Apr 04 15 2004

Big screens

Tim Bray has a new job at Sun, and it comes with a big 1920x1200 flat panel display to plug into his TiBook.

...a whole lot o’ pixels. I haven’t really figured out how to use ’em all effectively just yet, but it’s a nice problem to have. Mind you, when I lean back at my chair and look up at this horking thing, it occurs to me to say “On-screen!” in a clipped English accent and see if I can get the captain of that Romulan Warbird up there.

How much screen is enough? Perhaps 8'x6' at 72dpi? We spend a lot of time in the software world, talking about clever visualization strategies -- 3D, VR, fisheye views, animation, spatial hypertext. How will really big displays change this?

Comments don't belong in weblogs.

The measured pace of weblog response, and the distance between rival weblogs, makes measured discourse possible. Comments let idiots deface your weblog, and that's intolerable. Because you can't tolerate it, you have to do something. And that means the idiots have to do something, too.

For example, Kathryn has been doing some interesting sleuth-work on last month's mysterious African rent-a-coup, and so her weblog has been immersed in spam, bickering, and legal threats. (You know it's getting complicated with you see Comments (158) | TrackBack (0) )

They weren't all fools
Among the worst, and most popular, historical myths is the belief that people used to be idiots. It's always tempting for historians to fall into this because the historian knows how things turned out, and because our inner adolescents perpetually remind us of what fools are parents were.

One symptom of this mistake is sudden descent from something very interesting into unreadable politics:

Orientalism serves as a safety valve for gender friction by providing imagery for male wish fulfillment and idealization. This function was signaled early on by Hiram Power's The Greek Slave, which depicts a nude woman chained and sold into harem slavery by the Turks during the Greek War of Independence. This unprecedented sculpture came to epitomize the nineteenth century cult of pure womanhood, with its ideals of chastity, piety, and quietude.

This is Holly Edwards channeling Edward Said, and though her catalog of Noble Dreams, Wicked Pleasures: Orientalism in America 1870-1930 is full of delightful work, the text is often marred by this sort of carping. The artist's politics were not ours -- or at least not Said's -- and so we have to hold the artist (and the buyers, viewers, and critics who were all fellow travelers) in a certain contempt. (Edwards does do an important service in explaining why France's imperial issues were not America's, and that Orientalism in American painting does not have to mean the same thing that it meant in Paris)

But if of politics we now would speak, what were the artist's and the viewer's politics, anyway? No, the Greek Slave isn't the girl next door. If she were, this would be unbearable. Terrible things happen; art gives you some distance. Sooner or later, it is going to happen here , and it's easier to plan and to prepare if you begin by thinking about what's happening over there.

And somehow, speaking of politics, we've managed to forget that we're in New England in 1843 and we're talking about people who knew slaves, people who could (and did) say that some of their best friends had been slaves, people who were 17 years away from busting up the last, best hope of mankind and walking out of their New England parlors all the way to Georgia, in order to destroy slavery at any cost.

If you convince yourself that those old folks were fools and louts who just wanted to do the male gaze thing to the Orient, you're fooling yourself. They may have been wrong, there were lots of things they didn't know, but they weren't merely greedy simpletons and perverts.

They weren't all fools
In Carpeaux's Les Quarte Parties Du Monde (Luxembourg Gardens, maquette in the Orsay), we have four women -- Europe, Africa, Asia, and America, supporting the globe. Africa's leg is shackled, and America just happens to be stepping on her chain.

I spent a couple of hours this afternoon, sketching The Greek Slave in Poser, to sort out two different ways we're distanced from this girl. One is intentional -- she's a hunk of silicon dioxide. The other, though, is partly an accident of time. Her hair is antique, and that's a powerful source of distance. (Anne Hollander points out that, in historical movies, the stars almost always have modern hair.) And her body is antique, too; nude statues, from the Greeks on, are almost always distorted by the forces of the underwear that the subject is not wearing . (Poser has the same problem -- and it gives her the abs of a kouros -- but because it's modern we don't see it so prominently)

The unexpectedly interesting parts of this exercise in new media remediation were the hands. I wanted to change the right arm just a little, moving the girl a little away from the column and letting her put more weight on it. Hiram Powers probably couldn't do this because the marble wouldn't withstand the shear force, but we can benefit from the magic of virtuality. The Poser 5 slave isn't conscious of the corset she's not wearing, she's accustomed to jeans or skirts or to the little black dresses that Coco Chanel will invent for her granddaughter, and I wanted some muscle tension somewhere.

The left hand, I thought, was a bit of silly 19th-century nonsense. It turned out to be a headache, because small changes in gesture end up reading very differently, and I didn't want to wind up in Fleshbot. It was murder to get anything close to acceptable, and to do that I pretty much abandoned the attempt to mimic the original angles. (There's a reason you never see hands in user interface icons: nearly every gesture is obscene, somewhere)

It's a hell of a note that we're now more puritanical than the 19th century Boston Brahmins. But, let's face it, we are. Gender friction? Male wish fulfillment? The viewers were often women, the tastemakers were even more often women. Those women didn't have much political power, but they had some -- and they could easily have preferred something else. And the whole point of the sculpture is that you're obviously supposed to identify with the slave girl.

And this is becoming the sort of thing of which plain folks, in 21st century America, dare not speak.

Nov 03 5 2003


I've revamped this page to use CSS styles instead of nested tables. You should think about doing this, too, next time you dust and clean your design; it's cleaner and faster and better for the Web environment.

This might be as good a time as any to mention a delightful experimental site, the CSS Zen Garden, which demonstrates the enormous visual range of contemporary, standards-compliant Web design.

Over at Grand Text Auto, Andrew Stern posts that my question about games is obvious. In the comments, Wardip-Fruin says that it's just not an interesting question. (Jason Rhody, writing on his own weblog, found Stern's restatement more interesting, and his response, from Asheron's Call, is very useful indeed.)

It's interesting that Andrew's note has, in the course of a weekend, generated one of the longest comment threads in the site's history.

Another interesting observation is the inherent tendency of comment threads to become usenet flamefests (an observation Winer made at Bloggercon, though this thread isn't yet very toasty).

"You're all a bunch of fancypants. My response to the question: Syndicate" -- arch stanton

That's interesting, too. I know the work of many of the other contributors, but I don't recognize Stanton's name offhand. I don't know whether the post is written by the musician, or by an admirer, or if if the eponymous domain represents this musician or another one. Perhaps its another arch stanton entirely. Isn't Google interesting?

