Unusability of bluetooth is one of the worst plagues in Ubuntu

[Matteo Sisti Sette]

On 26/12/13 19:14, Gabor Toth wrote:
 > I would question though if bluetooth would be the most major issue.

Obviously not the most major issue in Ubuntu. There are bigger issues, 
like those that completely freeze the whole system obliging you to do a 
hard reboot and loose all unsaved data, or sudden xorg crashes/logouts 
(which also wipe out all your unsaved data) or not being able to 
hibernate (and if you do, take the risk you may not be able to resume).

But bluetooth and wifi connectivity (oh!! and mobile broadband! what a 
nightmare too) are the kind of things that one expects to be completely 
painless, as you won't easily find a smartphone that has the smallest 
wifi/bluetooth issue, so in comparison it's a little embarassing and 
irritating to find yourself unable to use bluetooth on a laptop.

 > I personally did not have much problem with that for what
 > I was using it
 > for - sending music over to a bluetooth device.  Worked with a
 > few click
 > no manual setting or scripting at all.

I personally use it mainly for transferring files (primarily pictures) 
from my android smartphone to the computer (which is only a small 
fraction of its uses, I realise), and it also used to work without and 
manual setting or scripting at all... not even a click! (except clicking 
on "OK" when you had received a file), on previous versions of Ubuntu. 
It occasionally would start to systematically reject all transfers, 
which used to be fixed by turning bluetooth off and on. Now on 13.04 and 
on my new computer, sending a file from the phone is completely 
impossible, so I've started browsing files on the phone from the 
computer, which works like 15% of the times.

 > I do think that a team to look into bluetooth would be helpful.
 > However,
> as there are few people who can work on sections such as the drivers
> for, e.g. nvidia, In the mean time, go complain to companies who do not
> have linux drivers for their hardware :)

Correct me if I'm wrong but I'm under the impression everybody here is 
assuming all or most bluetooth issues are related to drivers. The fact 
is most of the bug reports have not even been looked at, and I suspect a 
few of them are not even related to drivers, being either in the 
tray-icon applet or in Nautilus. In many cases the hcidump is completely 
empty while experiencing the issue.



By the way, I got the impression there is a kind of more general issue 
in the way bugs are managed, please take it into consideration, I may be 
completely mistaken of course. It looks to me like there's a gap that 
needs to be filled between users reporting bugs and developers working 
on fixing triaged and confirmed bugs.

That is, the typical lifecycle of a lot of bug reports (of course I see 
this mainly on bugs I report myself, so I don't claim this is 
statistically significant) is:
- a user files a bug report
- then either:
-- A: nobody (capable of fixing, investigating, triaging or digging into 
it) even looks at it until it's confirmed, which may be in ages or just 
never
or
-- B: - somebody (or a bot) asks for more infomration, like "please test 
upstream kernels", or whatever.
       - The OP simply can't, or doesn't want to, or hasn't the skills 
or the time to, or just WON'T, do that. Most of the time the required 
work of collecting additional information must be done by somebody with 
more experience and knowledge than you can expect by the original reporter.
       - so the bug expires.


Most of the bugs I report end like that, and many that I didn't report 
myself but I found out describing the issues I was experimenting, had 
the exact same fate.

If the strategy used to tackle bugs is:
- developers only look at/for confirmed, perfectly triaged bug reports 
with already enough information to start fixing them
- expect users who encounter the issues to file complete, perfect, 
already workable bug reports
- then have some people dedicated to testing who, of course, will file 
complete, perfect, already workable bug reports (and expect this to be 
enough to find and report most relevant issues)

then sorry but that's a wrong strategy and is not going to work. Those 
trained people dedicated to testing will only find a ridiculously small 
percentage of bugs affecting end users (I'm not saying there's anything 
wrong in that, that's certainly a necessary and invaluable work).
But bug reports filed by end users, finding bugs that have eluded 
"dedicated" testing, will either starve or expire.

Either developers who work at fixing bugs or testers who usually do 
thorough testing, should also look at unconfirmed, need-info, etc. bug 
reports where the OP hasn't provided (and will never provide) all the 
needed information, because often the bug report, whether or not the 
issue is easily reproducible, does contain enough information for 
somebody to investigate and test it further and take it to the next step.


As a user, like me, who uses some amazing software that has been created 
by the huge effort of thousands of generous people, when I find an issue 
and take the time to report it, of course I don't feel like I'm doing 
much to help. I do see I'm doing very, very little. But it's just as 
much as I can do and it is frustrating to see that even that "very 
little" gets wasted and turns into nothing when it could actually be used.