Last night, we went with the Landows to see Sly Fox, a fresh version of Volpone. It's that rare bird, a Broadway tryout in Boston, which means the production is dripping in familiar names. Written by Larry Gelbart (M*A*S*H), directed by Arthur Penn (Bonnie and Clyde), with Richard Dreyfus, René Auberjonois, Bronson Pichot, and Professor Irwin Corey. It's fluff, but it's magnificent fluff, and everything is done right. It's the story of a con in gold-rush San Francisco -- "that's GOLD, God with an L!"-- and it's headed to the Barrymore next month.

This is the second fine production we've seen this month; last week in Chicago we saw the final night of "The Light On The Piazza." That's a nifty musical by Adam Guettel, based on an Elizabeth Spencer novella from 1960. The story is a curious one to tell at this moment, an unsettling tale of a young southern girl who travels to postwar Italy with her mother and who falls in love with a boy to whom she can hardly speak. In 1960, this must have been about getting in touch with spirit and sex and leaving Eisenhower-era complacency behind us. Today, it means something else: our kids are damaged -- the Prozac generation -- but even so they can still love be be loved. Another fine, fine production -- the set was magnificent, the lighting a revalation, the singing fine. That night, I thought the music was nice, but I wasn't sure I heard much that isn't already clear in early Sondheim. And early Sondheim is, what, fifty years ago? Lahr wrote that

Guettel’s music and lyrics take nothing from the razzle-dazzle bargain basement of feeling; they represent, instead, a genuine expense of spirit. Rather than selling a cheap-and-cheerful redemption, the show offers only the prospect of repair. The Light in the Piazza doesn’t want to make theatregoers feel good; it wants to make them feel deeply.

That's exactly right, and if my friend the Composer is also right and if Guettel is simply reaching for musical effects he can't control, well, at least it's a lot better as a musical than Beauty and the Beast.

Feb 04 27 2004


Nathn Matias offers an intelligent review of Tinderbox at SitePoint.

It's like no content management system I've ever used..... Tinderbox is a truly open-ended writing tool. It can be used to build Websites, keep track of contact information, produce flowcharts, and visualize database records. And these are just the obvious uses.
Tinderbox, which uses XML as its native file format, stores information in "notes." These notes can be organized and linked many ways: in a hierarchy, in a two-dimensional map, or in any other structure you think of. It's as simple as drag and drop.

One nice thing about Matias's review is that he describes a variety of good ways to use Tinderbox for getting work done. It's common for magazine reviewers to limit themselves to working through the tutorials or trying the obvious, familiar tasks, but Matias does a lot of real work with Tinderbox. The color scheme of the example map view in the first screen shot, for example, shows care and flair. The discussion of different tasks -- taking notes on historical events, building a recreational web site -- makes the review worth reading, even for experienced Tinderbox users. It's not merely a consumer guide to making the purchase decision.

It's also nice to see that Matias gets the point -- even though he's not naturally inclined to the Tinderbox way.

Tinderbox takes time to understand. It's unlike any piece of software you've ever used, and, unless you have time to learn it, Tinderbox might not be the software for you. On the bright side, the process of learning Tinderbox is also unique. Much of the learning process involves freeing the mind from the beaten-down assumptions and flat limitations we get from using word processors.

He describes himself as " a hardcore free/open source software fanatic:, but concludes in the end that "this is one of the few software packages I believe is well worth the money."

Diane Greco makes an important observation (Feb 25: no permalink) about the nature of blogging.

By and large, the blogs tell success stories. They have to -- blogging as a literary form supports the idea of eventual success. When there's bad news from the bathroom scale, the open-endedness of blogging makes it possible to cast the gain as just a temporary setback, not a failure. Diet blogging recasts or reimagines the yo-yo effects of a diet as a surface, a space, a site for potentially endless re-inscription. Dieting as Etch-a-Sketch, very postmodern.
Another point about the rhetoric of diet blogs. When their weight changes, either up or down, most diet bloggers will see the change as something to be explained, and will search for reasons in their behavior. There is something quasi-religious about this.

This tells you more about the nature of weblogs as a creative medium than you'll find in a dozen newspaper stories. And Greco's keen observation on the narrativity of weblogs is obviously a key to understanding why weblogs are what they are -- and I don't believe this specific point appears anywhere in the major introductions to weblogs.

"Business is dramatic and business is everywhere, but in new media, business has been neither." That's my lead in Conflicts and Interest, a survey in the latest TEKKA of the remarkable failures of business games.

