Error database requirements

[Bryce Harrington]

On Thu, Aug 11, 2011 at 11:02:09AM +0100, Matthew Paul Thomas wrote:
> Christopher James Halse Rogers wrote on 09/08/11 06:37:
> > On Mon, 2011-08-08 at 00:34 -0700, Bryce Harrington wrote:
> >...
> >> Here's a start...
> >>
> >>  * The user should have some way to check back on the status of their
> >>    crash report; e.g. have some report ID they can look at to see
> >>    statistics and/or any associated bug #.
> 
> This one will be tricky without weirding people out ("what is this
> 'Launchpad' thing and why is it on my computer?"), but I'm sure there's
> some way of making it subtle enough.

Yeah, and in fact the use case I'm envisioning for this would be limited
to semi-technical folk, so it could be as simple as just giving them
some serial number on the Problem Reported dialog they can take note of
and load manually later on.

Even though relatively few people would use this feature, I think it's
important.  Too often with other crash database it feels like you're
sending things to a black hole.  By including a mechanism to allow users
to check back, it also provides a (subtle) path for sufficiently
motivated technical users to participate more actively in finding
solutions.

> > To this list I'd add: 
> >  * It should be brainlessly easy for users to submit this data.
> > Either a single "Yes, submit this crash" confirmation, or a check box
> > to automatically submit these crashes.  One of the features that the X
> > team really desire out of this sort of database is "how frequent is
> > this kind of problem", which requires the widest possible sample
> > space.
> 
> With my proposed design it's two clicks to submit, then another click to
> dismiss the error message. Probably I should simplify it further.
>
> I've added all of these [requirements] to
> <https://wiki.ubuntu.com/ErrorTracker>. (Not saying that they're right
> or wrong, just as a base to work from.)

Sweet, thanks for jumping into the design on this!


In looking at it, a few items spring to mind we should discuss further:

1.  Prior art in Linux?  Aside from Breakpad, do any other distros or
    FOSS projects have crash reporting systems like this?

2.  I notice the Windows crash reporter includes buttons to check for
    existing solutions.  Is that something we'd like to consider as a
    requirement for this?

3.  Privacy issues should probably be elaborated on further.

4.  How do us developers want to interact with the gathered data?
    I.e., what should the graphs or tables look like?

Bryce