The Occasional Masthead

I’d long suspected this, but assumed it was my own cynicism: Clueless individuals believe they know what they are doing. A paper by Justin Kruger and David Dunning of Cornell University documents empirical studies that unambiguously demonstrate that poorly skilled individuals consistently self-assess themselves to be highly skilled:

People tend to hold overly favorable views of their abilities in many social and intellectual domains. The authors suggest that this overestimation occurs, in part, because people who are unskilled in these domains suffer a dual burden: Not only do these people reach erroneous conclusions and make unfortunate choices, but their incompetence robs them of the metacognitive ability to realize it. Across 4 studies, the authors found that participants scoring in the bottom quartile on tests of humor, grammar, and logic grossly overestimated their test performance and ability. Although their test scores put them in the 12th percentile, they estimated themselves to be in the 62nd. Several analyses linked this miscalibration to deficits in metacognitive skill, or the capacity to distinguish accuracy from error. Paradoxically, improving the skills of participants, and thus increasing their metacognitive competence, helped them recognize the limitations of their abilities.

This has significant implications for the processes that bring neophyte coders into an organisation, and the processes (if any) that are used to train them.

Joel Spolsky writes about his new office. For someone who has worked in spaces that have been on the whole drab and not vastly comfortable (apart from the six months or so on the top floor of the Landcentre, Woolloongabba), Joel’s new digs look spectacularly lush. Several things struck me about the design, since it was designed from the ground up for code cutters.

Firstly, the simple things of providing privacy, quiet, a view out the window, and lots of natural light. Secondly, designing data and power cabling and access point solutions that don’t require people to crawl around on the floor. But most importantly, the design recognises that coders spend most of their waking life in the office. An unpleasant office makes for unhappy staff which makes for poor work results.

Joel quotes Philip Greenspun

Your business success will depend on the extent to which programmers essentially live at your office. For this to be a common choice, your office had better be nicer than the average programmer’s home. There are two ways to achieve this result. One is to hire programmers who live in extremely shabby apartments. The other is to create a nice office.

Joe’s article, and more importantly the one by Philip Greenspun, do make me wonder about something. There’s lots and lots and lots of writing around about how to manage coders, and how to get the best results out of having them on staff. So why do so many organisations do such a bad job of it? There’s a PhD in that somewhere.

Eric Burke points out some technological reasons why Word should be avoided for project documentation. It’s an interesting read for me, since I previously advocated against Word because it was a poor choice for document interchange.

Eric points out some problems with Word as a knowledge repository:

1. It doesn’t work well with version control systems;
2. Compared to other formats, retrieving and displaying Word documents is slow (although PDF is slow as well);
3. Linking is lousy;
4. Searching is lousy.

Worth reading.

When people find out I’m a member of the Richard IIIrd Society I am inevitably asked whether I thought he did “it”. In other words, did the hunchbacked monster sneak up the stairs of the tower one night and smother two innocent babes.

Before responding to that question, I’d like to muse on the question itself.

Here in Australia, the fact that that is the first, and usually only, question says some interesting things. Firstly, it suggests that most Australians’ mental picture of Richard IIIrd is dominated, or completely informed, by the stock parody figure. Like a number of comedia dell arte stock characters, the character of Evil Royal Uncle is one that can be easily be adapted for all sorts of purposes. An artist need only dress a figure in black renaissance costume, give him a hunch, and suddenly you have an archetype of Evil. It’s roughly the same as an archetypal Witch, with warts and pointy nose and hat—a set of quick identifying characteristics that remove the need to explain much about the character. This is good, basic theatrical technique, and we’ve been doing it for millenia.

Slightly more informed visions of Richard come from people who know the Shakespearean play, often recalled from dreary school days. I won’t knock the play—it’s wonderful, and the Richard character is one of Shakespeare’s best villians. The play is, of course, the primary source for the quick parodic sketch figure, even though the character in the play is a fully developed one. For one reason or another a handful of Shakespeare’s characters have escaped his plays to live independent lives as basic stock figures: Richard, Lady Macbeth, Oberon and Titania, Romeo and Juliet, and to a lesser extent Falstaff. Nothing wrong with that, it’s a testament to the strength of the original writing.