Let's face is: game studies are going to stay stuck until we have lots of real criticism of real games. Not just box-office predictions, not just the gossip of industry crafts ("great special effect! amazing costumes! the best gaffer in the business!"), but sitting down to look at what games are trying to do, and where their means are not aligned with their ends.

That we woke up one morning and decided to start a toilet factory seems a nightmare, but what must the game team have thought when they were handed this assignment? The writers (and translators) were, of course, nonplussed; it's tough to write a business game that will travel, but writing about toilets is even worse....
"Factory Mogul can't take itself seriously and so tries to defuse the audience's justified rage by acting the clown. Modest wit can work marvelously when a work intelligently critiques its own failings. Greg Costikyan's Paranoia, for example, hilariously resolved the contradiction in role-playing games between the heroism of characters facing mortal danger and the quotidian reality of tabletop gaming by making each character a member of a clonable family. When disaster strikes and the player's decisions prove calamitous, "Send in the clones!". But there's no ironic critique in Factory Mogul, nor any hint of slapstick satire. The game takes pratfalls in a desperately irrelevant effort to please or, at any rate, to placate."

The difference between a silly disaster like Factory Mogul and a really funny game like Paranoia is not a question of adding money or game mechanics or a better game engine, it's a matter of getting the tone and timing exactly right. Being able to recognize this is hard, and it's probably even harder when the disk duplicator is calling every day for the master, when the developers are wrestling with the code and the marketers are screaming at the developers and nobody has seen Quality Assurance since last Tuesday.

But, if we can't get a handle about this, then what exactly can we do with all those monographs about narrative and ludology and genre and gender?

Update: Ian Bogost responds that, 'before research can be useful for developers, we need to see more inspiration in the business. We need people who have something they want to say, and who want to say it in the medium of the videogame.'

Feb 04 25 2004


I'm going to be giving one of the keynotes at BlogTalk 2, July 5-6, in Vienna. Mena and Ben Trott (Moveable Type) will also be keynoting. Torill Mortensen, too! Last year's conference notes look interesting.

I've received six or more spam or quasi-spam ads for Mel Gibson's Passion, or tie-ins trying to leverage it. Including ads from Amazon and PalmGear .

One reason I noticed is that it's so rare to see spam that has any chance of succeeding. Almost all spam seems to advertise offers that only a fool would accept; why wouldn't spammers occasionally advertise something that you might possibly want to buy?

That's one reason the Passion spam is surprising: I probably won't see Passion, but it's not as impossible as half the spam suggestions I receive. Passion sounds too violent for me. Ebert, who liked it, writes:

The movie is 126 minutes long, and I would guess that at least 100 of those minutes, maybe more, are concerned specifically and graphically with the details of the torture and death of Jesus. This is the most violent film I have ever seen.

The experience of watching this movie sounds so unpleasant that I don't think I shall -- even though I'd like to judge for myself on the subject of the film's anti-semitism.

Still, it's not inconceivable that I'd see this movie, so it's surprising to get spam about it. Conversely, it's quite possible that people will be offended to receive ads suggesting they see it, which makes one wonder what Palm and Amazon were thinking.

Feb 04 24 2004


Andrew Sullivan stands up and acts like a real conservative:

WAR IS DECLARED: The president launched a war today against the civil rights of gay citizens and their families. And just as importantly, he launched a war to defile the most sacred document in the land.
...This struggle is hard but it is also easy. The president has made it easy. He's a simple man and he divides the world into friends and foes. He has now made a whole group of Americans - and their families and their friends - his enemy. We have no alternative but to defend ourselves and our families from this attack. And we will.

This is something I've been wondering for the last few months. It seems there's a decent chance that good sense and justice may prevail once more in the US, and when it has done so, we will all look back once again and point to some defining moment. "That we here highly resolve that these dead shall not have died in vain," in 1863. "At long last, sir, have you no sense of decency? Have you no shame?", in 1954.

Perhaps, this time, it'll be written in a weblog.

Feb 04 20 2004


I don't know what the greatest line in Meiggs might be -- and I haven't read nearly all of Meiggs. But I believe I've read a decent share of Samuel Eliot Morison, the paragon of naval historians, and his excursus at the end of "The Battle of Surigao Strait" (Leyte: History of US Naval Operations in World War II, volume 12, pp. 140-141) has been much on my mind lately. Historians often chronicle the first time, and we rarely know much about the last; here is one exception.

