May 2003 Archives

Schools's out for two of my three kids. The other one gets out on June 9th.

I spent most of my K-12 days in California, where school didn't start until after Labor Day, and usually got out around June 15. Here in GA, public schools seems to start soon after August 15, and gets out before Memorial Day. I'm still not used to it.

Following up on my recent posts [1, 2] about rsync, here's a very useful trick with rsync involving multiple source directories. All credit for this tip goes to William LeFebvre, my former colleague from CNN.

James McGee ties together a series of excellent posts on knowledge managment

There's lots of meat in here; I spent a good 20+ minutes reading the posts James points to and following up on those. The best find was arguably Dave Pollard's weblog How To Save the World. I haven't come across this weblog before, but just browsing across his front page was most thought-provoking.

The World as a Blog is an excellent reason to put a Geourl on your weblog. (A geourl encodes your physical location.) The World as Blog checks weblogs.com every minute or so, and everytime it finds a post from a weblog that has a geourl, it puts puts a little textbox with a bit of the post, and highlights the location of the weblog on a map of the world.

A few minutes after I post this, for example, I should see my post show up. Let's see ...

I'm always interesting in reading about the process of writing. Particularly non-fiction. I've never been a big fiction reader, and when I write, I don't have enough imagination to make it up.

Once in a while, I'll pull out one of my favorite books on writing, read it, and get inspired to write again. Books like Bird by Bird by Anne Lamott, Writing to Learn or On Writing Well by William Zinsser, or almost anything by John McPhee. (The Control of Nature comes to mind.)

In that vein, Mark Bernstein has a wonderful essay called 10 Tips on Writing the Living Web. To be fair, it's not meant to be in the same league as the books, but it is inspiring around the topic of putting keyboard to editor. Nice work.

Following up on my note about the wonders of rsync, here's a very interesting article by Mike Rubel about using rsync to create snapshot-style backups. (Thanks to Jeremy Zawodny for the pointer.)

Filesystem snapshots are a useful notion. I first came across them with Network Appliance's Filer products; their WAFL file system. Snapshots give you access to backups of files in real time. In the Netapp boxes, snapshots work with a copy-on-write technique - that is, when a file in a snapshot is changed, the original file is left alone (and is still accessible from the snapshot), while the new changed data consumes new space in the filesystem.

Quoting the abstract from Mike Rubel's article:

More proof that everyone remembers more than I do. We had dinner with Kurt and Kathleen Eiselt over the weekend. I asked Kurt about his story about my strange notetaking habits. As it turns out, I apparently got the story all wrong.

He hadn't noticed me taking notes in class; it was my homework that caught his eye. And apparently what he noticed was my strange, randomly-colored TTL logic diagrams. He'd being going along, grading page after page of monocromatic logic diagrams, and here would come my work. It was done with done with one of those old Bic 4 color pens; he recognized them from his days at Disneyland, where he used to sell them. He spent quite a bit of time trying to figure out how I decided which color to use for different parts of the diagram, and finally decided it was random.

I was apparently quite amusing to him and the other TAs. When one of my assignments would end up in a pile for someone else, Kurt would grab it.

Listening to him tell the story, I tried to claim that my choices weren't random: I was simply providing examples of the four-color theorem. He didn't buy it.

Oh, well. I'm a software guy, not a hardware guy.

Last year I wrote a post in sun tea jars that mentioned the company Hearthmark of Muncie IN as the vendor of these jars. (Hearthmark was also mentioned as putting out the Golden Harvest line of glass canning products.)

Someone found that note in my weblog, and wrote asking if I'd ever found a phone number or an address. Doing simple web searches and even yellow and white pages searches turned up nothing, though I did find a reference to Hearthmark with a partial address of "HEARTHMARK CORP/345 S HIGH POBox2729 , MUNCIE/Indiana.