But, if I understand the intent, this is real criticism of real games -- or at least a gesture in that direction. Like real criticism of real hypertexts, by other means.

And that's always interesting.

Almost two years have passed since I issued "Bernstein's Challenge" to the game theory community.

Let me try a probe, just to make a little tsimmes. Take the last twenty years of computer games -- the whole kit and kaboodle. Put them on a shelf. (Yeah, it's a big shelf) Now look over the shelf, and tell me everything we learn about, say, sexuality. (House rules: no arguments from silence, no metaphors, cigars are just cigars unless it's apparent to players who aren't Professors that they aren't. Chatroom romances don't count; I'm looking for what's in the game, not what the audience brings to the table, and yes, I see the theoretical shortcomings of the previous clause. You understand what I'm getting at. Play along at home; it's that kind of movie )

This inspired a lot of email discussion, which I enjoyed. Obvious extensions to the question were raised. Fathers and sons. For that matter, how about not having enough money? Or the setting of the best meal I've had this decade -- pace last week in Paris?

Bernstein's Challenge

Lots of questions. Few if any answers. And then everyone seems to have gone home and hoped the problem would go away.

Two years later, next month, and there's still not much of an answer.

Surely, if we can't answer the question, it's interesting to ask, "Why not?"

And, surely, with a question this big lying on the table, you'd think people in the field would engage it?

Scholarship proceeds through dialogue, by finding questions and then finding answers. We're waiting, folks.

As I'm walking home, a fellow pulls over and asked me for the best way to drive to Paddington. I'm from Boston, I have no idea where Paddington is, and I'm striding confidently down Gower Street because I narrowly escaped from being enmazed in Russell Square. I'm not that confident anyway: I've been up for 40 hours, I just walked about 15 miles or so through Bloomsbury and Notting Hill (see below), before that I flew across the ocean, and I have just enjoyed two of those excellent (but preternaturally large) pints of real ale.

Someone always asks me for directions. Lately, someone always asks on my first day. Often, in a language I don't speak, in a place I've never been.

Why, exactly, did we think hypertext navigation was a problem?

I had those two pints (memo: special is worth the 15p premium) at a very pleasant, unremarkable pub just down the street from where Foundling's Hospital used to be. It's a park now, for children -- Foundling's is. Not the pub. The pub has always been there, and it's always been a pub.

I did have some trouble, ordering a pint. even though I read an entire thesis on how to act in a pub. Spending more time in Britain and among Britons abroad has left me nearly deaf to accents. I forget that, where I hear a charming trace of an accent I can't place, the other person hears me speaking broad and incomprehensible Chicago with overlays of something nobody can identify anymore as a speech problem (thank you Mrs. Horowitz). Like the Minbari, Britons may sometimes sound like us....

The pub seems unremarkable, but Robert Elms did remark (in London Walks v2) that it's not much changed since his schooldays -- or since the 18th century -- "complete with snugs and engraved screens." So, back in 1775 or so, a couple of medical workers probably sat in the nook where I was reading Hiromi Goto and complained about their caseload and worried together about the troubles over in Boston.

Everything in this post is about hypertext. It has no links.

In the past months, many weblogs have pondered Iraqi blogger Salam Pax. Is he real? A fiction? Paul Boutin summed up the evidence  last month:

Speculation continues that Dear Raed , the weblog of a young man in Baghdad who posts under the name Salam Pax, is a hoax, perhaps even a disinformation campaign by the CIA or Mossad. A month after Computerworld published a story quoting a "terrorist" who turned out to be one of their former writers pranking them, it would be foolish not to wonder.

An interview with Salam Pax by Mark Stephen Meadows, to appear in Tekka this week, appears to close the book on the question. Salam Pax is real.

  • Meadows met Salam in Baghdad, at the Baghdad Sheraton. They spoke at length, on the record. Meadows also met Salam's father. Salam asked not to be photographed.
  • Salam blogged the meeting in some detail, on May 22.

So, we know that Meadows met someone in Baghdad who said he was Salam, and who has access to Salam's weblog. This doesn't completely preclude some off-the-wall scenarios -- perhaps Salam flew to Baghdad for this interview, having posted his weblog from Wisconsin through an Iraqi intermediary? But what would be the point?

As I've written previously, we worry about the authenticity of weblogs more than we ought. Kaycee Nicole was true, although she was a fiction. Isabella (new URL) revealed an important lesson about thrillers that the interactive fiction community had overlooked. Salam reminds us that people can be who they say they are -- that, sometimes, things are pretty much what they seem.

Of course, we still have all the questions about reality that face any reader. When Salam offers an opinion, or I do, is this precisely what he believes? Of course not. We write things we believe, we write things we want to believe, we write things we ought to believe even though we cannot believe them. We write about who we were, not who we are; we write about who we wish he had been and what we wish we'd thought to say. Isn't that more true, in the end, than a recording?

Dave Winer and Tim Bray each write today about the terrible damage we are inflicting on the software world. Winer writes:

Today software and music, software and writing, software and all kinds of creativity, are indistinguishable.

and asks, "Who will pay for software?" Bray observes that corporations today are unreasonably risk averse, that writing a small software order requires Permission From The Top, and that this stifles everyone. He's trying a shareware strategy for big-ticket systems.

My intuition is that trying to make brilliant, innovative software is hard, and trying to establish brilliant, innovative business models is hard, too. Trying to do both at once might be too hard for anyone. But, you never know, and it's worth a try. Nothing else is working…

I touched on this recently when I observed that the call for freeing artists from the need to program is really an effort to see scientists and engineers as subhuman.

The one mention of Buffy in the 2003 DAC Proceedings is in Tiffany Holmes' interesting survey, "Arcade Classics Spawn Art? Current Trends in the Art Game Genre." Unfortunately, it's wrong.

The online explosion of the riotous cyberpunk culture in the middle to late nineties was followed by a resurgence of a glamorous fighter chick in both television and Hollywood productions. Hollywood’s Lara Croft, Buffy, Zeena, the Matrix’s Trinity, and Charlie’s Angels are but a few examples of the new technologically adept warrior princesses.

The problem is that Buffy is not technically adept. (Is Zeena? Seems unlikely) And Buffy is hardly an outgrowth of cyberpunk culture -- any more than Charlie's Angels were. And cyberpunk culture wasn't online, and as Neuromancer appeared in 1984, "middle to late nineties" might be a bit of a stretch as well.

