Linux on iBook redux (not yet)

My son has been using my old tangerine clamshell iBook with the original 5.6 G hard disk. (There's a lot of water under that bridge.) Mostly, he listens to mp3s, but he also sometimes plays around with ECMAScript and perl. Has some strange interest in the three kingdoms of China, and games based on the old tales.

(Must be a teen-age boy?)

Well, my white iBook G4 died the death. It served me pretty faithfully for about three years after I bought it used and installed a 160G HD in it. Mac OS X 10.3, OpenBSD, and Fedora 12; Mac OS 9 under classic emulation, but not booting from power on, as Apple designed it. (Wasn't interested interest in ex post facto or whatever that was, it did what I needed. I'm kind of looking at the reverse problem, now, more below.)

Then, about a year ago it just suddenly started powering down for no reason. I'd read somewhere on the web about cold solder joints. Found out enough to identify the video controller (maybe?) and test it with pressure, and then re-solder it. (3mm by 1mm flat tip was not fat thirty years ago! It's fat now.) Careful work cleared the slop between pins and it was running again.

About three months later, I dropped the poor thing on my way into the train station. My shoulder-bag strap broke. It took about a meter drop, but it was in the bag, with all my books that I read on the train. Checked when I got to the school and it booted up. Checked it more carefully at home, and the chassis was bent in around the battery. Straightened that out, but now it was powering down unannounced again.

After another three months or so of taking it apart and resoldering again every now and then, it just was getting too much, too often, so I semi-retired it. Recently had a little time and tried one last time. Not enough money (I thought) for a real iron, so I found a cheap hobby iron (ungrounded) with a fine point tip and bought that. Also found some cheap hobby grade braid.

Not a good idea.

I should have been willing to wait until I can afford real tools. Somewhere, I either burned a short in a chip from the lack of grounding in the iron, or I left a piece of braid or a whisker of solder somewhere. A different chip bubbled, I smelled the smoke, and it does not boot. Period.

I wonder if I can find parts, other than a new motherboard. (I have found those listed on the web, used, but with 6 month guarantees.)

Anyway, my son gets the 160 G hard drive, but there's that problem with the limits of Mac OS 9 on the iBook. 120G max per drive. Not partition. Drive.

So, much to my son's distress, I started trying again to get Linux running in it. Tried Debian this time. But the version of gparted currently in Debian is worse than the one in Fedora 12 for walking on the Apple Partition Map. Seems to think formatting the drive with less than it has is an error condition or something. I need to file a bug for this, and add the information to the bug I filed on Fedora.

Dug into the partition map information, found some old stuff around, a somewhat inaccurate mactech article on the change from APM to INTEL's GPT junk, walked the map, and tried patching the map.

Arggghhh. The pre-Firewire iBooks have no target disk mode to allow you to directly access the disk. Can't boot from USB on them, either. (Blast you, INTEL, for expanding USB beyond keyboards and low-speed printers!) I'm going to have to set up a live CD of something really small (openBSD, I suppose) in order to be able to patch the drivers.

Tried various things, botched it all up, re-installed Mac OS X and 9, spent some time bringing it all up to speed. Still missing a few things. Need to finish cutting the partitions and install OpenBSD. (After I get OpenBSD running, I may try loading drivers for the Linux ext3 file system and using OpenBSD's tools to get a beachhead install of either Debian or Fedora 16's highly experimental stuff running.) In the meantime, my son has his iBook back. With the ancient browsers.

I suppose I can call this all lab work for when I go after the LPIC level 2.

But this is a big part of a certain daydream. (With apologies to Steve's family and friends, for touching on his management style so soon after he crossed the stream.)

I sure wish Apple supported their old hardware and software better. They have the money now, and have no need at all to push so hard to get their customers to move to the latest, greatest. It would be a common courtesy to their customers to keep the paths open to all the old data.

• First, Mac-on-*nix. That really should be an official product of Apple, and it should run on Linux, BSD, and Mac OS X. If they didn't set the price too high, it would pay for itself monetarily and more than pay for itself in customer good will.
• Second, the old versions of Mac OS X. They should try to to at least keep them road-worthy -- security fixes and a reasonably functional web browser. Maybe active support for the authors of some of the better alternative web browsers that still support the old systems.
• Third, driver support for non-Apple Libre/Open OSses. Nothing to do with the crown jewels, it's just selfishness, pure and simple, to use closed drivers to try to force your customers to stay with you. Also reflects a lack of confidence in your product. Most important, it prevents your customers from helping you find and fix bugs. (Okay, that's the closed-vs.-open argument and it applies to all technical works. But these are real principles and they sit at the center of the current economic implosion, and they also drive a lot of the excess industrial activity that promotes the global climate changes.)
• Fourth, the old system software and hardware. Apple is not using Mac OS 7 any more. Why shouldn't they let students play with it? Publish it for reference, license it for non-profit study use, if they don't want to put it under their APL.

I'm sure that there are "intellectual property" (that old oxymoron) issues that would get in the way. But that could be solved by setting up a sister company, maybe call it Crabapple Computers, or something, to "launder" the IP rights issues and manage the open source projects.

Just make open source projects for all the major elements, and support them the way Red Hat (the real leader in open source, you guys!) does Fedora. It shouldn't be that hard to do.