]

This ignorance driven by a simple choice of following what is mostly a lie that  MS is some sort of "prevalent desktop" or computer.. will then free them to support the wider and better real computer world.

I don't think this was ignorance, but rather marketing!

Linux has always sucked as a desktop probably because there is no centralized way to connect drivers with hardware; there is no Apple, no Microsoft, and years ago brands like IBM and HP ran their business with their proprietary Unix solutions, leaving the rest of the PC uses in a chaotic scenario where there was nothing supported, documented except that done by volunteers.

Linux is not Unix, do you remember the most famous meme? Well, Linus said it because he didn't like those Unix solutions (also because too bloody expensive).

Linux is done by hackers for hackers, and and recently it has somehow captured the interest of the mass!

It has evolved and it's keep getting better, but for sure you cannot put Linux kernel v1.0...v2.6 and their related userlands in the same basket of commercial holy Unix monsters like { IRIX, HPUX, AIX, ... }.

Back in 2000 I was a BeOS user, and I hacked every single version of BeOS, from 4.5 to 5.03-Netbone, and also ported a lot of text-based applications to it.

X11 apps were not trivial to be ported because BeOS has no X11 support and the graphis engine is completely different.

BeOS was great on its BeBox hardware produced and sold by the same Be.inc company. They were somehow a kind of Apple in miniature, uncommon PowerPC hardware (dual PPC60x processing!!!) easy to use and very intuitive software, they initially developed both hardware and  drivers, and this was it was "plug and play".

Then they had to face it out: they spent a lot of money and they weren't earning a penny becase the marketing was dominated by x86 PeeeeCeeee, so they ported BeOS to this architecture thinking they could have benefit more available hardware.

Result?

I spent months and months trying to adapt Linux drivers for PC's common hardware to BeOS/x86, and found the Linux code is not exactly friendly to be understood and recycled, and worse still, the most of Linux developer are not willing to help unless you pay them.

Which brings us to the question: how money and resources are allocated?

If I have to pay a Linux developer to help me to port a driver to BeOS, why cannot directly pay whom did the hardware to have a fresh driver rather than some adapted code?

That's what Apple and Microsoft do, but usually not Linux, unless it's something enterprise, for example Darpa and Cern pay IBM, who pays whoever is involved with POWER9-POWER10-approved hardware to have their drivers written as they need to be!

So it's like the BeBox, but on a larger money scale!

[Marco]

Linux has always sucked as a desktop probably because there is no centralized way to connect drivers with hardware

There's Chromebooks.

The Chromebook hardware model is a very nice intermediary between Windows PC's chaos and Apple's closed garden. As open source as possible, but using trademark and certification to keep configurations of Chromebooks tightly controlled so Google can do good QA.

If only Google had a different revenue stream than datamining people and served more hardware intensive markets. I hoped that Valve would do something similar to Chromebooks, but without datamining the users ... but alas.

[rsjsouza]

Anyone tried Q4OS  ?

I am using it. I posted about it in the first page

I am typing this in a former Chromebook (Acer C720) that is quite the specialized/non-standard machine, but it is old enough that all its kinks were solved in the process of removing the former Chrome OS (not supported anymore by Google) and replacing it with a Q4OS (a lightweight Debian). All this is running on its original 16GB of SSD and 2GB of RAM.

[madires]

Neither is the Windows ecosystem a driver heaven. If you don't get a Win10 driver for your old scanner it becomes e-junk also, for example. There are expensive special purpose devices which run only with XP/Win7 or are certified in conjunction with with XP/Win7 (like medical devices), while vendors don't provide Win10 drivers or don't wan't to pay large sums for re-certification of their systems for each new Win10 release.

[bd139]

Yep. There's a driver bug with the Intel graphics in my desktop which means every time it wakes up from sleep it resizes all the windows. But that's the only thing it does which is annoying (so far)

[DiTBho]

You are lucky. There are no bugs in my Apple Powerbook G4 with macOSX v10.4, but with Linux every time the laptop wakes up from sleep something somewhere inside the graphics chip starts to go wrong so much that the screen color is like the TV tuned to a dead channel and the snow falls like pixels of a random color.

There are no solutions, you have to force the laptop to go back to sleep again and pray that waking up will bring it back to working order, kind of leap into the dark you have no guarantees and no care, you don't even see what's on the screen. Which is so annoying that I removed all the backcalls in the kernel in order to remove the sleep/wake up functionality.

feature-- (because bugged). Problem solved 

[madires]

There are several kernel parameters to modify sleep mode settings, just saying.

[bd139]

No one wants to screw with that though

[madires]

Only for the brave?

PS: The kernel parameters are documented.

[rsjsouza]

Only for the brave?

PS: The kernel parameters are documented.

Yep, and that is one of the reasons why Linux can be extremely unpalatable for the general public. Computer goes to standby/sleep, never recovers, the person loses unsaved data. No worries, let's muck around with parameters in a very sensitive area that, if you do a typo or delete too much, your system may become unbootable.

For that you have the immense power of M$'s popularity: things get fixed much quicker and these "details" (which are very critical IMO) are left behind in Linux.

[PKTKS]

Only for the brave?

PS: The kernel parameters are documented.

Yep, and that is one of the reasons why Linux can be extremely unpalatable for the general public. Computer goes to standby/sleep, never recovers, the person loses unsaved data. No worries, let's muck around with parameters in a very sensitive area that, if you do a typo or delete too much, your system may become unbootable.

For that you have the immense power of M$'s popularity: things get fixed much quicker and these "details" (which are very critical IMO) are left behind in Linux.

The problem is not *NIX neither the kernel.

These suspend/resume problems are ACPI  only related ...

Blame WINTEL for the very mess they have done with the Super I/O
chipsets which now require specialized firmware..

The mercy bullet is called UEFI in which unless things match
your hardware may be a total or partial brick

Things are just converging as always:
- no driver support
- a messy obscure closed source of firmware and buggyware..
- ultimately rendering *ONLY* allowed shell partners to succeed

Alas same film again again... only those selected have undisclosed
information to make that shit really work

the rest..  well the rest will have to figure it out or just doom themselves..

nothing new..  same shitware they do since 80s

Paul