The Battle of Surigao Strait marks the end of an era in naval warfare. It was the last naval battle in which air power played no part, except in the pursuit. It was the last engagement of a battle line. Here an old sailor may engage in a little sentiment.
....Thus, when Mississippi discharged her twelve 14-inch guns at Yamashiro, at a range of 19,790 yards, at 0408 October 25, 1944, she was not only giving that battleship the coup de grâce, but firing a funeral salute to a finished era of naval warfare. One can imagine the ghosts of all the great admirals from Raleigh to Jellicoe standing at attention as Battle Line went into oblivion, along with the Greek phalanx, the Spanish wall of pikemen, the English longbow and the row-galley tactics of Salamis and Lepanto.

The romance works because Morison almost never indulges himself this way. His prose is crisp, clear, and detailed, preserving the data (19,790 yards) and avoiding embellishment, yet somehow still keeping the reader engaged and, as we should be, appalled.

It's impressive to find how much nicer a Tinderbox outline looks if you set the NameFont to Optima instead of the default (Geneva). Or Helvetica Neue, Eurostile. Or our old friend, Fabrizio Schiavi's screen-optimized font Sys.

One important aspect of tools like the daybook (or like a moleskine notepad, for that matter) is that you want them with you, all the time.

Another is that you want to know you'll never lose them.

That means you need to keep copies in several places, just in case something happens. I'm playing with ChronoSync, a nice little $30 application that makes it easy to synchronize a briefcase folder on my laptop with a corresponding folder on my office desktop. It's not going to displace Retrospect for backup, but it's great for this task.

This week marks the second anniversary of Tinderbox. Wheee.

I'm very happy to say that the Tinderbox community is large and growing, and that we've got lots of positive energy. My mailbox (and NetNewsWire window) has been overflowing with comments on the Daybook note earlier this week; Daybook is just about the simplest Tinderbox task you could imagine, but it also seems to be a frontier of knowledge management.

Save the date: we're planning a Tinderbox weekend in Boston, May 22-23. We'll have training -- from Getting Started basics to advanced template design. We'll take a look at the roadmap for Tinderbox for Windows and Tinderbox 3. We'll look at a wide variety of applications. More details, soon.

Feb 04 19 2004

LA Mac Help

We need some Macintosh help in the Los Angeles area -- someone who could, for example, move files from an old Powerbook to a new one. Windows/Mac networking experience a plus. Decent pay, gratitude, cookies. Email me.

I'm reading a fascinating book that tries to understand the archaeological evidence of ancient Ostia in terms of the Roman legal code.

There are strange gaps in our knowledge, things that really matter to us but that nobody thought to write down. The Romans had lots of slaves: where did they sleep? Where did they keep their stuff? We don't know.

But there are some things we do know, because real estate law was always a nightmare. We can figure out, for example, that the firemen owned some shops and rented them out, because the shops would not have been up-to-code unless they and the firehouse were jointly owned.

Ancient Laws

In Rome, property rights extended from the ground to the bottom of the sky. Except, if your neighbor had an ocean view, it was illegal to block it.

Lots of people are interested in the Daybook.

William Cole is intrigued, and he's just one of several people who wonder (a) is it real, or theory, and (b) can they have a sample document?

It's not immediately clear to me whether he is talking about an actual product made in/for Tinderbox, or if he is just theorizing.

Yes, it's real. And the version I described is such a simple Tinderbox document that there's not much point in distributing a prototype. Make a note for today: voila! Tomorrow, make another note. Now, make a container to hold both those notes. Perhaps sort by date. That's all you need to get started.

Of course, Tinderbox could do a lot more. How much additional work is going to be useful? We don't know yet! Jon Buscall has his own Daybook, a more elaborate system, and he's not quite sure whether the simpler version would be better or worse.

Knowledge workers frequently need to handle information that isn't routine and that doesn't fit established categories. Some of this information is glamorous:

While you were out: Hey, boss -- Comcast called, they want to buy the company.

but most is mundane. You print out the e-ticket receipt for next month's Chicago trip: now where do you put it? How hard will it be to find next month?

The ideal, of course, is to have a comprehensive filing system and to use it religiously and systematically. In practice, though, finding things notoriously depends on remembering where you put them. As Engelbart famously observed, the whole point of software is prosthetic: finding ways to help people do things they at which they don't naturally excel. If you always remember where you put things, good files are easy but good piles will do fine, too. If you don't, you'll file those tickets away and next month you'll spend hours trying to remember where you filed them.