It's an interesting paper, and these are minor mistakes that don't vitiate its argument. I'm picking on Holmes because it's a useful paper.

But there are errors, they contaminate the literature, they could lead people astray. (And how I wish the next sentence didn't concern Laura Croft's breasts: if you're going to do Freudian exegesis of her pistol, could you perhaps show some awareness of the role of latency, or whatever you want to call it, in Tomb Raider?) Do the humanities people just not care? It's not like this obscure knowledge is locked away in an unknown Anatolian archive. That's why we review completed papers: to find minor holes and fix minor gaffes.

Buffy is, obviously, a response to the teen slasher movie, and specifically to the genre's second body -- the pretty girl who enjoys her sexuality without earning it. Whedon says so, but it's not exactly rocket science to figure it out for yourself. Buffy is not Trinity and she's surely no Angel, and if you can't tell the difference aren't you arguing that all young women are interchangeable?

Note for Watchers: The Blonde Victim is attacked by monsters because everything comes easy for her. She's pretty, she's talented, she's popular, she enjoys her body. She's blessed -- and she does nothing to earn it. And that's exactly where Buffy starts, back in high school.

Now, very close to the end, Anya (of all people) says the words: "You're not better than us. You are just luckier. Than us."

So, we're doing Isolation Of The Hero, and we're making a clever little point about relationships by having The Night When Everyone Has Great Sex Except Buffy, we're courageously making TV safe for gay people, and all the time we're letting this childhood terror out one last time.

Not bad for about 20 minutes of footage. Or, maybe I'm wrong. I think I know where we're going, but you never know.

May 03 7 2003

Flight Risk

Wired reports on a new, unfolding Web narrative: Flight Risk. Leander Kahey calls it "the incredible story of Isabella V., a wealthy young woman who claims she went into hiding in early March to avoid an arranged marriage."

"On March 2, 2003 at 4:12 pm, I disappeared. My name is isabella v., but it's not. I'm twentysomething and I am an international fugitive"

Kahey gets all tangled up in asking whether or not Flight Risk is a "hoax". He's really asking, "Is it true?", which is something different. (With Kaycee Nicole, there was concern that the author might be seeking donations to help with medical expenses that, like Kaycee, weren't real. But Isabella is an heiress; she doesn't need your money)

What we have here, presumably, is a new media thriller unfolding in real time. That's a very clever idea. Thrillers are very closely allied with mysteries -- both kinds of stories describe how a damaged world is once more made whole -- and mysteries have long been a very difficult target for hypertext. Letting the thriller play out over time can be enormously effective (witness The Fugitive), Thrillers begin by forcing the hero to step off the sidewalk and into another world; the problem here is to choose the rabbit-hole:

  • If the danger the hero faces is too pressing, Web detectives will get to the bottom of the story and end it prematurely. Look at the investigative resources harnessed for Kaycee, or for Imagine what it would be like if our heroine faced credible danger -- if we were convinced, say, that the blogger was in imminent peril of being abducted by a drug cartel or shot by gangsters from New Jersey or Novogorod. ISPs and hackers would fall over themselves trying to help, and agencies from Interpol down to your local precinct house would dream of breaking the case. It would be madness.
  • If the danger isn't real and important, we won't particularly care -- and the weblog will be lost in the noise.
  • We don't want anyone to get hurt or to break the law, because we might be implicated as an accessory. This means we can't really attempt a heist thriller.

Isabella's story is perfect. She's a Girl In Peril. She's rich, which is always interesting, and gives the writes flexibility; rich folks can always catch a plane if the plot gets jammed. And her peril is precisely tuned so that you want her to escape, but you don't really need (or want) to rush in and save the day. Obviously, even if you figured out where to find her, you'd worry that her Evil Mastermind Father would have Minions on your tail, and you'd lead them to her.

P.S. Someone really needs to look up the creators of The Spot (anyone know where it's archived?) and give them an award (or something) for creating an entirely new literary genre. And someone needs to give all this a name. Suggestions? (Check here -- thanks Eric Scheid!)

Update: Isabella has sent me email; she says she is real.

Paul Graham's Hackers and Painters argues, correctly, that making fine software resembles painting more closely than engineering.

What hackers and painters have in common is that they're both makers. Along with composers, architects, and writers, what hackers and painters are trying to do is make good things. They're not doing research per se, though if in the course of trying to make good things they discover some new technique, so much the better.

I think Graham is wrong about scientists -- science is a lot like painting, too. He's got computer science dead to rights, though, when he calls is "a grab bag of tenuously related areas thrown together by an accident of history, like Yugoslavia." Like the other fields that call themselves a science -- Political Science, Social Science -- Computer Science isn't a science, and tends to waste energy hiding that embarrassing detail. And I think that he's exactly right when he says that

if you can figure out a way to get in a design war with a company big enough that its software is designed by product managers, they'll never be able to keep up with you

I certainly hope that's correct. But I think Graham is simple wrong about science.

Scientists don't learn science by doing it, but by doing labs and problem sets. Scientists start out doing work that's perfect, in the sense that they're just trying to reproduce work someone else has already done for them. Eventually, they get to the point where they can do original work. Whereas hackers, from the start, are doing original work; it's just very bad,

Labs and problem sets are just learning the mechanics -- how to turn on the machine, where the keys are, how to use the editor. They take a little more time in the sciences because some of the machinery is a little more dangerous than an iBook. You can't let people figure out Dewar flasks and vacuum lines in their dorm room, or you'll get lots of broken glassware underfoot. But it's the same stuff. (A lot of lab time is spent studying the equivalent of COBOL and JCL; there are reasons for it, but physicists and chemists dislike the situation just as much as you'd expect)

Scientists learn the hard part of doing science -- finding problems worth studying and learning how to find answers to questions nobody though to ask before -- in precisely the way programmers learn: by working right next to a real scientist for a few years, by studying good work, and by doing science.

May 03 3 2003

Narrative effect

Many people like to play down the role of craft in weblogs, and it's almost heresy to talk about craft in the digital storytelling world, where being "natural" is prized above all.

But craft matters. Look, for example, at Alwin Hawkins account of a rough weekend. The gist of the entry is: "I haven't been posting much, because work has been busy." You see this very note in weblogs all the time. Here's one from Kottke.

Good morning! It was a helluva weekend, as you could probably figure out from the paucity of postings here.

