This is BrainLog, a blog by Dan Sanderson. Older entries, from October 1999 through September 2010, are preserved for posterity, but are no longer maintained. See the front page and newer entries.

May 2003 Archives

May 22, 2003

By now most of us know that spammers send email addresses with forged headers; the email address in the From line is not the real From address. We also know that spammers think they're clever by using other email addresses on their spam lists as the From address, so they might be sending spam with your email address in the From line. Well, now spammers think they're even more clever by sending you an angry email complaining that you are sending them spam, that they have reported you to your ISP, and that they have put your email address on hundreds of spam lists in retaliation. Their goal? To get you to respond to their message, and in doing so, confirm that your email address is valid and is read by a human being, so they can put you on their spam list for resale. Isn't that clever?

I almost fell for it, too, until I received a second similar complaint message using a similar format but different wording, from a different domain. Thankfully, both messages were easily caught by SpamAssassin, demonstrating that the spammers aren't that clever after all.

Alas, other messages are starting to get past using the exposed address I mentioned the other day, with SpamAssassin scores equivalent to legitimate email. I believe I'm even receiving spam via other addresses I've used. I could only guess that these email addresses ended up in people's address books, which were then harvested by computer viruses. So much for paritioning as an anti-spam measure. It looks like Bayesian trained anti-spam email clients are in my future after all.

May 21, 2003

Copy Protection is a Crime..., by David Weinberger. (David's blog.), clips from old TV shows and commercials spanning many decades. (Thanks Brad!)

May 20, 2003

Gigantic: A Tale of Two Johns, the They Might Be Giants documentary, will be shown eight cities over the next few months. Seattle is last: 8/15-8/21 at the Varsity.

Open Source Initiative's Position Paper on the SCO-vs.-IBM Complaint.

The BugBlog looks useful. It could be better organized, but it stays current and covers a lot of products I use regularly.

May 19, 2003

World's Largest Scavenger Hunt [PDF], "(performed on 22 Floréal 211 in Ida Noyes Hall and her environs, unless indicated otherwise)". From the University of Chicago. Slashdot story on last year's hunt.

Yay! The Elliott Bay Water Taxi is back for 2003. It was uncertain last year if the water taxi was going to be extended. (Thanks Anita.)

How to Manually Uninstall Programs from Windows XP. When the uninstaller doesn't work, sometimes you have to take the registry into your own hands.

May 17, 2003


As we have seen, in order for the iPod to work with Windows, it must be formatted using FAT32 (the Windows format). The new iPods all come formatted using HFS+ (the Mac OS X format). Windows users are expected to run an installation program that reformats the iPod with FAT32 and re-installs the iPod software. But the included software didn't work for me, and doesn't work for many others (though lots of people have had success with it). I could mount the iPod as a hard drive, and I could format it as FAT32 manually, but this wipes out the iPod software. I could connect it to a Mac and restore the software, but doing so reformats the drive with HFS+. What I needed was the ability to install the software on a FAT32-formatted iPod without reformatting.

Linux to the rescue! Using Linux and documentation for the GNUpod project, I was able to:

  1. restore the iPod using the iBook, resulting in a Mac-formatted drive with iPod v2.0 software,
  2. copy the v2.0 software off the iPod,
  3. repartition and format the iPod with FAT32, and
  4. copy the v2.0 software back onto the iPod.

Presto, a Windows-formatted iPod with v2.0 software! I can now mount and use the iPod with Linux and Open Source software, use the iPod with the Mac and iTunes 4, and even use the iPod with Windows and EphPod if I wanted to. (A Windows user should never have to touch the CD-ROM that came in the box, if they can get their iPod into this state.)

I wonder if the Mac version of the Updater can do a software update (not a restore) with a Windows-formatted iPod without having to reformat. I assume it does. If it doesn't, hopefully sites like will provide images of the latest version, which can be installed with more instructions from GNUpod.