The Daybook is a very simple Tinderbox application that contains a fresh entry each day. It's kept in reverse chronological order, like a weblog. But the Daybook is meant to be private -- shared with your assistants, perhaps with your workgroup, but chiefly a personal note space. -- shared with your assistants, perhaps with your workgroup, but chiefly a personal note space.

Daybook: a new Knowledge Management domain

In the Daybook, we record informal notes about decisions made during the day, as they are made.

E-tickets for Chicago trip: filed under Travel Plans.

New disk arrived. Formatted for FAT32 and attached to Prosperity.

These are too boring for publication -- in fact, many of these are going to be write-only. Nor are they journal entries intended for personal enlightenment. Instead, these are short, random notes for which we might someday need to search.

Search is the key. Fast search means that, when we've got to a plane to catch and can't find the tickets, we can look in the Daybook and search for "tickets" or "Chicago" and not wait so long that we'll miss our plane anyway. Regular Expressions and boolean search mean we can look for things even when we aren't sure exactly what to look for. And incremental search gives us immediate feedback, so we know in seconds whether our search is going to turn up anything.

Can this work? It requires discipline. First, you've got to write it down; if you don't write in the Daybook, you can't find it later. Second, you've got to limit your Daybook notes to actions with consequences that you might need to review later. The Daybook lists what you acquired or installed, who you contacted, how you fixed a problem, where you put the widget. Plans, ideas, inspirations, appointments, and everything routine -- all of that goes somewhere else.

The reason for this discipline is that we want to keep the Daybook compact, so search will have some leverage, and so browsing among search results is practical. If we add a score of notes a day -- 6000 notes a year -- this is manageable. When you start working with 600,000 notes, as you would if you tried to put everything into One Big File, then neither search nor spatial hypertext gives you any traction.

PDAs are best at organizing appointments; if you try to use them to hold everything, the software can't manage the load.

Something like the Daybooks is a component of some CRM tools -- a sector I've not followed closely. But a good deal of CRM, it seems to me, is about enforcing a process through time: following up leads, sending birthday greetings. The Daybook records what you did, never what you should do.

Daybooks are tangential to the sort of information triage tasks that inspire much spatial hypertext work.

This sort of informal knowledge management tool has not been a common target for software innovation, in part because it's unglamorous and its ROI will be hard to measure. I do suspect it's going to be very useful to a lot of people over the long haul. For example, it's exactly what you need to help keep track of the myriad things that crop up when you're using the network to organize for a cause.

Feb 04 17 2004

Might Have Been

Imagine what the tech side of the blogosphere would be like today if, when Atom kicked off, the Atom folks had felt strongly that the new standard should minimize disruption and avoid hurt feelings -- even the feelings of people they might not want to invite to dinner.

One of the interesting questions about open source as a movement is, what do you do when the interests of the code conflict with the emotional interests of a committed contributor? In the business world, sometimes you just have no choice but to call someone into your office and say something like:

I know that you and Blodget have issues, and I know that Blodgett is our competitor. But we've decided the antagonism is doing us a lot of damage.
So, here's the keys to the company condo in the Carribean. Here are your plane tickets. Go down there, watch the sun set over the ocean, relax, have a drink. Next Monday, come back tell me whether you think you can swallow hard and become a Blodgett supporter, or if you've decided to change jobs.
Might Have Been

One of the key rules of running a volunteer organization is that yes, you can fire a volunteer. But it's hard, and I don't know that it would work in open source or if you'd just generate a fork every time you tried.

Linda and I have been watching the first season of 24. We missed it first time round, but when hypertext wiz Leslie Carr said we had to see it, well, we had to see it.

I think the writers on this show may have discovered the precise limit to the quantity of plot you can pack into 44 minutes without breaking anything.

Feb 04 13 2004

More on Atom

With all the talk about openness and innovation, wouldn't you think there'd be an open source, public utility by now that, given an Atom feed, would supply you with an RSS equivalent? That way, if you wanted to offer a link to an RSS feed and your tools generated Atom, you could just write a link

http: //

and you'd have an RSS feed, too. Sure, there'd be some impedance mismatch, but you could get Close Enough pretty easily.

Aaron Swartz has pointed out to me that some XSLT utilities, such as Aaron Cope's, do exist. But none are well publicized, and I'm not sure their public implementations inspire much confidence that the service they provide won't suddenly break, or vanish. It's not as if it's baked into Google or Blogger.