Last night was one of those Sunday Night Specials that cap off the WeekendFromHeck frequently. Hammered by admissions of incredibly sick heart patients, we gave as good as we got.

My own trial consisted of a woman who blew out a papillary muscle , causing her valve to flail wildly and putting her into cardiogenic shock . We took her to the cath lab and stuck a balloon pump into her to get her blood moving round and round again. While in there, we found that most of her vessels were nearly completely closed, forcing us to 'hold the fort' until we could roust the open heart team out of bed to do a combo bypass and valve job .

In the meantime, the patients continued to come, and we continued to fix 'em and put 'em to bed. Lovely time, but I was glad when I saw the sun start shining in the windows.

Yeah, we all survived, both nurses and patients. And that's all it takes to make it a good morning some days.

Now, this isn't fancy writing, it's not meant to be poetry, and if there's any affectation here, Hawkins is writing down, not reaching. But look again at how this is put together. We begin in the morning (title: 7:35, post timestamp 9 am), and we begin by revealing the end. Then back to the start of the weekend. Leap forward to last night. Then the main story, followed by a flashback ("the patients continued to come, and we continued to fix 'em"), and then the wrap-up.

You can write this well (and this with this kind of complexity) without thinking about it, without theory or study. But thinking is always good.

Gordon Meyer describes a unique experience: learning to use Tinderbox to take notes at a lecture by hypertext pioneer Doug Engelbart. (Engelbart's NLS/Augment was probably the first full-scale hypertext system; the same project also invented the computer mouse, the outliner, and groupware).

Sometimes, an application just comes together; it sounds like Tinderbox is at that stage. Meyer recalls that:

"Engelbart discussed how technology should be used to augment our abilities, to extend what we can do. Automating the mundane does not move us forward -- we should be building tools that augment our human capabilities -- making connections, thinking, communicating, and extending our knowledge.

"I laughed aloud as I realized Tinderbox had just done exactly that for me. Its "agent" feature was allowing me to gather a better understanding of this view of "smart agent" technology. How wonderfully recursive!

Powazek advocates raising the bar in an electronic community-- making it difficult to participate in order to deter irrelevance and casually malice. The Zoetrope writers community has a significant bar!

First, it's hard to find; this might be bad design but it's probably deliberate, recalling one version of that had a secret URL.

Second, you have to register.

Third, after you register, you learn you're only a probationer; until you write five reviews of recently-submitted short stories, you have very limited access.

This encourages active participants, makes newcomers share the work, and chases away freeloaders. In a field rife with wannabees who want the perks but can't or won't do the work, it's a very clever idea.

Mar 02 21 2002


The Treo is a new Palm-compatible cell phone from Handspring. Dan Bricklin liked the Treo a lot:

Many aspects of this device that are important to understand if you want to see where handheld computing is going.

I was so convinced by Bricklin's review that I bought a Treo. I've never used a cell phone before. It's very exciting. (The Treo is loud enough for comfortable use, even though my hearing is terrible)

Once again, the Palm proved a superb information appliance. My Treo works as a phone, out of the box. It works as a Palm, too. Better yet, a few minutes of beaming from my old Palm IIIxe to the Treo, and suddenly the Treo is my Palm: it knows my contact list, it knows the list of things I'm planning to write, it knows the itinerary for my upcoming trips.

Mar 02 19 2002

7 Issues

In 1991, Frank Halasz's 7 issues revisited updated his important 1987 lecture on the key hypertext research topics. A decade later, there are still well worth pondering.

Anders Fagerjord mentions Where are the hypertexts? and observes (March 7) that "Notably, Mark Bernstein himself doesn't write hypertext. He writes papers, slide shows and a blog. Even as he said himself, the trouble with blogs is that dates rarely are a good way of organising material."

This is a fair indictment, I think, but the matter is complicated. My Election of 1912, published in 1988, is now of purely historical interest but it's definitely a hypertext. Hypertext Gardens is a hypertext (and gets assigned in lots of Web writing courses). This web journal is itself pushing the envelope of weblog hypertextuality -- there's a lot of linkage and a lot of collage here, and there will soon be more -- while staying just inside the bounds of weblog convention. It's important that this page feel like a weblog -- perhaps a richer, linkier weblog, but a weblog nonetheless; the purpose here, after all, is to improve tools and techniques.

We're planning a Hypertext '02 panel on web logs, Wikis, and self-assembling hypertexts to explore issues raised by Web writing clusters, such as the Scripting News technojournalists and the Scandinavian media theory cluster. (Want to play? Write me) There's more going on here than links between documents.

Fagerjord quotes Aarseth on hypertext: "It's an ideology", he says. But partial differential equations work. So do French horns, and omelet pans, and game theory, and computers. Lots of people use none of these very effectively; some things take time, effort and practice.

Derek Powazek's Design For Community makes a fascinating observation in passing. Community sites and discussion forums aim to attract posts, to build big discussion threads on hot topics. You might expect that the best way to do this would be to make it easy to post; Powazek finds that it's actually a better idea to make posting difficult, to "bury the Post button" deep in the discussion. For example, in {fray}, you can post only at the end of multi-page stories.

The core problems community sites face are flamewars and graffiti -- casual posts that add nothing but noise, that sap the vitality and spirit of the discussion. Making it hard to post reduces graffiti; only committed people stick around long enough to post. By the time people reach the Post button, they've had a chance to absorb the spirit and the mood of the site.

Why am I only now writing about Powazek's fine Design for Community? It came out last Fall! But I cleverly sent it ahead to Singapore, so I'd read and review it in an ideal environment: an office, a good library, a good Web connection, and no temptation to rush out and write code to experiment with Powazek's ideas.

Unfortunately, the package never made it to Singapore. Or, rather, I never got the message that said "Come and get it!" In due course, the Singapore post shipped it back to the Boston office, using a slow but expensive surface rate.

Feb 02 27 2002

Hypertext 2002

This weekend, I'll be in Maryland for the Hypertext 2002 Program Committee meetings. I've been on every Hypertext Program Committee except for the first one (1987), and I've been co-chair twice. It's always an interesting conference, usually the most interesting conference I attend each year.

Hypertext '99 PC Meeting, Darmstadt. Hugh Davis, Uffe Wiil, and Joerg Haake

The program committee's work begins with the paper submissions. Most papers are already carefully vetted by the author's frienda and colleagues even before the paper is submitted, and every submission is read by at least three (and often by five or six) expert reviewers. The readers are looking, first of all, for original ideas and for outright mistakes -- errors in fact or methodology that could mislead future researchers. Finding mistakes in cutting-edge research can be incredibly difficult; that's why so many reviewers are needed.

