Ars Technica article

[Jason Bacon]

On 2020-04-13 01:19, Alexey Dokuchaev wrote:
> [ setting CC to a more appropriate -wireless@ list ]
>
> On Fri, Apr 10, 2020 at 07:20:14PM -0500, Jason Bacon wrote:
>> ...
>> I think [well-working gfx stack] a great long-term goal, but it will
>> take a significant investment of man-hours to get us there.  In the
>> meantime there's a lot of lower-hanging fruit, like minor improvements
>> to the bsdinstall UI, external media mounting, etc.
> What I find more frustrating when reading similar articles (they are
> all more or less the same) is that they rarely focus on FreeBSD's *real*
> problems (and no, that's not the awkward installer, having to manually
> find working DRM port + xf86-video driver combination, search for the
> scattered knowledge of which magic lines to add to /etc/rc.conf and/or
> /boot/loader.conf, read the Handbook N times before one can set up their
> bluetooth mouse, et cetera).  While installer's issues, better defaults,
> bash vs sh, sudo, modelines, and those other little things might seem
> significant for someone coming from Ubuntu, it would take them a day
> to learn, adopt, and get back most, if not all, from their previous
> environment.  Our real problems aren't solved that easily, and being
> solved painfully slowly.
>
> Leaving X.org/DRM mess aside, we lack a lot in our laptop department.
> Our WiFi stack is essentially maintained solely by adrian@, and he's
> currently not very active.  Some ~4 y.o. cards like BCM43228 are still
> not supported [1].  If you search for "atheros" in our Bugzilla, it
> returns 13 bugs, the latest action being on 2019-01-26 (reassignment).
>
> AR5B22 WLAN+Bluetooth combo does not work/broken, discussion [2] had
> ended nowhere.
>
> Realtek 5209 card readers are quite common and also do not work; the
> WIP OpenBSD driver porting effort [3] is stuck because apparently
> something is missing in FreeBSD [subsystems], although [the] driver
> seems to faithfully implement the OpenBSD [code].  No kernel hacker
> had chimed in to help. :-(
>
> Our WMI stack is unmaintained and incomplete: brightness and multimedia
> keys do not work on many laptops despite corresponding kernel modules
> being loaded.
>
> This list goes on.  Yes, these are deep, hard problems, not the low-
> hanging fruit, but if Foundation decides where to spend some money,
> I'd rather see it considers these rather than bsdinstall UI, external
> media mounting, or some other "lipstick on a pig" type of tasks.
>
> ./danfe
>
> [1] https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=202501
> [2] https://lists.freebsd.org/pipermail/freebsd-wireless/2019-April/008660.html
> [3] https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=204521
Just to be clear, I was not suggesting that the foundation spend money 
on low-hanging fruit.

What I'm saying is that we can make the desktop experience significantly 
better by fixing the easy issues before the foundation considers an 
investment.

I suspect that some developers might not think fixing a WiFi issue is 
worth their time when there are so many other problems with setting up a 
desktop system.  On the other hand, if it's easy to build a desktop 
machine that mostly works, and WiFi support is one of only a few obvious 
shortcomings, that perception might be different.

I'll also add that I'm not a proponent of of even trying to make FreeBSD 
attractive to people hung up on some of the trivia mentioned in articles 
like this one.  I think our primary focus should be on attracting 
potential contributors, not average Joes who will only make demands on 
our time.  For that, we don't need a desktop experience for the 
technically clueless, just one that a reasonably savvy user can easily 
manage so they can focus on their real work.

     JB

-- 
Earth is a beta site.