The big problem is that folk who know the play often make the mistake of thinking that it bears any meaningful relationship to history. Few people recognise that trying to understand the players in the Wars of the Roses period in England via Shakespeare’s plays is like trying to understand Edward II by watching Braveheart. Entertainment, yes. History, no.

Returning to The Question, I’ll begin by saying that it’s a question on the order of “how long is a piece of string”. The question isn’t framed in a meaningful way, because the enquirer doesn’t have the context. So the trite response must be: “Did Evil Uncle Dickon sneak up the stairs with pillows under his arm? Almost definitely not.”

The trouble is that it’s the wrong question. A set of better questions would be “Were Richard’s nephews killed? Did he have a hand in it? If so, why? Why did the boys wind up in the Tower anyway? Why did Richard grab the crown?”. Much better questions. Much more open ended, and the answers are more interesting. I’m not going to give them to you either: you’ll need to come and talk to the society!

Oh, alright: maybe, probably not, see previous, good question, even better question.

It’s been years since I read New Scientist, but I picked up the 13th September 2003 issue a few days ago to read on the train. I might start reading it again regularly, as it was a very refreshing and entertaining read.

This issue included a nice little article on current re-evaluation of the black death. Just to refresh your memory, this was an epidemic that ripped through large parts of Europe in the mid 14th century, wiping out something like half the population. Some areas lost anywhere up to about 80%. Entire villages disappeared, and social chaos ensued.

For centuries the disease has been blamed on a rat-and-flea borne bug with the catchy name of Yersinia pestis. Recently though a problem with this idea has shown up: no trace of Yersinia can be found in the remains of plague victims. At the fringes there has long been some dissent with the rat-and-flea theory. Partially because the epidemiology doesn’t fully support the idea, and partially because the reported symptoms don’t really match Bubonic plague as it now exists. The best guess now is that the bug, whatever it was, was some nasty ebola-like haemorragic fever.

Great stuff.

I’ve got a lot of respect for Joel Spolsky, based on the things he writes. While I don’t agree with all of it, he’s got a knack for displaying common sense, a rare skill in the IT industry.

In his article on interviewing job applicants he not only outlines a sensible way of finding good IT staff, he winds up highlighting what I believe is a fundamental problem with the IT industry: there are too many people in the industry who are not really very good at it.

I was particularly struck by his cardinal criteria for selection: “Smart and Gets Things Done”. I’ve worked with a few very smart people who can’t really get things done. I’ve worked with a few more people who are very good at getting things done (often doing things that later need to be undone, but that’s a different story). And I’ve worked with a lot of people who are neither smart, nor can get things done.

The best IT folk I’ve worked with have mostly come from other backgrounds. Engineering, arts, political science, anything but IT. The thing that has made them good IT practicioners is that they were smart, and could get things done.

Why are there so many people in the industry who aren’t very good? Like Joel, I think it’s the selection process. When I did my undergraduate studies in the early ’80s, the only “computer” courses around were primarily computer science courses, and it was generally held that the purpose of those courses was to generate academics. All well and good. The problem is that 25 years on, a lot of folk are coming through courses that are still oriented toward generating academics, and filling craftsmen roles in the industry. Worse, there are people coming through courses designed to create craftsmen which are horribly inadequate.

Join the generally poor quality of graduates, who have so far only demonstrated the ability to pass the course, with selection processes geared around the assumption that anyone who passes the course is smart. Not a good marriage. The selection processes are not good at identifying smart people. And they’re not good at identifying people who can get things done.

Which is why I’ve had the best experiences with folk from other backgrounds who’ve moved into IT—they’ve been smart folk who got things done in other industries, including getting themselves into IT.

Here endeth the rant.

Over at Accesify, someone Ian L has built a very cool tool to trivialize the effort of turning a set of <ul> list items into a nifty CSS+XHTML menu.