A couple more iPod links and I'll try to shut up about it for a while: If your iPod clicks between tracks, you're not alone. I understand this to be an acknowledged problem for which Apple is working on a software upgrade to fix.

Also: iPod developer documentation, including dimensional diagrams for case designers and technical information on using the new "NotesOnly" (Museum) mode.

May 15, 2003

Apple tech support posts an informative document on getting the new iPods to work with Windows. These notes are better than the official knowledge base articles and describe my problems exactly, which appear to be common. In particular, it confirms some conclusions I made yesterday. The tech support message even has notes on how to deal with their misbehaving software. A choice quote: "MusicMatch Jukebox may not be the best program to use with the iPod."

Because Windows saw the iPod as a drive just fine, I went ahead and formatted it as FAT32. Of course, this obliterated the iPod software (which now has to be restored by the Updater), but I was mighty impressed that the iPod became a nifty tiny portable Firewire hard drive. Stores files and everything, and transfers mighty fast. Furthermore, this appears to have been the right thing to do.

Unfortunately, the support message doesn't seem to be enough in my case, so I've got my own sob story in the iPod for Windows Apple support forum. But it's nice to see that my problems appear to be common to some (so I'm not alone) but not common to all (so there are functional Windows set-ups out there that I can aspire to). That I have a Mac with which it is known to work, that it works as a Firewire hard drive with Windows just fine, and that Apple tech support is recommending existing third-party iPod software to work with the new iPods, are all encouraging enough that I don't mind so much that Giggy can't play tunes at the moment. (Yes, its name is Giggy. I reserve the right to name anything that costs me more than $500.)

[I plan on being especially verbose about my iPod problems for the sake of incoming Google searchers experiencing similar problems. There's a dearth of iPoddish blog material at the moment which I hope to rectify. Please bear with me.]

May 14, 2003

When my iPod arrived, I plugged it in and within minutes had copied an iBook full of music onto its little 30GB hard drive. The new touchy-touch buttons and touch wheel work even better than I expected, and I prefer them to the old iPod's turny wheel and click buttons. I quickly got the shiny metal back all greasy with my fingerprints, and the intentionally sub-par (I always assume) carrying case that's included will surely motivate me to get a real case when they come out in June. (L's old iPod never leaves its denim case, and looks and works fantastic.) And I love the cradle, as expected.

The fun stopped when I tried to use it with Windows XP. For one, the software installer couldn't detect the iPod to "configure" it (even though Windows's "Safely Remove Hardware" panel did). After the manager and MusicMatch were installed, the installer (now a mere rectangle on the task bar) wouldn't go away even after many minutes. So I killed it, and InstallShield complained that the installation failed. It could have been a buggy but successful installer, so I continued as if nothing bad happened.

Of course, like the installer, the iPod Manager failed to ever detect the properly connected device. Again, WinXP saw it and even tried to set it up with a drive letter, but iPod Manager had no reaction. I was able to get the Manager to tell me it sees an iPod, but only under weird circumstances, including once when it actually wasn't connected. Of course, it could never actually mount it, again despite Windows's otherwise normal detection. And then the Manager crashed a few times.

As implied by item 1 on Apple's iPod For Windows 2.0 advanced troubleshooting guide, even though the new iPod is labelled "Win/Mac" and purports to work interchangably between the two platforms, this only appears to be the case when the drive is formatted as a Windows (FAT32) drive. Ostensibly Macs can read and write the FAT32 format, but of course the new iPod ships formatted for Mac (HFS Plus), which only works on a Mac. All that music I copied from our iBook now had to be wiped so I could format with FAT32. The iPod Manager could supposedly do this for me, but since it wasn't helping, I decided to format the drive using Windows itself. I kicked off the format, and everything seemed fine, until the iPod's battery died.

