[kubuntu-devel] work item style

Harald Sitter apachelogger at ubuntu.com
Tue Feb 28 12:11:59 UTC 2012


On Tue, Feb 28, 2012 at 10:57 AM, Jonathan Riddell <jriddell at ubuntu.com> wrote:
> On Tue, Feb 28, 2012 at 01:27:44AM +0100, Harald Sitter wrote:
>
> Morning, and there was me thinking I was the grumpy one :)

Don't try to out-grumpy the apachelogger, it does not work ;)

>> Random guesses on what the items *could* be:
>> [kubuntu-members] Check if disabling krunner has considerable impact
>> on memory/log in time, iff so discuss and disable it
>
> This comes from discussion at UDS where we weren't sure why you had
> disabled krunner on kubuntu netbook.  If it does save significant
> login time that's all good.
>
>> [kubuntu-members] Ask upstream KDE PIM on their recommendation of
>> mysql vs sqlite, change our defaults accordingly
>
> This comes from Kolab Sys going to trade shows and saying "if you have
> problems with akonadi and mysql that's because they haven't chosen
> sqlite".  It's nonsense as I and you have discovered so we'll stick
> with mysql.
>
>> [kubuntu-members] replace kopete with kde-telepathy by feature freeze
>> iff community is in favor
>
> This is done but of course a final decision is yet to be made.
>
> There's limited space in the work items lines for a full explanation,
> that's the disadvantage of the work item system over the full written
> spec system.  So if you don't understand anything do ask.

I do appreciate the problem of having to convey considerable amount of
information in a short line, but that is in almost all cases a good
thing as you have to focus on the important stuff plus big things most
of the time need to be different WIs, so splitting them accordingly is
a good thing too :)

Asking is unnecessary overhead (yeah I know, this sounds like I don't
like talking to people, but read on for why it produces overhead WRT
what I was complaing about ;)). Particularly since one would first
have to find someone who was attending UDS and attended the session
where this was discussed and still remembers the motivation behind the
WI. And then probably wait until this person gets home from work or
something. I really want to stress that it is not about putting all
the information into a WI, but the information necessary so that
anyone who has a bit of an idea what we do and how we do knows what
this WI requires him to do.

> There's nothing wrong with having a research item as a work item, it
> still takes time to look into and decide what to do.

Absolutely. But even so you need to outline what to do to conclude the
WI like in my examples. It does not help much if someone does research
but only publishes vague results on IRC, and that did happen (FWIW, I
myself did do that :P) though it is easy to prevent by reminding
ourselves that research results need to be published somewhere
specific. If we look into boot time performance and find that since
the first kubuntu release the boot time decreased, then that is a very
good reason to blog about ;) Equally more boring matters might better
be sent to the ml or documented on the wiki and then sent to the ml.

HS



More information about the kubuntu-devel mailing list