On a whim, I decided to see if I could do a trademark search for "Golden Harvest" or "Hearthmark." Bingo - I found a page suggesting that Hearthmark was now Alltrista. The street address matched what I had for Hearthmark, so this must be the one. It's not clear if Hearthmark changed names, or was bought, but there is an Alltrista web site that identifies it as part of the Jarden Corporation.

The most useful thing I came across is the US Patent Search database. It lets you search on the trademark name, owner name, and tells you something about whether the trademark is classed as "live" or "dead." And in this case, it looks like a useful technique for tracking down old brands.

It's been about two months since I built myself a new computer, so I figure it's time to brag a bit.

Paul Graham's essay on Hackers and Painters is making the rounds. It's a good essay, but I think an even better place to start is Tim Bray's own essay called Language Fermentation.

I'd approach Bray's essay bottom-up. Bray first references Graham, and then points to a very good essay by Bruce Eckel called Strong Typing vs. Strong Testing, followed finally by Are Dynamic Languages Going to Replace Static Languages? by Robert C. Martin. Read those three essays first, and then read Bray's take on them.

All four essays cluster around the general question about whether dynamically-typed languages like Python and Ruby (and Lisp, Graham's favorite) will beat out the old statically-typed stalwarts like C++ and Java. (Perl falls on the dynamically-typed side of the fence, but takes its lumps in these essays.) Moreover, Bray and Eckel come down on the side of saying having your compiler do 'testing' - for that's what static typing really is - is inferior to run-type tests that you create along side the code you're writing. This fits in nicely with Graham's musing, who says that programming is better thought of as sketching: working out your solution as you work, rather than thinking it all out in advance. Dynamic languages aid this process by keeping details such as type checking out of the way.

At 42, I'm old enough to remember a time when Big Box office supply stores didn't exist. I can still remember going into the local stationary store (as office supply stores used to be called) in San Leandro, CA, and just reveling in all the supplies.

If you like office supplies, one of the great pleasures is go in, find the pen aisle, and try out new pens. Places like Office {Depot|Max} aren't much good for this; they seem to stock pens without much flair or interest. Smaller stores are better for trying to find what's new in good in the world of cheap pens.

In Atlanta, the best place I've found for pens is Artlite. I first discovered them when I was considering buying a fountain pen, but it turns out they have a great selection of all kinds of cheap pens and pencils. $15 will get you a nice selection of low end writing instruments. ($15 won't even get you a clip on a nice fountain pen.) I'd tell you which ones I like, but that would mean I'd have to find them .. and as much as I like pens and pencils, they like to run and hide when I come near.

Along these lines, Mark Pilgrim came across a series of commentaries from Phil Agre about cheap pens. (link to Mark's article with the pointer). The commentaries are a few years old - ranging from late '97 to late '00 - but still interesting.

Our old friend Kurt Eiselt tells a story about me that involves pens. Kurt was in the computer science program at UC Irvine at the same time as me. I didn't know Kurt at the time, but he was the TA for a computer hardware class I was taking. At some point, he noticed that I came to class armed with an engineering notepad and a rainbow of Pilot Razorpoints: black, blue, red, purple, and green. Kurt noticed that I would periodically switch colors as I was taking notes. Growing curious, he spent considerable time trying to discern what my scheme was for choosing which color to use next.

Eventually, he concluded that I had no scheme. It was random. I switched colors for emphasis, and I picked colors based on the whim of the moment.

5/12/03 - I checked this story with Kurt, and I got the story all wrong.

One more pen story. A few years later, I was in a seminar that had something to do with AI and modeling knowledge about the world. I wasn't that interested in the subject, buy my girlfriend was; she was a Psych/Linguistics major, and she asked me to sit in with her. One day during the seminary, the professor walked over to my desk, picked up the purple Razorpoint and walked off with it. After a minute he returned it to me, and asked "Why didn't Paul protest when I took his pen? It's because of my special role as professor. No one else in the room could take Paul's pen without him protesting." At this, my girlfriend reached over and grabbed the pen. The professor grinned and bowed slightly.

We were married in 1981. She still takes my pens.