Testing Issues on KDE4 pre

KDE 4 isn't ready for general use. It is ready as a development platform.

That is what this release is for.

Derek (who wonders some days whether free software will survive the 
desktop)

> They actually weren't releasing early and often, but the opposite.  It

i agree; there was a too-long period where we didn't have a release 
management team looking after the kde4 process (they had remained focussed 
on kde3 instead).

of course, we rectify that and people complain about that too ;)

> In the past the improvements were self-evident and intuitive to an
> experienced KDE user.  This time, I found 
> new eye-candy but difficulties in usage, making it hard to spend more
> than a few stints testing. 

yes, some things have changed. that's a truism. some things aren't 
complete; another truism. however, many of the improvements in 4.0 are 
pretty self-evident. 4 is the largest divergence we've taken, since 1-3 
were all essentially "build a foundation, follow the taillights" releases. 
you could say much of the same things about the first versions of macos 
(though they spent a greater number of years working on things behind the 
scenes as their os9 crumbled into the ground; f/oss vs proprietary 
differences there).

even good change sucks, and then you get used to it and it's good again. 
rejecting change as a knee jerk reaction is actually an evolutionary 
advantage for an animal such as a human: it keeps us in the safe ruts of 
what we already know works. unfortunately, it tends to keep us there even 
when what works doesn't work very well and in the case of new technology 
it can keep us in the ruts of that which doesn't work for others at all. 
many kde users struggle with some of the interesting quirks of kde3; in 
the project, and esp as someone who travels about visiting people all 
about, i get to see this quite a bit.

that isn't to say kde3 sucks. it doesn't: it rocks =) there are things 
that could be done better, and that's what we're attempting to do. i'm 
sure we'll cock up here and there, but thus far we have far and away more 
successes than otherwise going on in kde4 and what cock ups we've managed 
are pretty temporary in nature.

> When this gets discussed in blogs, folk pipe up 
> about damage being done to developer morale, and  heated arguments in

that was not due to criticism raised, but the amazing lack of civility 
displayed time and again by people who would essentially show up to throw 
a tantrum. i could show you emails, BR's and web forum comments that would 
curl your toes. that was the problem, not constructive feedback.

over the years the free software user base has grown increasingly 
aggressive, abusive and generally mean spirited; this is because it's 
grown in size and thus is becoming more representative of the populace at 
large. we don't have many cultural ways of dealing with truly antisocial 
outbursts effectively (free software projects are often either too nice to 
say anything or too ready to flame back at full bore; i've personally been 
guilty of both sins) and so right now this new vocal minority are causing 
problems. it doesn't help that the traditional flame war often looks 
similar to this new whinging that happens, but the difference is that in 
most traditional flame wars the people on all sides were well meaning and 
looking for solution.

> Frankly I don't think my "participation" submitting bug reports, really 
> helped except in one case.

that's one better than none. =) the biggest issue i've seen with bug 
reports is that they would get fixed and then others would file the same 
one, which can feel pretty frustrating for the reporter, no doubt.

i honestly wish bugzilla was a lot better, that we had a better 
distribution mechanism for development releases (well, we have the daily 
live cd's now which are pretty good) ... i do fear that we (the f/oss 
community) don't spend enough time working on our tools. (ugh, gdb!)

> Looked like release managers keen to push 4.0 out the door to get 
> a "result" in time for the Google event, and using the release labels to 
> pressure developers late on delivering features.  That seems to cause 
> acrimony and now a round of justifications.

that isn't what happened at all. in fact, we created a contingency plan 
for how to do the release event even if 4.0 wasn't out by then. the plan 
was actually pretty good and the promo team was comfortable with it.

we have been careful not to let marketing push these decisions around; in 
fact, if marketing had, we probably would've done the releases very 
differently. it probably would've gone better for marketing but worse for 
the tech, the project and ultimately the users.

but it is funny that people would blame marketing for something that 
obviously isn't good from a marketing perspective ;)