The trick, for me, has been to develop a feel for how much rework is optimal. When you are a part of a scrum team and you understand what the people in control of the organization want to do with it, you can learn to sense how likely it is that code will be backed out, or that a feature won’t be recognizable in its ultimate form, and you can use that sense to gauge how much thought and drafting time you should put into a given user story at a given time. You have to accept that some text will have to be discarded, because you thought it would be needed and it wasn’t, and that some text will have to be created right up against deadline because you didn’t see the need for it until the last minute.

In the early stages there is inevitably some wasted work. But if you can stick with this approach over the life of a project, you can end up working far more efficiently, on the average, than you ever did the old way. (That’s not a trivial “if.” Even the most committedly agile manager will at some point give in to the temptation to ask for a “doc outline,” or a “doc freeze,” or some other artifact from the old days.)

The ironic thing is how much this resembles the way we’ve always worked, minus the pretense. When I told myself I was working from a functional spec, I was always really working with one eye on the spec and one eye on the real development that was happening, which sometimes followed the spec and sometimes led it. I saw some good specs, but I never did encounter one that I could put through the classic semantic-reversal exercise that in theory could have auto-generated the help content and put us all out of jobs.

Time frame: Web app development tends to work on a faster cycle than packaged software. You can’t ship a new book with every release. You learn to work incrementally. Your work is always ready to go, but never fully finished.

Structure: I’m not sure the long-form narrative approach was ever really that good for users, but in web apps it is intolerable. SaaS forces you to finally break with the book-on-a-screen model and prioritize getting the right bit of info into view at the right time, without regard to where the reader came from or where they are going. Kicking the habit of saying “above,” “below,” “before” or “after” has surprisingly profound consequences for how you develop and package information.

Audience: The first thing I did in SaaS was to move the help content out to an open site on the web, where Google could see it and potential users could stumble across it. Tech docs have always been part of selling a product, and with web apps you can’t avoid it. Also: on web app teams, developers have even less bandwidth to review docs, so you learn to get feedback other ways, especially by leveraging real-life users.

“Leverage:” There’s a word that hasn’t been around that long, as a verb at least. In Marx’s time they used the word “exploit” for almost exactly the same idea. During the 20th century, “exploit” moved closer in tone to words like “oppress” or “cheat,” and “leverage” has pretty much taken its place as a more or less neutral word for using someone or something to get something done. A lot of casual readers read “exploit” the 20-century way, and not the way Marx used it. Marx was no fan of mistreating people, but when he talked about exploitation he was really just talking about how a company can leverage the work of an employee to make money. If a pro basketball player earns a million dollars during a year when he enables his employer to earn 20 million dollars, he’s more exploited — more effectively leveraged — than a farmhand who gets paid 20,000 a year and enables his boss to make 50,000, even though the farmhand undoubtedly gets treated a lot worse.

Top four things I’m hoping to achieve at this conference:

• Make contacts with people who are into Drupal and DITA. (Had a chat with Kristof last week that got me all hyped on the possibilities.)
• Learn more ways to automate production processes. (I’m still spending too much time pressing buttons and watching stdout flash by.)
• Get my help.collab.net site evaluated by someone who knows what they’re talking about. (To help prioritize the long list of new stuff I want to do with it.)
• See if Mike Hughes has brought his Dobro. (Never rule out the possibility of some parking-lot picking.)

In the meantime, I’m liking this guy’s suggestion to help make it harder for the bad guys to shut down WikiLeaks. Basically, you provide a dummy web address with “wikileaks” in it. When people find it, via a Google search or whatever, and click on it, they are passed straight through to the real WikiLeaks site.

Mine is here: http://wikileaks.tedtedted.com

Anybody that has a web site can do this. Let’s all! Your hosting provider probably has a web control panel where you can tweak your hosting settings. Use the tools there to create a subdomain (you can have any number of subdomains for nothing if you’re paying for a domain) and have it redirected to 213.251.145.96.

The next step would be to actually mirror the WikiLeaks content. Doable, but more time-consuming for someone like me who’s not all that technical. I’ll let you know when I have it set up.

I marvel at how much of this stuff was wasted on the 20-year-old me. The ability to read things at multiple levels didn’t really come in for me until my, I don’t know, mid-30s maybe. I guess I was vaguely aware that different things were going on; it’s the interplay between them that’s new to me. Borges doesn’t just write a story that’s simultaneously a police procedural, a disquisition on the Kabbalah and a mathematical exercise; he builds each level so that it interlocks with the others. He makes them talk to each other.

How plain old funny this stuff is had also eluded me. As a literary undergraduate, you’re coached to take everything with total seriousness. It’s not until years later that you allow yourself to notice that this guy was, basically, playing. The more the narrator of “Pierre Menard” sucks up to his rich salon patronesses, the more his hate shines through. You can’t not laugh. These narrators are so unreliable you have no choice but to trust them, and so earnest that you can take nothing they say seriously.

In Borges’ world it only takes a few pages to put you into the habit of reading absurdly close. In a description of a found note on page 154 of my copy, the word “handwritten” is misprinted as “manuscristas,” rather than “manuscritas.” I don’t know what to make of this. It comes in the middle of a segment about obscure Talmudic reference texts. The error seems not quite random enough to be a real error, not quite suggestive enough to be intentional. The collection was first published in 1956; this was the tenth printing of just the paperback edition, after a long run in hardcover. (I got it in 1982; I know this because I solemnly, undergraduately, wrote the date on the title page.) How could a typo like that survive so many years of examination? What was Borges trying to pull? Right now I prefer to think he wanted me to write this paragraph. “Even this, perhaps, was foretold.”

Once I was in a seminar course on Latin American “boom” authors. This was the early 80s, when writers of political commitment were all the rage. You didn’t even pay attention to an author unless you could find some social or political agenda to celebrate in his work. Fair enough, I thought; when the writing is good, any motivation will do. But Borges, a rich, reactionary old bastard by all accounts, wasn’t doing well in this atmosphere. His action, such as it was, took place in Vienna and Dublin, not on the battle lines of the oppressed. His characters were aesthetes and pedants. His plots were puzzles. No one had time for that.

The brilliance of the work, therefore, presented a kind of a puzzle in itself. I remember one guy who came up with an ingenious workaround: the writer displayed, he said, “un compromiso vital” — a commitment… to life! Perhaps he wasn’t in exile like Cortazar or murdered like Neruda, but at least he was, um, alive. Yeah. I thought this was the most amazing bullshit, and I admired the guy for his cheek. (And for his ability to get through an entire pack of unfiltered cigarettes in one hour of class — I hope he’s still alive.)

I was, of course, wrong. Maybe the sadness that permeates these stories just isn’t visible to a 20-something. So Borges is all about the ambiguous interplay between appearance and substance, between word and thing. So what? How does that really work? Here’s how: the odd self-assurance, the Rubikian self-involvement, is the appearance that masks, and signals, the desperation, confusion and loss that make up the substance. These stories are set in a world that was falling apart, the world between the first and second world wars. Pierre Menard madly pounds out Quijote in Belgium while armies are massing to invade. Funes is condemned by his inability to forget; Lonnrot by his eye for the immanent patterns in everything. Everyone here is doomed; their only hope is to find some sort of order, some periodicity, in a world that shows no evidence of any such thing.