Each year, a handful of papers breeze to general acclaim, and a somehwat larger number are quickly dismissed as inappropriate for the conference, fundamentally unsound, or otherwise unsuitable. The rest are examined by the committee, paper by paper, hour after hour. The views of the readers are scrutinized, and often still more readers are recruited to resolve thorny questions. Calls may be placed to authorities who live half way 'round the world, just to make sure that a paper's claims are sound. (One year, wanting to make an argument within the research community, I planted some in-jokes and barbs in the footnotes. I knew that they might not be accepted, but calculated that everyone I wanted to see them would wind up reading them during the reviews)

What surprises most people is how easy it is to have a paper rejected. My first year on the committee, I was astonished by the carnage: major papers from big labs, leading universities, acclaimed projects were studied, scrutinized, and judged not quite good enough. Nobody cares that the paper comes from a big company. Nobody cares that is comes from an ancient university. It's got to be right.

Ten tips (in progress)

  1. Write for a reason
  2. Write often
  3. Write tight
  4. Make good friends
  5. Choose good enemies
  6. Let the story unfold
  7. Stand up, speak out
  8. Be sexy
  9. Use your archives
  10. Relax
Feb 02 22 2002

Weblog clusters

Tinka has a nice summary of the growth of weblog clusters (see also Torill), and a fascinating note about the ambivalence people feel about the size of their audience. If nobody is listening, why write? But, if too many strangers are listening, people feel inhibited -- a shared, private writing space can suddenly seem all too public.

Though people write about the people who read their web logs, I've been thinking lots about the people who don't read this page. When lots of readers started appearing every day, I assumed most of them were friends, family, colleagues.

That turns out not to be the case. My mother doesn't read this. (I wish she did) Neither does Linda. I reconnected with some old school friends a few years ago; we send regular newsy emails. Only one of them reads this (and he's the worst correspondent). My little sister doesn't read this.

Feb 02 19 2002

Now, it's yours

This morning, I woke a little before sunrise and didn't work on Tinderbox. Tinderbox 1.0 shipped yesterday. It's finished.

Until yesterday, there was nothing much like it, anywhere. Yes, it's a little like Storyspace, VIKI, Blogger, Agenda, PageJockey, and a host of other systems old and new. But it's a very new kind of software, and until yesterday, if this is the kind of software you needed, you were in a bad place. Now, it's done.

Tinderbox is a personal content management assistant. It's a wonderful tool for making, understanding, and sharing notes.

I've been working on, or thinking about, Tinderbox just about every day since Sunday, August 15, 1999. That morning, as Linda's Toyota headed out the driveway for a photography workshop in Vermont, I settled in to write a new foundation for a pair of hypertext systems. You know these today as Storyspace 2 and Tinderbox. Through the week that followed, Linda photographed and I coded non-stop, working almost around the clock. It was the most extreme of extreme methodologies: an idea, a plan, and a deadline. It was the best run of coding I've ever had.

Tinderbox itself started on April 20, 2001. Since then, it's been 304 days. Roughly 581 files of code, resource definitions, working doc, and test jigs (not counting libraries). About 80,000 line of code. 28 developer releases. The first Tinderbox weblog, the Development peekhole, was published on June 1.

In Hawaii for a conference, I took advantage of the jet lag to watch the sun rise while writing the code that adjusts links when you add or delete text. "It's much harder to concentrate in on a tropical beach than you'd think," my notes read, "although the code seems to be good." Those notes are in a text file; for months, all my programming notes have been in Tinderbox. Now, Tinderbox is yours.

In Denmark for the Hypertext meetings, I took notes in one window and fixed bugs in another. Some of those notes led to features like the Nakakoji view that made it into the first release. On the trip home, the guy in the seat in front of me had his seat all the way back for ten hours. Mine wouldn't budge. I scrunched my half-open laptop sideways in my lap, and fixed a show-stopper from Australia.

In Berlin for p0es1s, I stole moments late at night to fix text editing and adjust HTML templates. On early Malden mornings, I spent birdwatching time hunting bugs and tuning code. I've lost a lot of sleep.

Every line of code had to be written, revised, checked, and tested. Everything has to be just right. Lots of things ended up on the floor -- entire families of classes were written, debugged, rewritten, and then refactored out of existence. Software is craftwork; every joint and bevel in Tinderbox was carefully wrought and polished by hand.

I've wanted this tool for years. It's not everything I want. Yet. It won't be everything you want. But it's new, it's here, and now it's yours.

A holding company sent out lots of warning letters this week, claiming that they own a patent covering all modern metadata formats. This includes the simple RSS syndication format we (and everyone else) use to share headlines between web sites.

I've only glanced at the claims, and I'm not much of an expert on either law or metadata, but the patent is written in esoterically amateur language that seems tantamount to gibberish. I doubt if anyone knows what this patent is trying to say. Of course, if a judge somewhere happens to think it means something, the patent can be enforced for a few more years. But this seems unlikely to happen; there's a lot of prior art lying around in the early AI literature, and even more (I'm told) in the database world.

Meanwhile, the holding company can use the threat of lawsuits to make people pay fees to avoid lawsuits. The lawsuits won't succeed, but would be expensive and bothersome to defend; it's tempting to just pay the money. This is an abusive of the public trust, verging on a protection racket.

Jan 02 4 2002


Diane Greco, who rarely writes ill of anyone, denounces "the latest thuggery about cybertext at EBR."

Then, Anja Rau weighs in on the responsibilities of the hypertext critic, doing a nice job of showing how van Looy's 23 charges against We Descend amount to little more than a general dislike of the author's prose style and language.

It's worth noting, too, that the reviewer gets upset about font size, something he could easily have corrected himself. It's a little like complaining that an 18th-century book is bad because the typeface is old-fashioned and you were sitting in an uncomfortable chair when you read it.

Tis hard to say, if greater want of skill
Appear in writing or in judging ill,
But, of the two, less dang'rous is th' offence
To tire our patience, than mislead our sense.

Jan 02 3 2002

Bad reviews

Jill Walker savages van Looy's overwrought critique of We Descend.

A review should take a work seriously and consider it on its own terms. That doesn't mean not being critical. I'm happy for a reviewer to decide that the work's premises are flawed, or that it doesn't fulfil what it's trying to do.