Instead, we're going to have a format war. James Robertson writes:

It's been my contention all along that Atom is nothing but a tax on aggregator developers; it serves no useful purpose as a syndication format, but does manage to create tons of extra work (with more to come - the spec is only 0.3, and people are implementing). Wait awhile, and we'll have as many versions of Atom - and of the Atom API - as we have of RSS.

Dare Obasanjo figures that, with various people rushing to deploy the Atom 0.3 draft, we're likely to end up with a bunch of different version of Atom in parallel, indefinitely.

It's amazing how geeks can turn the simplest things into such a mess. I'm definitely going to sit it out until the IETF Atom 1.0 syndication format spec before spending any time working on this for RSS Bandit.

Also: Dwight Shih:

When I say Google owes us an explanation, I mean that Google needs to demonstrate good intent if they want to portray themselves as doing no evil. I am for applications, not against Atom. Breaking current applications is bad. Providing new applications is good. Where are my new applications?

You'd think a company like audible would make it easy to link to specific titles. Their CMS gives you messy URLs, but that's OK; one reason Tinderbox macros are nice is that they hide the complexity of things like amazon urls.

But, no -- the only way to link to audible titles, according to their tech support people, is to sign up for their affiliate program.

To do that, you just need to agree to a long document and sign up on this web page. But the javascript on this page is broken; it won't accept that my phone number has ten digits, even though it quite clearly does. Another long, long call with audible tech support yields no answer.

I'd have expected that whoever runs the affiliate program would want to know this right away, but that doesn't seem to be the way things work at audible. In any case, when I suggested that whoever manages the affiliate program might want to talk to me, they gave me the number of the PR director. Strange kind of org chart.

The PR director was not, in fact, the right contact. But at least he knew who could solve the problem, and once he got involved everyone was very nice.

Jonathan Peterson suggests that Apple's strategy is now steered more by knowledge work than by its visual heritage:

Less obviously (and more recently), the Mac is becoming very compelling for some of the knowledge work tools that are currently Mac-only: Spring, Tinderbox, OmniGraffle, OmniOutliner.

More response to the TEKKA piece on My Life With Master: Emily Dresner-Thornber.

Feb 04 12 2004

More Atom

If the Atom rhetoric about openness and innovation is sincere, wouldn't you think there'd be an open source, public utility by now that, given an Atom feed, would supply you with an RSS equivalent? That way, if you wanted to offer a link to an RSS feed and your tools generated Atom, you could just write a link

and you'd have an RSS feed to. Sure, there'd be some impedance mismatch, but you could get Close Enough pretty easily.

Instead, we're going to have a format war and more food fights.

In what is said to be the best book on the Microsoft Foundation Classes , the discussion of OLE begins at chapter 18 on page 1061.

Ninety-one pages later -- on the last page of the following chapter -- Jeff Prosise gets around to mentioning that, before you do anything else, you've got to call AfxOleInit().

If you write an application and find that calls to DoDragDrop or other OLE functions mysteriously fail, make sure that AfxOleInit is called when the application starts up. You'll save yourself a lot of grief.

Now, they tell me!

My Windows laptop, Progress, has been told by Microsoft that it wants to fix the Bookshelf Symbol 7 font to "remove unacceptable characters."

The characters, presumably, are a pair of swastikas. The other glyphs in this strange little font include a Star of David, esoteric accented characters (e.g. capitalized consonants with accents), hearts and spades (but not clubs or diamonds), and eight Chinese characters.

I bet there's a story here....

Massachusetts politics are in an uproar right now. It's a complicated story.

I wrote my state senator and my state rep brief emails. Neither has replied.

This is weird, and inefficient. When I was in school, my friends and I spent afternoons and evenings working for reform candidates in Chicago. The city, in those years, was controlled by a staunchly conservative Democratic machine of legendary power; in a few years, of course, that machine was a memory. Back then, when I wrote typewritten letters to Washington, my letters were answered by (the staff of) Sid Yates and Chuck Percy and Everett Dirksen and Adlai Stevenson.

None of their staffers had email, or Word, or databases, or Web search; they did all this with typewriters and paper files. They managed to reply, despite much larger constituencies than the state representative for my town.

All politics is local. The problem with the Kennedy's, LBJ said, is that none of them ran for county sheriff. Look at what Dean accomplished. You'd think they'd get this right.