This is definitely one to keep in the toolbox. It’s also a great example of the intersection of open standards and open source coding. Four or five years ago, I would have expected this solution to only really be available through a proprietary (commercial) web design tool such as Dreamweaver. I would have expected the tool to generate swathes of JavaScript, most of it devoted to doing it this way for that browser, this way for the other browser, and throwing up under all other circumstances.

Instead we have a solution built completely with open standards, that is guaranteed to work for any browser that implements the standards, with access to a front end tool given away for free. Sweet.

Over at Baghdad Burning, RiverBend talks about terrorists.

Irrespective of absolutely any other issue, one indisputable fact is that prior to the US “intervention” in Iraq, suburban families generally didn’t have to live in fear that they would be dragged into the street at 3AM at gunpoint.

I have this awful feeling that Iraq is going to go horribly wrong. What part of this picture doesn’t make sense to you: gather a lot of (young, mainly inexperienced) troops, give them lots of weapons, spend six months telling them that anyone looking like a towel-head A-Rab is The Enemy, let them bomb several cities into the dark ages, then ask the same troops to be peacekeepers.

America’s military forces, from inception, have relied on having more guns and more troops and more tanks and more planes than anyone else. Culturally, their armies are attuned to going places and creating havoc. The best thing that could happen to Iraq would be for the UN and EU to send in peacekeepers, and send the US home.

Hey, John Howard, here’s an idea: lots of media folk have commented that Iraqis are very much like Australians in attitude. Why not win a real place in history and dispatch the Australian army to help rebuild the place. It’s what we’re good at.

Tim Bray has written a thoughtful article on software patents from the point of view of someone who has gone through the process of obtaining one or more.

Apart from making some valid critiscisms about the process, in the USA, of obtaining the patent and the flaws in the search for ‘prior art’, he points out that there are some circumstances in which patents on software may be a good thing.

I’m ambivalent about software patents. Personally, I suspect that a copyright or creative commons approach makes more sense when applied to software than a patent. My understanding of a patent is to protect innovative ideas, and prevent others from implementing that idea without attribution. Unfortunately patents are being sought, and granted, for fundamental software ideas and ideas that are neither innovative nor precisely defined. There is good evidence to suggest that a lot of patents are being sought for vexatious reasons, seeking potential financial gain, rather than to protect innovation.

There is a very real risk that either the patent process, or software development, will be harmed and hampered by this. The nightmare situation on one hand is that software development cannot proceed due to risk of litigation and licencing costs for fundamental ideas. On the other hand, patents may become devalued and worthless because they are ignored. It is incumbent on both sides of the fence to think about this and come to some sort of sensible common ground, otherwise nobody wins except the lawyers.

It’s a bit like the old visions of thermonuclear holocaust where nobody survives except the cockroaches. Actually, the similarities are very strong, aren’t they.

This is absolutely wonderful. An interactive do-it-yourself Beayeux tapestry. This is both sublimely silly and silly but sublime.

You don’t want to visit this site if you’re on a modem or with an elderly computer as it is (a) flash intensive, (b) full of big fat images. For everyone else, knock yourself out.

The site allows you to assemble an 11th-century style cartoon, and save it in a variety of formats, including emailing the results.

This could get strange and ugly. As mentioned by Zeldman, and discussed by Tom Coates, Eolas has won one round of a court battle with Microsoft which could see MS, and other browser manufacturers, altering their products so that multimedia is not displayed within the browser window. Yes, folks, Eolas, who have apparently trademarked the phrase “invented here”, claim to have invented and patented the whole idea of object embedding in 1993. So the courts have ordered the web wound back to 1994.

While this is much more daft than Apple being sued by Apple, it may not be silly enough to collapse under its own weight.

There’s a very scary trend here: a US company goes to court to contest a very vague patent claim over very basic and broad concepts, wins (over $US500 million in this case), and forces technological change across the world. Apart from the CEO of Eolas, who exactly wins here?