Firewire, i.e. IEEE 1394, has two kinds of connectors. One kind, a tad shorter and a tad thicker than USB, as seen on most recent Macs, has 6 pins. Data travels on 4 pins, and 12 volts of luscious juice flow down the remaining two, powering Firewire devices and charging any rechargable batteries the device might want charged. Then there's the 4-pin IEEE 1394 connector, a teeny version found on PC laptops like mine. Converters are available to make a 6-pin cable work with a 4-pin jack, and indeed one is included with the new iPod. Before I got this new toy I asked myself, how could a connection standard that uses 6 pins possibly work over just 4? I'm usually proud to figure things out without having to be told, but learning the hard way is never easy.

Needless to say, I'm disappointed that my new iPod will not charge in the cradle when connected to my PC, and while I'm annoyed that iPod Manager 2.0 for Windows does not appear to work properly, I can't say I'm terribly surprised. Hopefully there will be bug fixes, and perhaps it will behave once I've manually formatted it for FAT32. I'll let you know. I'm otherwise quite thrilled with the thing, and hope to get a lot of use out of it.

Oh, one more thing: The iPod connector disconnects easily if you squeeze the sides, and not so easily if you just pull on it, or try to pry it off with your pen knife.

May 9, 2003 has schedules and links to streams in various formats of hundreds of public radio stations around the world (including BBC and CBC).

So You Want to Write a Book?, from O'Reilly and Associates.

Yet another note for C++ programmers: O'Reilly is coming out with a C++ Pocket Reference.

A little humor for C++ programmers: Interview with Bjarne Stroustrop.

The same joke has been made of C and Unix as well.

May 8, 2003

An explanation of leap seconds, from the Time Service Department, U.S. Naval Observatory.

Historically, the second was defined in terms of the rotation of the Earth as 1/86,400 of a mean solar day. In 1956, the International Committee for Weights and Measures, under the authority given it by the Tenth General Conference on Weights and Measures in 1954, defined the second in terms of the period of revolution of the Earth around the Sun for a particular epoch, because by then it had become recognized that the Earth's rotation was not sufficiently uniform as a standard of time. ...

The Earth is constantly undergoing a deceleration caused by the braking action of the tides. Through the use of ancient observations of eclipses, it is possible to determine the average deceleration of the Earth to be roughly 1.4 milliseconds per day per century. This deceleration causes the Earth's rotational time to slow with respect to the atomic clock time. Thus, the definition of the ephemeris second embodied in [Simon] Newcomb's motion of the Sun [described in his Tables of the Sun] was implicitly equal to the average mean solar second over the eighteenth and nineteenth centuries. Modern studies have indicated that the epoch at which the mean solar day was exactly 86,400 SI seconds was approximately 1820. This is also the approximate mean epoch of the observations analyzed by Newcomb, ranging in date from 1750 to 1892, that resulted in the definition of the mean solar day on the scale of Ephemeris Time. Before then, the mean solar day was shorter than 86,400 seconds and since then it has been longer than 86,400 seconds.

... Over the course of one year, the difference accumulates to almost one second, which is compensated by the insertion of a leap second into the scale of UTC with a current regularity of a little less than once per year. Other factors also affect the Earth, some in unpredictable ways, so that it is necessary to monitor the Earth's rotation continuously.

The historical list of leap seconds.

May 7, 2003

Indiana Jones will finally be on DVD by Christmas.

Speaking of long-awaited DVD releases, I finally got my replacement Back to the Future discs last week. So if you've been waiting to buy, the fixed version is probably in the stores by now. It'll be difficult to tell if you're getting one of the new ones or one of the old ones, so you'll have to watch BTTF II and look for the dry-coat-button gag to find out, I suppose.

MisterHouse, talking home automation in Perl.

