Tuesday, 20 February 2018

Re: More diagnostics data from desktop

On 19 February 2018 at 13:55, Matthew Paul Thomas <mpt@canonical.com> wrote:
Will Cooke wrote on 14/02/18 15:22:
> We want to be able to focus our engineering efforts on the things that
> matter most to our users, and in order to do that we need to get some
> more data about sort of setups our users have and which software they
> are running on it.
> We would like to add a checkbox to the installer, exact wording TBD,
> but along the lines of "Send diagnostics information to help improve
> Ubuntu".  This would be checked by default.

I've just drafted a design for this. <https://goo.gl/yJ86Qa> It's
basically a subset of the System Settings screen.

Thanks, we will take a look at that, but on first scan it looks good.
> The result of having that box checked would be:
> * Information from the installation would be sent over HTTPS to a
> service run by Canonical's IS team.  This would be saved to disk and
> sent on first boot once there is a network connection.

Information from the installation would be fascinating, for improvement
of the installer in particular. However, I don't think it would give you
an accurate idea about the "sort of setups our users have", for
improvement of Ubuntu in general. It could lead you to think that, for

*   Internet connection is less common than it really is. (Because of
    things like proxies, as mentioned by Ernst Sjöstrand, or
    not-yet-installed wi-fi drivers. And because if people still can't
    get online later, they might uninstall Ubuntu and you'll never get
    a report.)

*   Wired Internet connections are more common than they really are.
    (Because they're being used temporarily during installation, while
    wi-fi isn't working.)

*   Typical screen resolution is lower than it really is. (Because
    people don't tweak the resolution until after installation. And
    because if they fail to do so, they might uninstall Ubuntu,
    resulting in a report for a system that soon stops existing.)

*   Bluetooth devices are much less common than they really are.

I think it would be much more interesting to measure these things month
by month.

Yes, general trending is what we're aiming for here.  Thanks for the input on those specifics.

>                                                         The file
> containing this data would be available for the user to inspect.
> That data would include:
>    * Ubuntu Flavour
>    * Ubuntu Version
>    * Network connectivity or not

If I understand "on first boot once there is a network connection", that
would exclude devices that were offline until the second startup or later.

We were thinking along the lines of something which would try to send the data at login a number of times, let's say.... 10, and then give up.  So if the machine never comes on line, then the data never gets sent.  If the machine travels between various locations before arriving at a working internet connection, then it should eventually be able to send it.  I think that would cover the vast majority of cases.

>    * CPU family
>    * RAM
>    * Disk(s) size
>    * Screen(s) resolution
>    * GPU vendor and model
>    * OEM Manufacturer
>    * Location (based on the location selection made by the user at
> install).  No IP information would be gathered
>    * Installation duration (time taken)
>    * Auto login enabled or not
>    * Disk layout selected
>    * Third party software selected or not
>    * Download updates during install or not
>    * LivePatch enabled or not
> * Popcon would be installed.  This will allow us to spot trends in
> package usage and help us to  focus on the packages which are of most
> value to our users.

This effectively singles out .deb package installation as the only thing
that should be reported periodically, with everything else reported
one-off. Is that just for ease of implementation, or is there a reason
not to report the other things periodically too?

Mainly ease of implementation.


For example, if we could see how often people change their reported
location, we'd have info on how accessible the time zone UI should be.
And if it turns out that only a tiny fraction of Livepatch users turn it
on during install, vs. afterwards, that would influence future installer

Wouldn't that involve us being able to track a person/machine to know when it had been changed?  Or would something in the location picker send a signal?  I don't like either of these options, I've probably misunderstood the idea.


> Any user can simply opt out by unchecking the box, which triggers one
> simple POST stating, "diagnostics=false".

What is the purpose of this?

To try and measure engagement rates.  This would be important in the "opt-in" case I think, how representative of users is the data?  < 10% of people are submitting data, then probably not very useful.

Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel