My brother pointed out this link to a story about the Haskell Free Library, which sits, intentionally, part-way in Stanstead Quebec, and part-way in Derby Line, Vermont. New border crossing and security rules are set to cause it some serious problems. It's a good illustration of the kind of disruption that happens when security concerns meet the reality of border communities that have been open for a very long time.

Stanstead is very close to where my grandparents lived, and I've been in that library a couple of times as a child. It'll be a shame if they have to set up customs controls between the stacks and the reading room.

Jim Robertson continues on with his ranting about DRM, which reminds me of my own recent experiences both with DRM and with people just carrying on with old design decisions in places where they don't make sense.

I have a DVD recorder. It was surprisingly not that much more expensive than a reasonable DVD player, and it also supports DVD-RAM recording, so you can watch things as you're recording them. Sort of the poor man's DVR. We've used it occasionally, but not that much. Then the actual DVR broke, and we were a few days without one. And during a Champions League week. So it started getting a lot more use, and its features really started standing out.

One is its tendency to develop errors over time and require resetting to default settings. We see this regularly with DVDs, where it decides everything is out of its region. Last week I saw it decide it wouldn't record a normal television channel because it didn't have permission. This is one of my main objections to DRM, that almost any failure mode is going to involve making your content unavailable. And this is controlled by software, and we know how dependable it is.

The other thing that struck me is the old VCR-ish feature that to make it start recording on a schedule, you need to turn the machine off. That never made a whole lot of sense, even on a VCR. When you've got a DVD-RAM, it's completely stupid. So, suppose you have two timers, to record two sporting events that are on consecutively. And you start watching the first one after a half-hour's delay or so. Now the machine is on, so the second recording won't start. @#$@#%@.

As I mentioned in the previous post, I've started using DabbleDB recently. I had used it in the beta period before to organize last year's Smalltalk Solutions presentations. It worked all right for that, although it had some problems dealing with large text fields like abstracts and biographies. But I didn't see that much use in it for me, at least not enough to pay a monthly fee for it, even if it was a Smalltalk application, so I let the account lapse when Dabble went out of beta.

Smalltalk Solutions is being held (April 30 to May 2nd in Toronto, if you're curious) in conjunction with the IT 360 conference (renamed/upgraded from last year's Linux World/Network World). The main conference organizer wanted the information in a spreadsheet. So I rather tedious cut and pasted the appropriate fields out of the sorts of e-mails you get from web forms into the spreadsheet. And then I needed to do some edits. That's really tedious, and if I thought early versions of DabbleDB had trouble with large text fields, that's nothing compared to trying to edit them in Excel cells. So I decided to give DabbleDB another try, signed up for a 30-day trial, uploaded the data in there, and was happy to find out that large text field handling was much nicer.

Another thing that's been going on is that I've been trying to become more organized. I've been trying out the "Getting Things Done" method for doing that, which seemed relatively simple and was described as "geek-friendly". I think I'm liking the general mechanisms, but I started out keeping tasks as pieces of paper in folders. That doesn't work that well for me - there's not a lot of desk space for folders with three computers, and I found it awkward to look through the folders. The obvious thing to do is computerize this, and the obvious way to do that is to write my own program to do it. But, attractive as that idea is, it takes time, and part of the reason I'm trying to get organized is shortage of time. And having, signed up with DabbleDB, it seemed like a worthwhile experiment to try managing tasks that way.

So far (only 8 days later) I'm quite happy with it. It's online, so I can get to it from anywhere. So far that mostly means "anywhere in the house", but with five machines total floating around, that's still convenient. And the various views and filters work out pretty nicely (I could really wish for the ability to do "OR" in filters, but I managed without it). Now I have a variety of nicely organized lists of things that I should have already gotten to, but haven't. I'm not using all that much of Dabble's capabilities so far, but I'm thinking about other things I might be able to use it for.

One of the things that always catches my eye are questionable or misleading uses of statistics and other numerical data. Sometimes I don't even have the facts to know if they're questionable, there are certain warning signs. "Between March 14 2002 and June 27 2003, X happened". Why were those dates chosen, and how different would the numbers look if they were slightly changed.

A couple of favourite examples. In one of the major Canadian newspapers a couple of years ago there was a big article bemoaning the state of Canadian labour productivity, in particular relative to the US. Leaving aside any questions about e.g. the role of relative currency values in such measurements, there was a wonderful anecdote about the decline of Canada's economy . I don't remember the exact numbers, used but in general it said that at the end of the Second World War, Canada had the Nth largest economy in the world (N<5), and since then it had declined to M, (M>10). This might sound bad, until you think about the state of the world economy at the end of the Second World War, and started making a list of countries whose economies might have been smaller then, but which might reasonably have passed Canada since then. You know, Germany, Japan, France, England, Russia, etc.

Another favourite example comes from sports. It works at any time during the recent run of Super Bowl victories for the New England Patriots. The specific example I remember is from when they were approaching their third in the 2004/5 season. A sportscaster illustrated how amazingly good they were, saying that they had won eight out of eight playoff games, and won two out of the last three Super Bowls. Granted, they were very good, and went on to win, but when you think about that statistic for a moment, you have to wonder how they could fail to lose a playoff game, yet only win two out of three trophies. The reason, of course, is that they failed to make the playoffs at all in the year that they didn't win the Super Bowl. The statistics are absolutely correct, but not nearly as impressive as they initially sound.

