[Testing] Today's/Tonight's QA Phone Call Conclusions & Unofficial Minutes

[Tim McNamara]

On Thu, Feb 25, 2010 at 7:07 AM, Mel Chua <mel at melchua.com> wrote:

>  Seconded, as we are all volunteers an asynchronous model might make most
>> sense. Especially when there is little time to do things one has to be
>> organized well and the teams must have a process they can follow and be
>> sure others get to the results.
>>
>
> Yep. One thing I think might be helpful is making the weekly
> test/development cycle more explicit - sounds like we already have one (we
> have weekly dev meetings, there are weekly QA meetings in Welly) but I'm not
> exactly clear on the timing of the two, or when each group is reporting
> results back and forth to each other.
>

Just to update everyone on NZ's progress.. we now have two testing groups
meeting weekly: one in Auckland (led by Tabitha), one in Wellington (led by
a small consortium/tbc/is that important?) & another small cadre in
Christchurch. We meet every Saturday our time, which is late Friday
afternoon in the USA or about midnight Friday in Europe.



>
> So being able to say "at each week's development meeting, we'll look at
> <list>, and email <this stuff!> to the testing list", and vice versa for
> each week's QA meeting, may be useful for getting us into a rhythm.


I don't think anyone from here (NZL) will ever be able to regularly attend
any of the Sugar/OLPC weekly meetings. They're generally on around 1300-1500
UTC, which is roughly 1am - 3am here..



>  I would love there to be one webpage for test requests that has some
>>> visual indicator of how many people testing it and how close to
>>> finished testing it is, with a big tick next to it when it has been
>>> tested thoroughly.
>>>
>>  Yes, that sounds very good indeed. I guess we do not have any
>> infrastructure for the working scenario you describe (yet)?
>>
>
> Not that I know of - though I'd love this as well. I'm setting up a
> Semantic MediaWiki test case system similar to the one we had at OLPC, just
> to have something to get started on. No firm date on when that's going to be
> completed, though I'm going to be tackling getting that infrastructure up
> this coming Friday.
>
>

I'm against bespoke development in this area, unless the alternatives are
untenable. I would prefer if we evaluate some of the current UAT testing
tools - if they're unsuitable for our needs, then building something
ourselves makes sense.


>
> I know Mel wanted to do QA for her upcoming pilot and was involved in
>> the QA efforts at olpc. Mel, do you have any thoughts about testing
>> tools that allows what Tabitha describes? Testlink [1] was mentioned
>> already, for example.
>>
> Not off the top of my head... general consensus of the past year or so
> asking QA folks about FOSS testing platforms is "there aren't any *really*
> good ones, FOSS or non-FOSS." Something we should keep pushing on, though.
>
> My questions:
>
> 1. Where's the mailing list for Sugar Labs testing (on any platform - XO,
> SoaS, other)? (I'm guessing testing at lists.laptop.org - that's fine, we
> should let everyone know that's the canonical location.)
>
> I think most testing discussion happens on the sugar devel list. Don't know
if there is a specific Sugarlabs testing list.  In keeping with my message
to coordinate things between OLPC/Sugarlabs/Activity developers - I promote
huddling around the testing at lists.laptop.org.


> 2. What's the weekly report that SL dev will give testing (and testing
> wants/needs to get from dev)? (Someone on the devel list should have the
> responsibility of pinging the testing list once a week to relay this over.)
> When will this be sent?
>

I don't know if there is a weekly report.. this might generate quite a lot
of admin overhead.


> 3. What's the weekly report that SL testing will give dev (and dev
> wants/needs to get from testing)? (Someone on the testing list should have
> the responsibility of pinging the devel list once a week to relay this
> over.) When will this be sent?
>
>
Since testing is often more than Sugar, I don't know if this could be
achieved. I support each group/individual reporting back once she or they
have something to report. Once we have a more formal structure established,
then we can consolodate on email traffic.


> 4. Is everyone hanging out in #sugar as they have questions on Sugar QA? If
> not, how do we get them there?
>

I've found #sugar to be pretty unresponsive when I'm active, because few
other people are. Planning around an asynchronous model was suggested
earlier in the thread.



> Basically, I'm trying to figure out the rhythm and the back-and-forth we
> already have - because we do already have it - so we can use it more
> effectively.
>
> --Mel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.laptop.org/pipermail/testing/attachments/20100225/fc75607a/attachment.htm