Received more than 500 emails on the original piece published at OSNews a couple weeks ago. Reactions to “Tales of a BeOS Refugee” includes further clarifications on my ideas about the Creator code and application binding, plus dozens of miscellaneous notes and continued comparisons between BeOS and OS X.

The piece I wrote for osnews.com came out today. 15,000 words on the experience of a BeOS user migrating to OS X. Got slashdotted later in the day. Mail is rolling on, most of it supportive. I got a few things wrong, as always, but overall the response has been pretty good – still waiting for the flames to start – I was pretty harsh on Windows, Linux and pre-X Mac OS. In fact, reading again, I sound like I hate everything, impossible to satisfy. Nope – I just think we won’t get perfection from anyone if we don’t demand it.

19 days of sweet uptime on the OS box, then we have to go and have a five-second power outage today. After reboot, most of my apps won’t launch. The dock icons just bounce eternally. Try to Force Quit and they won’t quit. App launching is stuck in limbo.

Reboot in single user mode and run fsck -y. One small error found, but that doesn’t fix the problem. Try to reinstall X 10.1 but it won’t let me – says it can’t find a valid OS X installation. Which is ridiculous because the OS will still boot.

Call Apple Support for the first time. Friendly and patient, but I felt like I knew more about OSX than the rep did, and I’ve only been using for a couple months. He suggested installing over again back from OSX and going through the upgrades, which is what I was hoping to avoid.

So install OS X. Install OS X.1. Then do all the online software updates to get back to X.1.1. Some of the downloadable updates are upwards of 10MBs. I’ve already done these – aren’t they cached somewhere? Why do I have to download them all again?

Problem solved, but there goes half my Saturday. As if there weren’t other things I wanted to be working on today.

Lessons:

1) HFS+ needs journaling badly.
2) Apple Support ain’t that helpful. Could have gotten better answers for free online.
3) HFS+ needs journaling badly.

This is pretty amazing. Amazing that there’s finally a near-complete and total usenet archive for the first time, and amazing to see some of this old stuff. I remember some of these posts and scandals, but not most of them. Got a kick out of the first usenet post from an AOL account – I helped to beta test that service. That message mentioned the fact that CompuServe (CIS) was just starting an Internet gateway. At the time, none of the proprietary services had connections to the net – AOL, CIS, and Prodigy were all self-contained universes. I remember when my CIS ID (CompuServe was command-line only at the time) — 72241,1777 — suddenly got access to the rest of the world and I became 72241.1777@compuserve.com . Everything opened up at once. A few months later, CompuServe got a GUI. Saw the Mosaic browser for the first time later that year. Everything was happening so fast…

In the ID3v1 spec (see end of document), the genre “psychedelic” is misspelled as “psychadelic.” Which means it’s misspelled in the id3ren tool I recompiled for BeOS, which means that RipEnc has probably created tens of thousands of misspelled attributes on people’s BeOS systems over the years. I could have recompiled it with it spelled correctly, but didn’t want to affect compatibility with other tools on that genre. I should have done it anyway. Now the misspelling shows up in a screenshot that’s going to end up in my BeOS/OSX piece on osnews.com on Monday. And I’ll look like an idiot, even though that institutionalized misspelling has been bugging me for years. I wouldn’t have thought of it except that mneptok spotted it and called me on the carpet for it.

How much of history is mistakes compounded by time and propagation?

Whew. Finally finished with orphan removal on the j-school site.

As of Nov. 13 there were:
12,260 files
998 MBs

As of now, there are:
3,798 files
113 MBs

… and without breaking any links ;). For years, the students have been able to drop pretty much anything they wanted anywhere they wanted. What a mess. The staging server I’m building will prevent that from happening again. All of this cleanup has been just a preface to the things I’m actually here to do.

I can’t recommend HTML Rename highly enough. It helped me lowercase all files and corresponding links sitewide (so the site will work when we move to Apache), and to dig up all these orphans. I ended up corresponding with the developer quite a bit, and beta testing builds after I started sending bug reports on false positives.

Tomorrow we build a new server. The current one is NT4/IIS. Looks like I talked them off of IIS, but we’re keeping Windows for ease of use. So it will be Win2K/Apache/MySQL/PHP. I guess that’s WAMP rather than LAMP ;) Once that’s up and running, I can dig in for realsies.

Over time, I’ve slowly been expanding the content on a page I maintain, Why HTML Email Is a Bad Idea . Today TinyApps.org (which is dedicated to simplicity in computing) linked to the piece in their email newsletter (archived here ).