The syndication war is once again in fyll swing. Google has switched its users from RSS to Atom. Winer explains this as the first step toward introducing a new Google client; by moving lots of users away from RSS, Google ensures plenty of feeds will work with their new software but not with Radio Userland and other software that's already deployed. covers the story in what Winer describes as a food fight.

Bad for users, and very bad for developers. (Have you noticed how little new syndication software is showing up lately?)

Feb 04 10 2004

Girl, Earl

A striking portrait turns out to be the third Earl of Southampton, wearing a dress. (pdf) Interesting ramifications for Shakespeare, Marlowe, and more. From the Guardian. Thanks, Meryl!

Girl, Earl

When Linda mentioned this discovery over dinner, my comeback was, "Well, someone had to be the first Juliet." Her question: "But was he acting?"

One thing they don't teach you in school is, what to do at a conference. Some people never seem to figure this out, and in fields (like hypertext research) where conferences are central, that's a real problem. Susan Karlin, who is both an entertainment writer and a tech journalist, contributes some fine tips in the new Tekka.

I learned the essentials of conference work at DuPont, chiefly from observing a complete stranger, Alan Huang (then at AT&T) at a Gordon Conference on optical computing. The key lessons all came during an evening discussion in the lounge of the residence hall, in which Huang and a senior manager from IBM went toe-to-toe on the direction of research in the field. Important work happens away from the auditorium: that's the first lesson. Huang crammed 60-odd slides into his 20-minute talk, an extraordinary performance that captivated everyone and that packed far more information into each minute than the expensively-prepared canned presentations we made at DuPont. I only saw Huang speak once, but I went home and completely reformed by speaking style. And Huang, who cannot have been much older than I, was already talking about his responsibility for the direction people were taking in the field, for not leading people into barren areas because they were currently fashionable. That responsibility was breathtaking: here I was, all outfitted with my nifty doctorate and my office at the Experimental Station, and nobody had told me, "It's time to stop thinking like a student."

In the new TEKKA, I review Paul Czege's role-playing game, My Life With Master as a hypertext tool. That's a surprise, in a way: we normally think of tools as software programs like Flash and Tinderbox and Code Warrior, not as games. But I think this could indeed be a useful tool for writers looking for an exercise or a way to leap over a block:

If we sit down to play a solitaire game of My Life With Master, the framework poses a series of writing problems, inviting us to sketch each scene effectively and economically. The framework takes care of some details of plotting, letting us focus on how we reach the inevitable dénouement, and also forces us to braid together a variety of plot lines as individual minions pursue their individually tormented existences, now aiding a companion, now luring away or destroying those their fellow minions cherish.

Anja Rau picks up a thread by game designer Greg Costikyan, to ask whether closure is tragic:

So, is it tragic if an RPG (or, for that matter, a hyperfiction) ends? It should be old news that the absence of beginng and/or end automatically creates an "open text" as imagined by U. Eco. But is endlessness a tantamout to quality? Is closure tragic?

Sure, closure is a suspect quality. But I don't believe this is a problem, much less a tragic circumstance. In Lost In Translation, we always know that the central characters will go home and go on with their lives. This isn't a Charade or a Roman Holiday. We don't build suspense in that direction; instead, we know how it ends, and we want to find out how we get there.

Tekka Tools: My Life With Master

Ed asked, sensibly, whether updates really are more likely to help than hurt. After all, we all have stories of updates that causes all sorts of problems! Does it make sense to stay up to date, or should you wait to jump until things break?

Here's a quick-and-dirty study: the last few things I've upgraded, and a rough idea of what I experienced.

