On Sat, 14 Dec 2013, at 11:25:40 -0800 Adam Williamson <awilliam(a)redhat.com> wrote:
One thing they've floated as an idea is to have a separate 'installation
environment' you could boot into from the live images - so you could
either boot into 'try it out' live mode, or 'install it' installer
mode,
but not run the installer from within the live mode.
That is pretty much what I had in mind. There would be no "live
installer". The Fedora stand-alone installation medium would be large
enough to have a live system that offered "try" mode, "install" mode
(full anaconda, with local repository), and maybe "rescue" mode. The
same system image would be used: options built into each choice would
direct which mode to execute.
I think Ubuntu (or one of its variants) uses a scheme like this, but I
do not know if it copies the live image or offers multiple
installation variations.
It is very crude, but I looked at how many models of USB flash drive
Amazon sells directly and see:
512MB & Under 4
1GB 39
2GB 176
4GB 378
8GB 518
16GB B 407
32GB & Up 465
Two thirds are models with 8GB or larger capacity. That does not mean
this many flash drives sold have that capacity, but such capacity is
readily available. There still are many machines that do not boot
directly from USB, but that is a condition for which GRUB and friends
are made.
I cannot quantify how much a single installer for Fedora saves over
two (live and multiple-environment). If one installer is adequate, it
still may not be worth the disruption caused when the other is
dropped.
The single installer choice could be made either way. The alternative
is only a live installer, which becomes the starting point for
subsequent installation by yum of the desired environment(s).
if you're interested in making [a limited resource Fedora]
happen...
I think one of my problems is too much hardware, not too little.
Somehow, I seem to have acquired the notion a new release of Fedora
justifies a new machine on which to run it.
I don't really see a lot of evidence that many other groups test
their stuff at all in any particularly organized way
There are many reasons this is so. One may be a perception there is a
QA group that will do this, therefore little need exists to do first
what will be done again. This is not the most important factor, I
believe, but a consequence is the more QA does, the more this attitude
grows. You become a victim of your own success. Greater clarity
about what QA will not test may help.
I'll note two design choices in storage which make this problem
exponentially harder...
Yes. Test the untestable. I do not think those choices are caprice;
they allow users, most of whom are no more than casually familiar with
Fedora installation, to find a more comfortable path through the
installation process, one that permits minimal rework (avoids the
"Something is wrong - start over" event) as understanding grows and
bad choices are improved. Fedora testers, who usually have an
abundance of installation experience, likely do not appreciate the
value to less experienced users of this flexibility that makes tests
so difficult.
for upgrading, we only test upgrading a very clean installation of
the previous release
Upgrade is a can of worms. Yes, I too am beguiled by how much easier
upgrade is than a new installation, and have often chosen upgrade, but
something (many somethings!) sooner or later bite. There are just too
many initial conditions, and too many distinct ways appropriate for
different users to handle these conditions, to make possible an
accurate understanding of what results from an upgrade. In lucid
moments, I know upgrade should be exorcised, but it just is *so*
convenient.
Has Fedora QA discussed how much effort they should or can invest in
organization and facilitation of others' test activities? Direct
testing scales (approximately) linearly with number of people, but
education, organization, and leadership has the potential to scale at
greater multiples.
|-----------------------------------------------------------------------|
| Great opportunity! Become a Fedora test franchisee. We'll provide |
| directions, training, and all the materials you need, so you too can |
| participate in this fast-moving field. Gain experience, recruit |
| others, become a Test Manager and train new Testers. With 10 or |
| more Testers, select your own Test Managers (each of whom will direct |
| at least five Testers) and advance to the Test Director level. With |
| five Test Managers, become a Test Executive... |
|-----------------------------------------------------------------------|