ARRGH! I have received my first piece of spam to one of my personal, unpublished email addresses! I try to keep the addresses I use for personal correspondence out of the hands of commercial services, which means never putting those addresses on the web (a sure-fire way to get spam) or Usenet, and never using those addresses for services, accounts, and especially contests. Even services with sensible posted privacy policies could change their agreement in a desperate search for income, get bought up by a less scrupulous company, get hacked, or simply violate their agreement and hope nobody notices. And mailing lists that keep web archives but do not obscure email addresses, shame on you-- though a well-trafficked list might have a subscriber that collects addresses for resale, so any list probably deserves using its own address. (Personal domain owners like myself will often use as the email address for a service or mailing list, so if spam is ever received at that address, the channel can be closed and the appropriate companies can be sued.)

But I can't control what addresses my friends and family use with electronic greeting card and send-this-to-a-friend services. This alone makes address partitioning futile. Not everyone is as spam-conscious as I am, or even aware that the act of entering an email address into a web form from an untrusted source can be signing the address up on countless mailing lists. Nor would I expect to be able to convince everyone I correspond with to know and adhere to my wishes about how my addresses are to be used, to read privacy policies of sites as diligently as I would before entering in my address. Some people think spam is inevitable and protecting an address isn't worth it. Maybe they were right after all, though I've been 99.99% spam free with 0 false positives using address partitioning and SpamAssassin together for years (SpamAssassin primarily for emails to info@ and; the only reason I haven't been 100% spam free is because a few managed to slip pass SpamAssassin this way).

This leak, however, is probably not anyone else's fault (other than the spammer's). The address at which I received this particular spam is one Pine has been using for replies to messages (other than those to mailing lists), with the username at my webhost followed by my domain name. This could have been any of the dozen or so random web people that have contacted me personally for whatever reason, to which I've responded politely with an answer they've requested. Perhaps I slipped on an innocent-looking query from a person or service and hit reply without changing the reply address.

Even more aggravating is that this spam also slipped past SpamAssassin and into my inbox. With a rating of three stars (my current configuration is set to six), this little mail is quite ingeniously crafted and probably rigorously tested against the very software I'm using to block it. UCE clichés are either worded around (not "remove" but "removal"), or when specific to the actual product, munged with underscores (100%_money_back_guarantee). The product itself is quite innocuously named. Other clues seem obvious, however: it's HTML email, it contains an HTTP URL with an IP address instead of a domain name, and the header contains "X-Mailer: Shizzel Mailzer", for Pete's sake.

I, of course, feel silly spending so many words over one piece of spam when people receive hundreds if not thousands of unsolicited emails every day. But I've had a 100% success rate with 0 false positives for years, until now. Thankfully, I don't see any legitimate email scoring higher than 1 with SpamAssassin, so maybe I just need to lower the bar.

May 6, 2003

This to That tells you what kind of glue to use to attach one kind of thing to another. (Thanks Zannah.)

How to Dispute Credit Report Errors, a page from the Federal Trade Commision. Mostly linking for the names, addresses and phone numbers of the three major national credit bureaus (Equifax, Experian, and Trans Union).

P.S. Hooray for Google for managing to list all of these pages in the top 10 results for a search for "credit report". I expected a pile of junk insurmountable even by Google, but this is what PageRank does best.

I've been using Mozilla Calendar for calendaring lately, and while I'm not too crazy about it (it doesn't pop up reminders reliably), it seemed reasonable enough. I made the mistake of trying to upgrade to the latest version, and now I'm stuck without it due to this bug that's been around so long, that the latest version of Calendar has a special pop-up message describing the issue when it occurs. The problem for Linux users, it seems, is that the libical library used by some other programs, such as Ximian Evolution, is different from what Mozilla Calendar uses, and it won't be until the next version of libical that this will be resolved. Uninstalling Ximian Evolution completely and installing a fresh libical would probably fix it, but I'm not going to bother... *sigh*

May 5, 2003

When I blogged about jazz pianist Johnny Costa the other day, I felt I didn't give him enough name promotion. I gave him and his work for Mr. Rogers' Neighborhood credit for my love of piano jazz, and I linked to the official Johnny Costa web site. I didn't mention that four of his five albums appear to be purchasable at

I'd like to amend my previous mention by saying that I finally acquired Classic Costa, and it's incredible. I expected some fairly traditional stride piano jazz, but I didn't expect it to be so smooth, so original, so engaging, so elegant. Maybe it's just been a while since I've added a new jazz pianist to my collection, but I'm so wonderfully grateful to have it. I will be eagerly collecting the remaining few albums Costa has recorded.

