[Warning: muse in progress.  Excuse the length.]

der Mouse bemoans:
> Do we really _want_ the point-and-drool market?
> 
> Do we want to be a research system, one aimed at people who don't mind,
> nay, even enjoy, getting their hands dirty grubbing around inside
> kernels and compilers and such?  Do we want to be a "solution", an O/S
> aimed at people who have trouble telling the mouse from the keyboard
> and are lucky not to stuff the CD-ROM into the floppy drive?  Just whom
> _do_ we want to be for?
> 
> I don't know.  I haven't heard any clear direction from core - indeed,
> from anyone - on this matter.  I'd like to know, if anyone can say,
> what the intended target audience for NetBSD is ... if only so that I
> know whether I should go my way and let it go its if I'm not part of
> said audience.  Certainly _I_ have no use for a "just click here to
> upgrade" release, and consequently consider time spent specifically
> making life easier for such things time wasted.

I think there's a big difference between the "point-and-drool market" and
making life easier (and more convenient, less time consuming, etc.) for
the technically savvy "market".

Kernel hacking and source hacking isn't for everyone.  For some folks, the
thrill of having a source code OS (just in case you *do* feel like some
hacking, or just local changes/fixes) that looks, feels and has those warm BSD
squishies is good enough.  In my case that's mostly true.  I do like some
sort of stability, but I also like the fact that it's evolving forward in
real time rather than in stutter-steps like temporally-separated vendor
releases.  But I also like functionality, since I sit in front of a NetBSD
system all day at work, and play on one (upstairs on the Pentium 120) in
those extremely precious moments that pass for "spare time" for us old-timers.
(I'm 37 now, with a wife and overhead and other mid-life concerns that you
wiry young 20-somethings who can stay up all night hacking on Mountain Dew and
donuts/candy bars just don't have to contend with.  Things Change.)

Part of the reason I say this is because I think NetBSD is sort of a happy
medium.  A lot of what der Mouse says above makes it sound like he'd be
much happier off playing with, say, Plan 9.  My own experience with Plan 9
(I have it installed on a 486DX/33 PC upstairs at work) says that although it
has a lot of Really Neat Things about it that I'd really like to know a lot
more about, in practical terms it's really hard to use as a day-to-day system.
Anybody who's used 8 1/2 and "9wm" (the Plan 9 window system) is not going to
be real happy if (like me) they're used to using "tvtwm" with lots of virtual 
windows and title bar shortcuts and triple menus and pixmap icons etc. etc.
(Does anybody remember "uwm"?  Well, that's what the GUI interface of "9wm"
reminds me of.  Sherman, set the Way Back for 1986 ... )  And there's lots of
other little things, like programs without switches for functionality I
can't live without (what, the Plan 9 "ls" doesn't have multi-column "-C"
output?!?).  It's a *much* more "bare-bones" type of environment.

Conversely, I see ways that NetBSD could easily be improved that would be a
boon to us techies, without pandering to the point-and-drool crowd.  Like
being able to install from installation sets in a DOS partition on a PC
without having to jump through serious hoops to figure out how to get it
to work.  Like figuring out how to make the PC Ethernet driver(s) probe more
than the default IRQ/DRQ combination so that you might be able to FTP or
NFS mount your installation sets if you don't happen to have your Ethernet
card set exactly like the GENERIC install kernel wants.  Or a way to start an
install from a floppy on a SPARCstation, much like on the PC.  And so on ...

I don't consider myself remotely close to a point-and-drool person, but I do
know this: the Solaris 2.5 installation program is a damn sight easier -
and *better* - than the SunOS 4.1.x "suninstall" ever was.  There's no reason
why "better" has to mean "pandering to the point-and-drool crowd".

	- Greg