Kubuntu Testing and KDE

[Valorie Zimmerman]

On Tue, Mar 25, 2014 at 8:54 AM, Harald Sitter <apachelogger at ubuntu.com> wrote:
> On Tue, Mar 25, 2014 at 4:53 PM, Myriam Schweingruber <myriam at kde.org> wrote:
>> On Tue, Mar 25, 2014 at 3:46 PM, Harald Sitter <apachelogger at ubuntu.com> wrote:
>>> On Tue, Mar 25, 2014 at 10:27 AM, Myriam Schweingruber <myriam at kde.org> wrote:
>>>> Ciao Luigi,
>>>>
>>>> On Mon, Mar 24, 2014 at 10:07 PM, Luigi Toscano
>>>> <luigi.toscano at tiscali.it> wrote:
>>>>> Valorie Zimmerman ha scritto:
>>>>>> Hi folks,
>>>>>>
>>>>>> We in Kubuntu are trying to organize our QA efforts, to make them
>>>>>> easier for testers to use, and make the results useful for the
>>>>>> developers. Of course most QA for us happens in KDE itself. We've
>>>>>> asked Sayak to think about creating a webapp that we could use to host
>>>>>> our stuff. We'd love to see KDE chime in, and make this webapp useful
>>>>>> to all of us.
>>>>>
>>>>> Hi all,
>>>>>
>>>>> why not reuse a real test case management system like TestLink?
>>>>> http://testlink.org/
>>>>>
>>>>> It is used by OpenOffice.org: https://wiki.openoffice.org/wiki/QA/Testlink
>>>>> I took a brief look at it some time ago and it seems quite complete, even if
>>>>> it could be complex for newcomers; on the other side it is a good choice for
>>>>> the long term.
>>>>
>>>> Sounds like a very good idea, and it is not overly complex, as the
>>>> tester doesn't have to configure much, and with clear instructions on
>>>> how to use it this shouldn't be a problem. Since it is Free Software,
>>>> it can also be adapted to our needs.
>>>>
>>>> So yes, why not go for TesLlink and eventually adapt it to our needs?
>>>> Would probably spare some work, and we don't have to reinvent the
>>>> wheel :)
>>>
>>> The last time I used TestLink, the administrative (i.e. test case
>>> management) side of things was nothing short of complicated madness,
>>> that was quite a while ago however. Not being simple on the
>>> administative reduces the amount of people who are willing to work
>>> with it, eventually again decreasing the usefulness (greetings from
>>> bugzilla and all the devs who never ever look at it :'<).
>>
>> Erm, sorry, but the bugzilla argument I simply don't buy, if software
>> developers think this is difficult to use, they should maybe start
>> working in some corporate environment and get an idea how much worse
>> stuff exists out there. While I can imagine that one would like to be
>> able to combine bugs with a task driven system, they seem to be
>> completely oblivious of the amount of administration such a system
>> generates, and there simply is nothing perfect I know of. For me
>> bugzilla is quite simple to use, and I am not a developer, so that
>> argument is just crap, sorry to be so blunt.
>>
>> When we were starting to talk about moving to git there were a lot of
>> voices expressing concerns about how complex it would be and I heard
>> from KDE core developers that they would never change, they were too
>> old for that, etc. With such an attitude we will never get a testing
>> system that will be used by anyone!
>>
>> People, just get real: there is no perfect testing system, and writing
>> one by a single person is totally unrealistic. I had a quick look at
>> TestLink today and it doesn't look as nearly as complicated as JIRA
>> (which has be repeatedly mentioned as being better than Bugzilla, I
>> still fail to see how we would ever get a single bug report from a
>> user in there...) would to a newcomer. Every time we try something new
>> there are voices against it, we want to make things better, and
>> TestLink is a start, let's try this and see if we can improve it
>> before reinventing the wheel
>
> Sure, just wanted to note my experience with it.
>
> HS

I would love to hear Sayak's feedback here. If he has time to look at
Testlink and thinks it is about the same quality as he could do, fine.
If he thinks he could fork and make it better, all to everybody's
good.

If he has no time right now to code something for us, then at least we
have a free alternative. Thanks for that, Luigi.

IMO the goal is to have a testing framework that is easy for devels,
bug triagers and users. We might have to settle for less, but in that
case, we'll get less useful information in the end.

For instance, I found the Google Docs OK for one text run, but it was
difficult for me to find the page with all the testcases listed, after
running a test. Also I couldn't exactly tell if I was done. So as a
tester, I probably did fewer tests than I would have, in the time I
had available.

Valorie
-- 
http://about.me/valoriez