  • Tinderbox: OK, I'm prejudiced. But I use experimental builds for everything -- including this weblog (1623 notes) and TEKKA. If I'd burned my fingers doing this, I'd surely be using release code to get work done. I don't, because new features (like the new list-making feature I'm using here) matter more.
  • Snapz Pro X. This update hasn't mattered, yet. The previous update was needed to get the widget to load reliable with Panther.
  • Retrospect 6. No surprises. Little glitches relating to network problems seem to have gone away. Backing up the new Windows laptop turned out to be easy.
  • Safari 1.2: A minor drag-and-drop oddity with Tinderbox needs investigation. Otherwise, no problems. No observed changes, either.
  • Eudora 6: Spam filter was a big win for several months, though the spammers are now learning to evade it. Windows version seems solid, reliable, and allegedly more secure than Outlook.
  • BBEdit 7: the built-in HTML preview is fine. No complaints.
  • NetNewsWire: small cosmetic improvements are fine. No problems under the hood.
  • Panorama 4: esoteric user interface ideas from old version have now been replaced; old habits cause confusion, but the new interface is better.
  • Office 2003/Win: it is what it is. Installation went better than expected.
  • Stuffit 7: mixed results. Good news -- better icons, better contextual menu in finder, said to fix some MacOS X issues, Bad news, I've got some esoteric headache that launches the wrong engine when I try to unstuff. Or something like that. Slightly munged, but I can live with it.
  • iTunes and iPod software: I updated with some trepidation, but everything went fine. Some iTunes upgrades have effectively removed features in the interest of security or closing loopholes, but on balance the application improves.
  • Airport: I upgraded with great trepidation, since it's a firmware update and since the Eastgate airport network is now absolutely critical to getting anything done. But it worked fine, and the network has been rock-solid since the update.
  • Railroad Tycoon III 1.0.3: tricky installation glitches, but once the installer actually ran, everything seemed fine.

That's the latest batch. No big wins, unless one of these routine updates happened to avoid a big loss I won't even notice now. But no disasters, and some everyday improvements.

There's a new version of Ambrosia's screen capture utility for Macintosh, SnapZ. 

It's a really good idea to upgrade your applications often. New versions are almost always better -- faster, cleaner, richer, and more reliable. Depending on stale software is risky, and eventually it leaves you high and dry.

The new issue of Tekka leads off with a fascinating look at cybercafe society in Nairobi. Playwright John Makeni takes us along on his morning rounds:

A small food counter offers coffee, tea, and mandazi. a Kenyan cake. The middle-aged man who is working next to me orders a cup of tea. I also have an urge to order a cup of tea, but I don't have enough money, so I forget about it. As more clients come in, more attendants also arrive. There are now seven: four women and three men, all clad in blue t-shirts that read LAZARDS CYBER CAFE. It's a fairly good job.

Makeni comes from a village in near Busia where

There is hardly a telephone booth. There are no city buses, and the occasional matatus are not as fancy as those in Nairobi. To get around, people walk or bicycle, or if they have no bicycle they take a bodaboda, a bicycle taxi, for about ksh 20. You could easily count all telephone booths in Busia Town, but today there are also several internet cafés.

In the 90's, it was easy to think that the net was Palo Alto, Soho, with a few outposts in Tokyo, London, and Stockholm office towers. It's a bigger, richer, and stranger world than we knew.

A useful $10 eBook: Take Control of File Sharing in Panther by Glenn Fleishmann.

Feb 04 5 2004

Nothing Done

I've managed to get nothing at all done today.

I've shipped some Tinderboxes to eager customers, as our office manager is unwell. I've fixed a couple of minor Tinderbox/Mac problems for other customers. I've cleared the accounting backlog that built up when I got nothing done yesterday, either. I've tried to take care of my ailing mother, and tried to find my AWOL kid sister, and tried to figure out whether my president was AWOL or what. I've tried to convince my state senator not to turn state politics upside down until 2006 in order to overturn gay marriage, and tried to make some suggestions about the future of RSS. I've tried to study the complex technical issues of whether Tekka lanyards would be better than Tekka baseball caps, tote bags, or coffee mugs.

I've backed up Progress, the shiny new Windows laptop, with the new Retrospect 6 upgrade. I've tried to explain to a Tinderbox user from London that upgrading is a really good idea. I've tried to explain to a student from Germany that those kids aren't the problem and that the sensuality of images and sounds in electronic texts won't corrupt them forever. I've tried to explain to a journal editor that, if you say you're peer-reviewed, you have to ensure that competent peer review takes place.

I've engaged in real-time, intercontinental exegesis of Jill's latest weblog entry. I've tried to find a way to attend an interesting meeting in Barcelona, to visit some colleagues in Texas, to visit my ailing mom, and to get dinner on Valentine's Day (which is also my 30th anniversary, though the wedding came much later). Here we are in a foul economy and I can't get a reservation at a fancy restaurant ten days in advance. Two Americas indeed.

I've learned one of our servers was down. In consequence, I learned, our email was down, as were our main Web site, our revenues, and my mood. I've learned some interesting facts about referrer spam. I've learned why armored bears don't get lonely . I've learned how to tell the ruins of a Roman grain warehouse from a Roman wine wholesaler: the jars for storing wine were buried to keep them cool. I plan to need to know this next month and the clock is ticking.