That's the key point: reviewers need to do the work. Sitting back and inventing fun wisecracks is easy work, and it produces something that looks like a review. But something that looks like criticism, but isn't, is actively harmful: it crowds out good criticism, debases discussion, and replaces serious scholarship with cheap thrills.

It's easy, when a critic gets worked up, to go overboard and lose perspective. Writers reaching for an unfamiliar voice (and I suppose van Looy is trying to do Dorothy Parker, not his usual mode, in a second language -- quite a challenge!) can stumble. I've done it lots. That's why editors matter, and shouldn't be scared to do their work.

Jan 02 2 2002


In "What Cybertext Theory Can't Do", Kate Hayles presents a graceful reply to Markku Eskilinen's "Cybertext Theory and Literary Studies: A User Manual."

As a follower of Aarseth, Eskelinen is more dogmatic and ideological than the theorist he has adopted (a well-known phenomenon observed with Marxists who are more dogmatic than Marx, Freudians who are more doctrinaire than Freud, Derridians who are more inflexible and less subtle than Derrida, etc.). Ironically Eskelinin is quick to claim other disciplines want to "colonize" cybertext, while engaging in a rhetoric that in its ideological excesses is as imperialistic as anything I have read in recent years.

It seems to me that the only things worth reading in EBR these days are rebuttals and corrections. Is this a sensible way to work?

Dec 01 28 2001

Jane's spaces

In hypertext parlance, a Jane's Space is a part of a hypertext that you can't find in the usual, link-following way. A Web page that's not linked to your site and that's hidden from the search engines is a Jane's space; you can only get there if you happen to know the URL. Jane's spaces are named for Jane Yellowlees Douglas (it's a long story)

I recently wrote a small program that scans Storyspace documents, looking for spaces with text but no inbound links. Of 28 published hypertexts, at least 16 appear to have Jane's spaces. I knew some of these, but the overall total seems extroadinarily high.

Dec 01 21 2001

I've been reading a variety of Singapore web logs lately. My favorite right now is by Pearl Pan, a student here at the University. It's curiously titled "Smelling Good Today". Pan write with energy, and she paints with even more energy -- and nearly every web log entry is capped with a cartoon of her eponymous character. Just when the web journal settles down into the mundane, it takes a sharp detour:

"hey! did you know that we went to chua chu kang grave yard to plant two trees yesterday morning just before the rain started pouring there!! well, it's nick's late grandpa. he was killed by some hostesses in his 30s... sad drama right? "

Dec 01 20 2001


Lou Rosenfeld has lots of intriguing ideas about Information Architecture Components -- elements "that get users to content". His newly revised taxonomy includes Browsing Aids, Search Components, Content and Tasks, and "Invisible" (infrastructural) components.

The problem here, as in the Rosenfeld's earlier book, is that the architectural elements are chiefly limited to signage -- labelling systems that identify where you are and tools for instrumental navigation. Architecture isn't about applying labels to spaces, it's about building spaces that label themselves. Louis Sullivan (for follows function) is the classic source, and he famously insisted that the windows of large buildings should indentify what happens inside them, rather than following some arbitrary Palladian scheme.

Trying to separate content from navigation is usually destructive of both. The most important content a hypertext offers is often its self-image, its mental map, the promise it offers the reader. What can I learn here? What can I buy here? Am I welcome? (Who are you? What do you want?)

Even more interesting in Rosenfeld's recent writing is a speculative combination of links and search. Morville's ideas are close to the essence of what Microcosm called Generic Links, although the actual flavor feels like HyperTIES (and, de facto, a little like Ward Cunningham's WIKI). I suspect the oldest implementation was Rosemary Simpson's Gateway, for the LMI Lambda.

Nov 01 15 2001


Danish media scholar Lisbeth Klastrup noticed some .mil and .gov visitors in her referrer logs after 9/11, and wondered whether the US government was watching her.

Checking a month of Eastgate's logs, I find that roughly 0.1% of our traffic comes from .gov and .mil respectively. That's about the same amount of traffic we see from Ireland and New Zealand.

The US government employs about 1.8M people. The work force of New Zealand is 1.8M people. The Irish workforce is nearly the same size. So it's probably just the effect of lots of people with diverse interests, and the effect of building a successful Web journal.

Nov 01 13 2001

Rau Row

For many years, I've been calling for real criticism of real hypertexts. Caitlin Fisher's "These Waves of Girls" is a real hypertext, of which Anja Rau is really critical.

Rau attacks on three separate fronts. First, she finds the work lacks craftsmanship. Load times are unnecessarily long. Significant links are broken. Borders come and go, sounds seem to be applied arbitrarily, and Rau attributes these inconsistencies to mere inattention. She notes that better solutions are available for the challenges that seem to have baffled Fisher; Flash sounds, vector graphics, better window management, and PHP could all (she thinks) have been deployed to good effect.

Next, Rau finds that the various elements -- text, images, sounds, animations -- that comprise the work fail to cohere; they are thrown together without forming a coherent whole. Strunk and White suggest that writers omit unnecessary words; Rau applies the same rule to tiled images, rollovers, noises. (Diane Greco gave a fine talk on this common failing to the Canadian Library Association a few years ago)

Finally, Rau believes the work was misconceived because it "utterly ignores its predecessors in both print and the digital medium."

The compulsion to adopt a unified identity and a linear story of ones life that came and still comes with coming out, might be represented in a text with a hypertextual layout. Waves is obviously not aware of such possibilities -- let alone of the tradition of homoerotic and especially lesbian (sub)texts in hyperfiction from Victory Garden via Patchwork Girl to Desert Mauve.

The hazard of this argument is that the critic can be tempted to argue about the work that might have been, not the work at hand. Occasionally, ignoring the past is a good idea. What particularly annoys Rau, I think, is not Fisher's turning her back on the past and its possibilities, but rather that ELO judge Larry McCaffery (who described himself as "an interested novice") failed to observe or address the issue.

Rau concludes by suggesting that these shortcomings indicate that multimedia is too hard for any one person to do alone, that collaborative teams are now indispensable. I think she's mistaken in this. Certainly the shortcomings of one work don't prove that better work is impossible. "Modern, commercial Web sites," Rau observes, are "built by teams of .... specialists, often in several months of hard work." This is, of course, indisputably true, and collaborative production of new media might yet prove important even if the first steps seem unpromising. Still, a host of influential and technologically complex sites are crafted by individuals, often working in their spare time (see Independents Day for polemics about this, or look at Rayseo, 6amHoover, or Praystation for examples)

The fewer hands, the fewer the barriers to real feeling, intensity, and expression. We read to make contact, to touch another mind. Two is company; a collaboration is sometimes a crowd.