Logo type, a database of logos. (Thanks Brad.)

I enjoyed seeing Brian Hatch's presentation, Linux: The Securable Operating System, at LinuxFest Northwest 2003 recently. Despite the 60-minute blitz through dozens of slides (only a little over half of them), the topics that did get covered solidified my understanding of some concepts I had only heard about in passing before. Brian is the primary author of Hacking Linux Exposed, and the talk made me very interested in buying the book. Brian also has a nifty Linux security newsletter.

I'm also grateful that Brian mentioned the name of the presentation software he used, which I admit I was admiring during the talk: MagicPoint for Linux. Not that I give many presentations these days, but some day...

May 3, 2003

If for whatever reason you wish to buy your new iPod from, here you go. As of this writing, a search for "iPod" doesn't bring these up, so I thought I'd mention it.

Also: New Apple store in Bellevue opens May 10.

May 2, 2003

Here's the Honda commercial that everyone is talking about [Flash 6 version]. I like it too. An article on the commercial proclaims that it's all real, and took 606 takes to film.

TiVo's been good about delivering ads and preview materials I might want to see into the main menu. Those BMW Films are kinda fun on the medium-size screen. How 'bout it, TiVo, can you deliver this one?

A review of Transgaming's WineX 3.0 and Point2Play.

Ooh: Perl enums.

May 1, 2003

Apple's big music-related announcement is a $.99/song download service and the new iPod. Slimmer, four new buttons, bigger hard drives, a dock, and USB 2.0, and the new models are at the same price points as the older models ($299, $399, $499). If I were to get one, I'm not sure if I'd prefer a Mac-formatted one to use with our iBook (with L has filled to the brim with her music), which would also give me Audible access, or if I should go for the PC format and try to use it with Linux. I could always wipe the drive and switch formats if I change my mind, so it's a workflow (playflow?) issue, mostly.

Update: Jerry (in comments) and this WSJ article have clued me in to the better Windows support of the new models; a single iPod works with both Windows and Mac interchangably. The article also mentions that USB 1.0 is supported as well as USB 2.0, which, while much slower, opens up many more compatibility options. (I knew USB 1.0 devices worked in USB 2.0 ports, but I didn't realize USB 2.0 devices work in USB 1.0 ports.)

Also: The new cradles have a line-out jack for piping audio from your iPod to your desk speakers, which pretty much clinches it for me.

Brad critiques AppleMusic. I had hoped that the recent practice of music publishers imprinting track data on the CDs themselves (recognized by any CD player capable of displaying it) would have cleaned up the mess made my CDDB users, but I know enough about reconciling vendor data to have little faith that official track data is any more consistent than a hodgepodge of different CDDB users.

CDDB is so bad that multiple discs in a boxed set often have different data styles from different people for each disc in the set, which means without intervention upon ripping, half of a set ends up in a folder for "Glenn Gould" and the other half ends up in a folder for "Bach." I, of course, have my own opinion on who the artist actually is in that scenario, but the wrong way worked for somebody well enough for them to share.

Why EMusic is better than AppleMusic.

For C++ developers: Smart Pointers - What, Why, Which? Smart pointers are so useful for memory management in C++ that it's amazing that there isn't an STL implementation yet. The Boost smart pointer implementation is a good common implementation, though not an official standard.

The C++ committee is in the process of updating the standard with a Technical Release for this year. Smart pointers are on the agenda, along with regular expressions, polymorphic function pointers, meta language features (type traits, typeof, an auto-type, perhaps even a meta language for better templates), and a bunch of other things I don't completely understand.