Hi,
...
This definition works well if there is a finite group working on the
release cycle. But here, and with other OSS, everyone has the
opportunity to be involved in the progression of the software through
its release cycle.
Remember the discussion about re-establishing our release committe? We
could have (and imho whe should have) a "close to finite" group working
on the release cycle. At the moment to me the situation looks like most
people think "anybody cares, so why should I" .. but it's unclear, who
is that "anybody" .. could be "nobody" in some cases.
showstopper has been raised yesterday at the releases list. It has
already
been marked as fixed, so I'd expect to see a new RC within the next few
days.
So, at this time, we *know* that our RC is not ready to be released. A
Yes, we will, so we will have another one. But don't you think it is
better that many more people are able and willing to download and test
the "releases that are not yet production" ?
Of course, it is better to have more people downloading and testing ... but:
- in theory our QA project has enought members do build this set of
"more people"
- members of the qa project ar (or should be) trained to raise issues,
so that they can be handled immidetely. If you move release testing to
the end user, we waste time. We waste our time and the user's time
Measured and experienced testers are vital, but end-users that are
willing to give it a go are also invaluable as they often do things in
a way that is difficult to include in regression and smoke tests.
Agreed .. but I had given some workshops on release testing and one of
the messages always is "don't follow the test cases to closely .. and
*think* what you are doing".
So .. the problem with the announcement is, that we cannot be shure
who ist
testing what RC ...
The files have dates and it is the first rc.
Well .. you might see this in another way, if you had tried to confirm a
lot of end users issues. they are of mmixed quality, some even fail to
report the proper version, if it is not a rc. It is not, that I think,
it's impossible .. but it takes time to find such issues, evalute them,
see that these are stoppers. And all this might happen by accident.
Or do you think, an end user will pop up here at the releases list and
raise a showstopper?
We have always announced the release candidates widely assuming that
the QA team are in the thick of it and the most aware.
No, that has been started with 1.1.2 or 1.1.3. Before that time we used
to announce RC within the project but not to the enduser.
Perhaps we can re-visit the process on the QA project list. I really
would like to invite more people to join the QA project as I find that
when a release is made this is when we have the discussions of "I
didn't know it was going to work that way" when it has been
progressing that way for months :(
http://qa.openoffice.org/servlets/BrowseList?list=dev&by=thread&from=489067
Has been stopped (imo), as:
- discussion abut re-establishing the release committee slowed down
- current (formal) QA-Leads don't care
André
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
