kindergarten; panel-less plasma?; unit tests; spatial filemanagers

my head is hurting right now. i haven't felt quite right since getting home from aKademy to be honest. i didn't feel quite right there, either. while the amount and type of work we got done was terrific and the people organizing everything really went the extra mile (from driving us to and from the airport to staying up until 5am one night so we could hack in the peaceful cool of night), i found the location troublesome. i know lots of others had an amazing time there, and were i to go there on vacation it would probably be a great place. but it wasn't a great location for me to get work done.

besides the heat and network issues, there was also the fact that for those of us who don't eat animals it wasn't the greatest. not to mention that if you don't speak spanish (i don't) it was considerably more difficult to get around or get much of anything done at restaurants and shops. so, great for vacationing (the old town is beautiful and i won't complain about ocean beaches ;), but maybe next year we can consider that the location should also be conducive to working.

i'm about to head out in half an hour or so here to go pick up the p-man from kindergarten, which commenced this week. it's an odd thought for me that my son is actually in school. wow. he's loving it and the school we have him enrolled at is a truly great place for him and the other kids.

today on irc we were discussing some plasma design issues and the idea of ditching the concept of panels altogether came up. what if, instead, everything was an applet and they simply interlocked together? the idea is seductive, and i'm sure it would make danalien happy too. for the default scenario it would mean that the "panel" would probably "grow" towards the middle from each side. code wise, once the interlocking was implemented, it would remove a whole host of problems associated with multiple panels. but i'll have to look at the various applets and extensions currently in use to see if this is even an option.

i had an odd problem today with unit tests in plasma where i needed the test app to process test data (a couple of .desktop files to be exact) which obviously doesn't work with srcdir!=builddir. after consulting the build gods, coolo suggested putting this in my Makefile.am:

check-local:
        opwd=$PWD; cd $(srcdir) && $$opwd/testAppletInfo

and 'lo, it worked! qttestlib is pretty slick as well. glad we're moving to it for KDE4; unit tests are so important and this makes it pretty easy to do. a nice gui over top of it in kdevelop that integrates with the class explorer would kick ass. (this is where one of the kdevelop hackers chimes in with, "what, you haven't installed the unittest kdevelop plugin?" ;)

last night i was reading (again) about spatial filemanagers. my biggest question is how fine grained and to what extent does kinesthetic memory scale? it's a no-brainer to exploit that for common elements such as menubars, toolbars, etc and the items within them... but does it scale to hundreds of elements in small (say, 17") space? at what point(s) does it fall off? i tried to find some information on this online last night as well but came up empty.

i also have to say that until we have a rational spatial file open/save GUI for use within applications, a spatial filemanager looks a lot less appealing.

aaaah.. the ibuprofin (sp) is starting to kick in. just in time to wade through 15 kids who will inevitably all be running around happy and with great boisterousness to collect p.