Problems Linux Enthusiasts Refuse to Address

[David Gerard]

On 5 April 2011 12:12, Liam Proven <lproven at gmail.com> wrote:
> On 5 April 2011 08:50, Avi Greenbury <lists at avi.co> wrote:

>> Really? At the risk of coming across as some sort of an uber geek I thought
>> most of those problems were fairly well solved by now? It's been a long time
>> since I've had a bit of hardware that didn't work in Linux out of the box,
>> and I've never got into the habit of checking compatibility first.

> You're lucky. I'll give you the last example to try my patience: the
> MSI Wind Top all-in-one touchscreen PC. The touchscreen works, but the
> directions are reverse: move your finger left, the pointer goes right.
> Move it down, it goes up. Completely useless.


J. Random Hardware on Linux either works perfectly or there's no hope
whatsoever, with very little in-between.

The Linux hardware problem is that the hardware manufacturers don't
care about Linux the way they do about Windows, so don't bend over
backwards to get a driver into Linux and userland controls into the
distros the way they do to have a Windows driver and controls ready
for Windows. This is not something the distros can do a huge amount
about.

Binary drivers on CD for Linux would duplicate the workaround used for
Windows, but binary drivers for Linux are an example of something that
people who think they want them are actually wrong about wanting,
because they will temporarily solve a small problem at the serious
expense of the whole thing.

Note that this is mostly a problem on the desktop - a server
manufacturer whose x86 box doesn't work immaculately in Linux is
*dead*. So your new server will work pretty much perfectly, with very
few exceptions.

The solution:

* more devices using Linux
* more GPL enforcement on them, meaning the code is available to the rest of us.

This is not fast, but it does have the advantage of actually working.


- d.