Reading this essay, I find that at times I agree with Rau. At other times I disagree, and on occasion I'm not sure. But we're always talking about ideas, not arguing about taste gossiping about box office. This is what criticism is supposed to do

Jill Walker responds..

Nov 01 9 2001

Economy of Links

Realism and a General Economy of The Link, by Adrian Miles. Much to ponder!

Nov 01 6 2001


The Proceedings of Hypertext 2001 are still not available from the ACM, apparently because of a technical copyright issue involving short papers which lacked an individual copyright notice.

The deadline for Hypertext 2002 is approaching swiftly.

The ACM exists to advance knowledge, not to hoard copyrights. If Hypertext 2002 papers cannot build on the results reported at Hypertext 2001, the ACM will have broken faith with its members, betrayed those who participated in the conference, and damaged the growth of the field and the ideals to which it once was dedicated.

The most important step in writing hypertext is writing the hypertext. Find the stories, do the work, trust the reader.

(Conclusion of my Southampton University talk, 6 Nov. 2001)

Nov 01 3 2001

Next Move?

Torill provides the first really good answer the the game challenge, proposing Adeline Software's Twinsen, "in which we are fighting the evil tyrant Dr Funfrock... In the end of LBA I the mother-figure rises and envelops Twinsen in her love when he has conquered the evil tyrannic father-figure..." This isn't quite what we were talking about -- it sounds closer to power fantasy than to discovery. But no quibbles, and drawing attention to specific good work is good.

Gonzalo Frasca sends word of Rune Klevjer's promising work on Computer Game Aesthetics (RTF). "A computer game is only a game. If nobody cares about the dubious ideology of chess or Monopoly, why should we care about the ideology of a Rambo-like and gory-realistic shooter game as Soldier of Fortune (2000)?" Frasca's own Ephemeral Games: is it barbaric to design videogames after Auschwitz? (PDF) is an interesting response.

(If you're joining us in progress, see A Game Question first)

Act II: You (yes, we mean 'you' -- don't miss Walker's wonderful piece on the 2nd person in games) are a comfortable professional, working perhaps in a University in a pleasant university town. One day, unexpectedly, Something Happens, and you discover that your father (or your daughter, or your lover) is not the man you thought he was. You find he's capable of terrible things. You find he's horribly, irredeemiably weak.

It's a dramatic moment. A narrative moment. A very human moment.

Can we do it on stage? Sure: Aeschylus. Plautus. Shakespeare. Wilde. Tony Kushner. Can we do it on screen? Ask Cary Grant, or Groucho Marx, or Buffy creator Joss Whedon. Poems? Epic or light, from Mortu Arthure to W. S. Gilbert. Can we do it in hypertext fiction? In spades, hearts, or diamonds (WOE, Charmin Cleary, Victory Garden. Plenty of others)

So let's step over to that big bookshelf that holds our 20 years of games, and start picking out the games that tell us about Fathers and Sons. (Same house rules) What've I missed? (Email me even though I'm on the road this week.)

Nov 01 1 2001

A game question

There's lots of interesting critical discussion of games these days. Good stuff from Walker, Klastrup, Frasca, and lots more.

I should remind everyone: I do like computer games.

Let me try a probe, just to make a little tsimmes. Take the last twenty years of computer games -- the whole kit and kaboodle. Put them on a shelf. (Yeah, it's a big shelf) Now look over the shelf, and tell me everything we learn about, say, sexuality. (House rules: no arguments from silence, no metaphors, cigars are just cigars unless it's apparent to players who aren't Professors that they aren't. Chatroom romances don't count; I'm looking for what's in the game, not what the audience brings to the table, and yes, I see the theoretical shortcomings of the previous clause. You understand what I'm getting at. Play along at home; it's that kind of movie )

I think I could construct an answer, but it's not a good answer. And I think I could find an excuse, but it's not a good excuse. Do you have a better one? What've I missed? (Email me even though I'm on the road this week.)

It's been a slow autumn for High Design on the Web, but there's some fascinating ferment at the edges. Christopher Baldwin's Bruno took a sudden dark turn this summer, then packed up and went to Europe; Baldwin's meticulously crafted pen-and-ink is particularly ill-suited to Web comics, and that makes his work singularly fascinating. Scott McCloud has started a series of daily one-hour improvs, an interesting counterpoint to his wonderful exercise, the 24-hour comic. Finally, thanks to McCloud, a gorgeous, atmospheric, and ambitious Web strip: Justine Shaw's Nowhere Girl.

eNarrative 3 in San Francisco is 3 coming in January. Join us!

Scientists have a responsibility to keep mistakes out of the literature, and to correct any that slip through. If you don't fix mistakes, people might spend a lot of time doing experiments that cannot possibly work. If you don't fix mistakes, people might build bridges that fall down.

Artists, on the other hand, have no responsibility to correct critics' mistakes. If a reviewer or a professor has poor taste or dim understanding, that's par for the course. You can't argue about taste.

This makes life here on the border of art and science especially tricky. In the last few weeks, I've bumped into at least two papers about hypertext issues, published in respectable Web journals, that appear to me to be simply unpublishable. One is methodologically unsound, resting upon an experiment that doesn't support its conclusions. The other is poorly researched, failing to take into account any of the four or five key papers on its subject.

What's our responsibility here?

Jessica Mulligan describes the corruption of the software trade press. (She's talking about the game industry, but the same problem pervaded the software trade magazines while they had influence and circulation. Now, they just don't seem to matter much.) "Just as I've always thought it is a conflict of interest for the game mags to rely on advertising dollars from the very companies whose products they review, I've always felt that those who write . . . about games should have helped make at least one." Thanks, Lisbeth Klastrup!

Mulligan writes for Skotos, a developer of text-dominant Web games. Many seem intriguing; has anyone played them? I'm trying to figure out their business model, thus far without success.

Here's a snapshot at part of the pile of books next to my bed: the books I'm planning to read right away. The picture is a week or two old; the pile's bigger now.