And if you think this won’t affect folk, think again. Want to embed a movie in your webpage? You won’t be able to. Want to embed music? Flash content? How about a Java application? Nope. The courts have ruled that’s not on.

Dumb. Dumb. Dumb. Dumb. Dumb.

Dave Barry has upset the telemarketers by suggesting people ring them up.

I’m just thinking out loud here. I’m sure you have a better idea for how we can resolve our differences with the telemarketing industry. If you do, call me. No, wait, I have a better idea: Call the American Teleservices Association, toll-free, at (877) 779-3974, and tell them what you think. I’m sure they’d love to hear your constitutionally protected views! Be sure to wipe your mouthpiece afterward.

He also mentions, completely at right angles to any of that delicious rant, that September 19 is international Talk Like A Pirate Day. The official web site even has piratical pick up lines. My favourite?

How’d you like to scrape the barnacles off of me rudder?

It seems to have become a trend, or a fashion, for blogs to have a two or three day hiataus once every so often, followed by an entry containing a handful of links and brief comments, with a title something like “While I was out…”

Herewith, remaindered links.

John Gruber catches up on a few things, including making the interesting observation that

…developers tend to volunteer to write developer software, which is a very different genus than software for normal people.

Diametrically opposed to such deep thoughts: cute little kittens. Aww, ain’t that cute.

Not quite so cute, but still feeling feline, Mirabilis notes a lion-headed statue that takes pride of place as the oldest known statue, at something like 30,000 years old. One interesting, and so far unremarked, feature of this is that this is considerably older than any other piece of non-representational art. Other art of similar age shows things which are real, people and animals, not abstract figures from someone’s imagination. Remarkable.

Sequing into other art: the Tate has put its collection on-line, there is a slew of Matrix comics graphic novels on-line that I’m slowly working through, and an open-source comic character Jenny Anywhere (via Tom Coates) looks intriguing. Oh. My mental editor has just tapped me on a metaphorical shoulder (right or left, your choice) and mentioned that should be “graphic novel”, not comic. I’m showing my age.

Still over with Tom, an interesting muse on mass amateurisation which has received a certain amount of comment around the blogoshpere, but doesn’t seem to have emerged from that dark sea yet. I’m not going to comment, beyond strongly recommend you read both the article and the comments appended to it. Tom has noted something interesting—it’s just hard to tell yet what it is. Food for thought.

Last, but far from least, Joel Spolksy’s test of software team quality. In his own words, its a “highly irresponsible, sloppy test to rate the quality of a software team.” It’s also a frighteningly accurate measure of how sloppy software teams are. Looking back over the last few places I’ve been, I’d say they were 4/12, 3/12 and 3/12 respectively. And I’d challenge you to prove that your programming team is much higher. Provocative, and very useful.

Somewhere along the line, as part of running up MovableType, I installed LWP. I didn’t realise at the time that the installer cheerfully created a file /usr/bin/HEAD.

The big problem was that the standard Unix head command was installed at /usr/bin/head—and the HFS file system for Mac OS X is not case sensitive! So my head was lost. It took me a week or so of weird things happening for me to realise what had happened.

Fortunately, thanks to the crunchy goodness of the internet, I found that David Wheeler had had the same problem, and wrote up his solution. Significant ouch resultant from this one.

Tim Bray muses on rolling up his sleeves and hacking old C code.

I recall one of the basic tutorials for those new to Unix some twenty years ago, it contained the immortal line “For any serious programming, you pretty well have to use C.” This is no longer true, except when it is; I note that a lot of important pieces of infrastructure are still written in C, and I don’t think it’s going away any time soon.

This is a thought that has been bubbling in the back of my mind for the last little while, seeing lots of advertisements for .Net and other high-level environments. I won’t get into the whole “is pushing buttons on Wizard-based IDE really programming?” debate, but limit myself to asking where the gurus of yesteryear are. How many “programmers” out there, particularly recent graduates, could sit down and knock up vi or emacs? More worrying, how is the Unix kernel going to be maintained 20 years from now?