> Now I know what the name of the list is, but to be honest I doubt most  
> of us charged out and got an HP just to write an OS for it.

I did, actually.

> The unfortunate truth of the matter is that at the moment is isn't worth  
> anyones time to rescue one of these machines. For a start there is no  
> way to boot one unless you already have NetBSD working on another  
> platform.

I am booting off of SCSI tape.  It's true that there is no way to
install the FS from tape (yet), but I *can* boot without NetBSD on
another platform.  The "conv=osync obs=20b" options necessary to
write a "good" tape are annoyingly undocumented, but it works.

> I think the problem is that there is a massive difference between the  
> guys running the porting effort and the first time user. From reading  
> the group I would say that people like Jason have 2 or 3 different  
> machines all net capable and at least a working HPUX or 
> NetBSD machine to boot off. In addition they have several years of  
> HP experience and are frankly of Guru or SuperGuru status in Unix
> hacking :-) The way the OS installation is handled is designed for these  
> people. IF you don't have the capacity to netboot then you are out of  
> luck.

Yes, I agree that the NetBSD effort has not put much effort into this
area -- the area you refer to is called "productization".  It's probably
the main reason for the FreeBSD x86 beating the NetBSD x86 port in the
general "popularity contest".

There are interesting organizational/group-dynamics issues that need
to be resolved before any of the BSD camps can really proceed in this
direction (whole hog coss platform productization/commoditization).  I
frankly don't have an answer for them, and that's the main reason I
haven't made a posting like yours.  The "core team" organizational
architecture is a self-limiting trap, but I don't have any structural
replacement (yet) which scales any better, which you could use in
its place.

Personally, I have little direct HP experience (well, non-PA/RISC
experience, anyway).  I'm keeping notes on what is causing me to
stumble, but until I'm done stumbling, I can't give you a comprehensive
"hit list" of things to fix.


> The installation notes say that this will be fixed but unfortunately this  
> has NO priority since the leaders of this effort already have working  
> systems. I would offer to help, but I can't get the OS on my machine.

Write down each of the issues that bites you as it bites you, and the
soloutions, as you implement them (that's what I'm doing).  When you
are done, go through and fix the issues that bit you, and start over
to make sure the fixes took.  For instance, my list includes three items
on disklabel, and the ability to install the "inst-11.fs" from SCSI tape.
So far.


> It's ok adding all this great stuff and I'm sure that things like the 4k
> page binary compatability are really important to someone but until a  
> guy can pick up an HP and boot with a tape or a floppy or a CD  
> without a second host I don't see this port as being viable.

I think CD won't work unless we recode the roms.  The floppy might work,
but I have yet to see a system with a floppy.  That basically leaves
tape, and you must have another system to build the tape in the first
place (or to pull a floppy image from the net or to burn a CDROM, for
that matter).

All of these are distribution/packaging issues.


> Anyway I have an offer. If people will send me what little hardware  
> data they have I'll start to compile a hardware FAQ. I'll need help  
> because I've only seen 2 or 3 of these machines but I'll do all the admin  
> and compilation and posting. This is all I can donate to the project  
> because I can't get the damned OS on my machine. Now once we have  
> some hardware data AND an OS you dont HAVE to netboot I think  
> we can start making progress.

I have *ZERO* hardware data, other than what comes up on the console
during boot, and a history with 68k based Amiga and Atari hardware;
sorry, but I think your plan isn't terribly workable (for me, anyway).
That's not to say that a hardware FAQ would not be auseful resource --
especially one which dealt with all the ROM issues for people unlucky
enough to be running Domain/OS ROMs.

On the other hand, I think binary compatability is a critical issue,
if NetBSD is to be the #1 upgrade path for people with hardware which
HP is no longer supporting very well (the primary role I see for NetBSD,
actually, even on Sun hardware: it upgrades SunOS, which Sun has
replaced with Solaris -- ie: Sun is not supporting its hardware very
well).  If you can't run their software, well, you're not an upgrade.


					Regards,
					Terry Lambert
					terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.