The heroes of Fedora 18 Final testing - wiki matrices

[Robyn Bergeron]

----- Original Message -----
> From: "Jóhann B. Guðmundsson" <johannbg at gmail.com>
> To: test at lists.fedoraproject.org
> Sent: Monday, January 14, 2013 5:24:12 AM
> Subject: Re: The heroes of Fedora 18 Final testing - wiki matrices
> 
> On 01/14/2013 12:16 PM, Michael Schwendt wrote:
> > On Mon, 14 Jan 2013 09:56:11 +0000, Jóhann B. Guðmundsson wrote:
> >
> >> On 01/14/2013 09:54 AM, Kamil Paral wrote:
> >>> I'd like to thank all who participated in Fedora 18 Final
> >>> testing. I have published some statistics here:
> >>>
> >>> http://kparal.wordpress.com/2013/01/14/the-heroes-of-fedora-18-final-testing-wiki-matrices/
> >>>
> >>> Thanks again.
> >> Can you please remove any paid Red Hatter from that statistic
> >> since no
> >> community member can compete with that thanks.
> > That doesn't matter here at all. It's more important that actual
> > tests are
> > performed, and painstakingly. It's great to see people working on
> > this.
> >
> > A bit of background for these statistics might be enlightening,
> > however.
> > I only remember the first Wiki "matrices", which are not everyone's
> > cup of tea,
> > and certainly there are lots of community people reporting
> > installer bugs
> > into bugzilla separately from this Wiki based activity.
> >
> 
> Which is another point this only collects what got written in the
> wiki
> not what's reported in Bugzilla which excludes a largest group of our
> reporters and some of our reporters might also be writing QA related
> wiki pages and doing another QA related work. Dont those indviduals
> deserve credit as well?

I woke up in the middle of the night Saturday night wondering how much time folks spent in blocker meetings this release.

Admittedly this is another item that could probably be at least partially automated - I spent about 2 hours collecting some stats manually from ALLLLL of the blocker meetings, which means there are probably some minor math/rounding errors here (but would probably still be looking at a 90% unfinished script had I tried that route) - but here's a snapshot:

* 4955 minutes of blocker meetings (that's more than 82 hours, almost 3.5 solid days of blocker meetings). This doesn't count anything done in a go/no-go, etc - with the longest clocking in at 291 minutes, unless you count Beta 2 + 2.1 meetings, since the first meeting got capped just before the 4 hour mark - which came in total at nearly 6.75 hours.)
* ~380 proposed/accepted blockers/nth discussed - this is individual bzs - does not repeatedly count blockers discussed more than once, twice... even 5 times, like our friend 855526.  (The ~ here is because I literally wrote these down manually in a notebook rather than typing them out - when I got to the second page I started questioning why I did it this way, and by the fourth multi-columned 4th page let's just say that I believe that I am acquainted fully now with writing w/my new fountain pen.)
* # of lines said in meetings - just from what meetbot adds at the bottom of meeting minutes - tflink with 7882, adamw with 5276, kparal with 2149, vikingice with 1315, jreznik at 1077, many others numbering in the hundreds - jlk, nirik, martix, dan408, cmurf.

So there's some other QA work that deserves credit as well. :)

As for the point regarding removing red hat folks from being counted because nobody can compete with their ability to have dedicated time - I think that's a mistake. And here's why:

* Context.  Inevitably emails with *any* type of statistic will work their way around to all sorts of places - whether it's some internal pointy-haired-boss at red hat because someone on one team wants to ensure that the work of someone on another team is acknowledged by that person's boss, or someone thinks it can be used as a measure of "community health" as far as volunteer vs. paid work on a release to release basis (whether from our community or someone from another community), or someone thinks it might be useful to look at "what kind of man-hours would I be looking at to roll my own distro," or etc. 

* Because people deserve to be thanked and acknowledged and get credit. Especially this release. Regardless of their where their paychecks come from. For those who are paid - it is thanks for not turning off the computer at 40 hours, or 60 hours, or more, for working on holidays for which they'll never get credit, etc.  The folks in QA, paid and unpaid, collectively put a crapton of their spare time into this release. 

Everyone on the QA team deserves a ton of credit, and gratitude, regardless of "how many" or "how long" we can calculate anything - I seriously salute all of you, for not pulling out your hair, for not walking away, for not lowering the bar.  Saying thank you seems trivial because it doesn't even begin to cover or express everything, but I thank you all nonetheless.

And I will hereby volunteer that I am surprised that nobody has yet proposed "I survived F18" QA team shirts :)

-Robyn

> 
> The wiki was only supposed to be a a short term solution because we
> never found a testing system to that quite suited our needs.
> 
> I've ping Adam and Tim to see if it's on the plan to look ( again )
> at
> test systems and see if we find something that suits our need.
> 
> If anyone knows of any feel free to share it.
> 
> JBG
> --
> test mailing list
> test at lists.fedoraproject.org
> To unsubscribe:
> https://admin.fedoraproject.org/mailman/listinfo/test