But I've not managed to write a Windows version of the Tinderbox DrawingPolicy class, and that's the first thing I've got to get done today.

Nothing Done
Interior, H. H. Richardson's Austin Hall, Harvard.

A new issue of TEKKA is just out, by the way. I'll be talking a lot about it in the next few days. Some really good stuff: see for yourself.

From this morning's Boston Globe:

In the fictitious village of Eastgate, soldiers from the 11th Armored Cavalry Regiment portrayed villagers used as human shields late last month. The training exercise was played out at the National Training Center at Fort Irwin in California's Mojave Desert.

I'm testing some new RSS tools in Tinderbox. For the next few hours, the RSS feed may break in surprising ways.

Much better now. For the present, I've included the entire post in my RSS feed instead of a brief summary. If you find this better, or worse, let me know.

Feb 04 3 2004

Don't do this

Yesterday, I had the unpleasant task of reviewing a study that had taken months and months of hard work by several professors and a bundle of students and that wound up, after all this effort, with less than nothing. I hate to lose a day trying to explain to people that their work for the past year is unpublishable. This problem crops up again and again, and I'm tired of breaking the bad news. It's easy to understand the problem.

Let's say you're working on a new software widget, and you want to demonstrate that it's useful. A very common approach is the following:

  • We recruit the 20-odd students from our CompSci 301 class to be our test subjects.
  • The students are randomly assigned to two groups. One group uses the Wonderful Widget all semester, the other group uses the plain old widget.
  • We measure some good outcome -- the number of students who ace the course, the number of epiphanies, whatever. We hope the Wonderful Widget group will get more good results. For simplicity, we'll call the good result a "win".

Now, it's good to actually do the experiment to confirm the widget's value. But this experiment is probably doomed: if the result is going to be statistically significant with 20 subjects, the widget is so good that its value will be obvious from one or two subjects -- or from simple inspection.

Let's say that, over the course of the experiment, we expect about half the students to have a good outcome: the control group should have about 5 wins, and we hope the group with the widget has lots more. How many wins do we need to be confident the widget is good?

Every single subject needs to win! As a rule of thumb, if you expect n wins, pure chance accounts for about sqrt(n) wins more or less. So, if the control group gets 5 wins this time, next time it might well get 3, or 7: to be confident the widget is better, you need about ten wins in the widget group.

Many people have excellent statistical intuition in sports. Take baseball: imagine a team in spring training that's a solid contender, a team with a good chance to make the playoffs. In other words, you're envisioning a team that you expect to win about 90 games. The Oakland A's, or the San Francisco Giants.

Now, if this team turns out to go 95-67, are we shocked? Not at all! It's a long season, that's why they play so many games. On the other hand, if this team that you expected to be competitive actually winds up having one of the best seasons in history, going 110-52, then you are surprised -- something you didn't expect happened. A swing of 5 or 10 wins, we know, might well be luck in a 162 game season; a swing of 20 needs an explanation.

Take a veteran left-fielder who usually hits .300 with 25 home runs. You expect him to get a hit about three times in every ten official at bats; if he bats 500 times in a season, you expect about 150 hits. The square root of 150 is about 12. If he finishes next year with 138 hits (.276) and 20 HR, you say 'He had an off year, but it's probably just bad breaks.' But if he finishes with 126 hits (.252) and 15 HR, you're pretty sure something is wrong. And if he gets 174 hits (.348) with 35 homers, everyone is going to be amazed. Same thing with pitchers: if a #3 starter who usually has nine or ten wins every year gets twelve wins, it could easily be good luck. If he gets 15, you're pretty sure something changed.

Can Senegal win a World Cup game against, say, France? Sure. Do you fancy them to win it all? Not likely.

Same thing with American football: if you expect n wins, then a difference of sqrt(n) wins is not remarkable. Last year, the Patriots went 9-7. This year, people expected them to be a little better, so pencil in 10-6. They actually went 14-2 (and won the Super Bowl). Next year, if they win 10 games instead of 14, their fans will be sad but not shocked. If they go 7-9, people will all be asking, 'what happened?'

When you set out to prove that your widget is best, the sqrt(n) rule is a terrific rule of thumb to keep in mind. An empirical study is a terrible thing to waste: be sure you have enough subjects.

Feb 04 2 2004


Ted Goranson's exemplary study of outliners continues with a close look at the development of the disclosure triangle, the note icon, drag feedback, and other user interface widgets.