I have two more of these piles of books at home, and another pile at work. And there are lots of books I've been planning to read, wanting to read, that I've promised to read, that aren't even on the stack. I've got weeks and weeks of hypertexts, too. Some of these (like the Simon Schama) have been simmering for months. Others (Tournament of Shadows) joined the pile because of The Current Situation. It's a big, big pile. And it ought to be bigger.

How does anyone keep up with this? I know, I know: it's a sophomoric question. But it's a real problem for me, and I'm always falling farther behind. Triage seems out of the question; what now? (Amazon lets you call up a list of everything you've purchased, ever: my list makes interesting reading....)

I've been thinking recently that it would be great fun to have a weekend meeting about Serious Hypertext, Education, Scholarship, and Storyspace. Storyspace II is out and flourishing, Ceres will be out soon, we have lots of exciting new hypertexts and lots of exciting new courses and project.

We're thinking about March 23-24. Probably in Boston. Everything is subject to change; it's just an idea. But if you think this would be interesting, or if you have an idea for the program, or you'd like to lend a hand, please email me right away.

Oct 01 11 2001

Bad Law

Michael Fraase's Arts and Farces may have been the first industry ezine -- a hypertext periodical, distributed on floppy disk, years before the Web.

Arts and Farces is back, and Fraase weighs in with an intelligent (and scary) speculation on how the US Congress, in the name of Security, is planning to make copy protection mandatory for everyone and everything. I don't follow Fraase's argument that this would criminalize PERL and Apache, but it clearly would give Gates, Bertelsmann, and Murdoch constructive (and permanent) ownership of our culture. Also important: Fraase's latest essay argues that, when it comes to software development teams, small is good.

Oct 01 10 2001

Copy protection

Dan Bricklin's latest essay argues that copy protection robs the future, endangers our cultural heritage and subverts the work of librarians, archivists, and curators. He also observes that this is, at heart, an issue of class. "Like the days when 'art' was only accessible to the rich, two classes will probably develop: Copy protected and not copy protected, the "high art" and "folk art" of tomorrow."

I suspect that many artists would find Bricklin's last clause puzzling; High Art, in today's net art world, seems closely tied to the academy. It's the Mass Media that worries about Rights Management. In the long term, though, I think it's important to remember that saving literature from commerce would reduce the artists once again to the status of servant to the Prince and the Priest, to the state and to established ideology. In the growing medievalization of our culture, this is one of the more worrisome developments.

Oct 01 4 2001


In my Hypertext 2000 paper, I discussed "intergrammatical" hypertexts that use overlaid, translucent elements for montage and collage. Here, from the Pergamon Museum in Berlin, is one of the oldest such hypertexts.

I don't know what the text says, unfortunately, but notice how it plays off the underlying image. The image doesn't illustrate the text, the text doesn't document the image, but the two combine to say something that neither, alone, could say.

The colors on this page are changing often, both because I'm reading Cailin Boyle's Color Harmony for the Web and because I'm working to find better ways to make flexible sites with Ceres.

In this version, I've added attributes to my Ceres File for things like BackgroundColor and BorderColor. The default values for these colors define the overall design. If a page wants a different color scheme, of course, it's easy to set different colors. Better yet, if you want an entire section to share a different color scheme, you can simply create a prototype for the pages in that section.

But there are many other ways to do this in Ceres! Perhaps you'd rather set the colors in style sheets: go ahead! Instead of defining a bunch of color attributes, we define a string sttribute containing the name of the style sheet. The default value is the site's default stylesheet, but we can easily set a different style for a single note or an entire selection. Or, we might define an attrbute called BodyTag, containing the entire Body tag in one chunk.

The point is not that Ceres makes it easy to play with colors, but rather that Ceres makes it easy to adapt your entire site -- manually, with power tools, or automatically with agents.

Ceres is simple, but it's also deep: you can get started quickly, but there are lots of ways to do things.

Oct 01 2 2001


A few weeks ago, I asked Donna Leishman for her thoughts about Jimmy Owenn's, especially her (or his) "Peau Nue." Her first reaction was very negative; now, Leishman's taken a second look.

The new discussion is extremely interesting. Notive how a change in frame rate seems to inspire new narrative interpretation: Eisenstein in action! By viewing the work on a faster system, Leishan sees new things: "headless torso lying on her back, someone kissing? (Did I see that or want to see that?) So now I feel something, some emotion, something more than watching loops of a semi-naked body."

In baseball, most great players start out with great tools -- speed, dexterity, a good throwing arm. A few players (most famously Pete Rose) manage to thrive even though they lack the tools; they find a way to succeed even though they aren't naturally capable.

Among writers, is there any better example of a no-tools player than Gary Gygax, the originator of Dungeons and Dragons? D&D, after all, is just a book. The rest of the marchandising came later, and hardly matters. Gygax is not a talented writer. His sentences are often ungraceful, his organization is frequently awkward. Yet the original D&D pamphlets had tremendous impact on culture (and cyberculture); few books have had greater long-term impact on cyberculture.

Jan 04 1 1904


After dinner last night, I was talking with open source software engineering guru Walt Scacchi, and with Doug Engelbart, the guy who invented outliners, the mouse, and lots of other things you use all the time.

It was a Memory Lane kind of evening -- Frank Halasz was there, too, a guy who had a huge influence on my career though I'm sure this would surprise him.

Someone recalled the early personal computer flea markets. "I was there," said Doug. "I saw Bill, in a little flea market booth, selling software off his tailgate," Scacchi recalled. "I saw Adam Osborne on his soapbox, telling us to give away the computers, to give away the software. 'How can we do that?', people asked him. 'We'll all make money selling them the manuals!'"

"Well," I said, "it turned out that it was Bill who got to buy the big house." That might have been the real start of open software, right then and there. Scacchi's almost convinced me that the open software economy might work, even without the cost-shifting arbitrage that currently seems to supply so much of its fuel.

And that's still, in a way, the core of the problem. 'Indirect economies suck,' said Diane Greco, and she is not wrong. Computing may have become a service industry, but the needs of a service industry are not the needs of research and innovation and progress. It's a puzzle and a problem.


Adam Osborne was a wonderful tech writer. He built a terrific business on his model, writing a wonderful series of books on microprocessors back when microprocessors were new. I still own mine. It was brilliant: from sentence structure to production, everything was right. Osborne, and Brian Kernighan, and maybe Bob Horn (who was at Hypertext '87, though not at dinner) pretty much defined modern tech writing. Or so it seems to me.