So why does this have anything to do with computers? An example came in my mailbox the other day, the most recent shot in Oracle's rather aggressive advertising campaign against SAP. The page was mostly blank, with something made to look like a clipping in the middle with the words "SAP Customers are 20% less profitable than their industry peers". That certainly sounds bad, and the idea of the campaign is to get people to switch to Oracle Applications. It seemed like an effective argument. I'm sure one could argue about the methodology and the specific results, but what struck me a few minutes later was, "What about Oracle Applications users"? If I were going to say that my competitor's customers are 20% less profitable than their peers, then capping it off by saying that mine were e.g. 20% more profitable would really be compelling. And if I don't say how profitable mine are, it might make you wonder why. Perhaps Oracle just don't have any statistics, but you'd think they'd be able to get some pretty quickly. Or perhaps the reason to not say anything is that the answer wouldn't actually sound that good. To say that your competitor's users are 20% less profitable than their peers, while yours are e.g. only 15% less just wouldn't make a great ad campaign.

Assuming that the SAP numbers quoted were remotely accurate (and you always want to be careful getting your facts from advertisements), and that my guess about the reason Oracle Application users' profitability wasn't quoted is broadly right, the whole business makes me think of a point Jim Robertson likes to make about the tradeoffs between corporate standardization of applications and custom applications for specific purposes. Perhaps the users of all such large company-wide integration applications actually suffer relative to those with more ad hoc solutions. That would be an interesting thing to have measured empirically. Even if it turned out to be true of course, it might mean any number of things. For example, it might mean that companies with a strong emphasis on central control, which led them to try to standardize software, did less well in general than more decentralized companies, for reasons that had nothing to do with software. Or any number of other things.

My main point, insofar as I have one, being that whenever someone quotes "facts" at you, you always want to be very careful to think about what important details they might not be telling you.

OK, nothing to do with Glorp, but Via The Fishbowl comes this absolutely brilliant skewering of Paul Graham's "Hackers and Painters" essay, and a few other things along the way, by Maciej Ceglowski.

I'm going to be at OOPSLA this year, and I'll be on a panel, which should be lots of fun. The topic is "The Great J2EE vs. Microsoft .NET Shootout", and it includes some heavyweight people from those communities. There's a brief description here.

My role in this is to be the designated "wildcard". There are more than two useful technologies in existence, and I get to represent everything else. That should be particularly fun, since I don't represent just Smalltalk, but the union of all alternative technologies. Linux and LISP, objects and open source, functional programming and FORTH, all rolled into one.

In this, I could use a bit of help. I work in Smalltalk, and spent a good bit of time in Java, so I'm pretty familiar with those areas. I'll need to brush up on some of the other technologies out there. Any advice, pointers to good references welcomed. I've done some reading on .NET and heard a few talks, but never done any serious programming with it, so that definitely needs some work. You can either post comments here, or email me as knight@acm.org.

Thanks in advance....

I spent last week in Cincinnati, at Cincom headquarters, in a variety of meetings. The meetings were quite good, as meetings go. The last day was spent in meetings with a large customer who has some very important systems dependent on VW. They had some very nice things to say about Smalltalk being the most stable piece of their information systems. So Friday evening I was due to fly home, from Cincinnati to Ottawa, via Pittsburgh on USAir. About midway through Friday I get a message that my flight is cancelled. I get re-arranged to fly on Delta to Washington Dulles and then from there to Ottawa on Air Canada. That's flying south to go north, but at least I can get home.

There are a few minor complications. I had an e-ticket on USAir, because these days everything is e-tickets by default. But if you need to switch airlines, those don't work, so you have to go to the USAir counter, wait in line for half an hour for a piece of paper, then go to the Delta counter (at the extreme other end of the airport) and wait in line half an hour to give them the piece of paper. Of course they then tell you that this is the wrong piece of paper, and you need a different one. However, they were nice and didn't make me go back to the USAir counter, but phoned and made the USAir person come bring me the right piece of paper. So far so good, and I've still got 40 minutes before my flight. They even gave me a coupon for a free drink for my trouble. Woohoo. Off to security.

At security, they grouped us into lines. I seem to be in one which is extremely slow-moving and longer than all the others. They don't, of course, explain, but I assume that it's changing my flight at the last minute, perhaps especially so with Washington included in the itinerary, that has entitled me to extra bonus security screening. By the time I get through the line and through this, my flight was scheduled to leave five minutes ago. The security people were very nice, and explained that I really shouldn't have to go through all of this. As it turned out, however, I was lucky. My plane was delayed, and I got onto it in good time.

It was a short flight. They didn't serve drinks, making the coupon rather useless. And since it was late leaving, it was even later arriving, and I missed my connecting flight and had to stay overnight in Washington. Fortunately, the next day I allowed lots of time, and the security lines were much more sensibly organized, with the ability to do extended searches at any of the checkpoints (the Air Canada agent was unimpressed by my explanation that the security people in Cincinnati said I really didn't need to go through the extended search). It was an uneventful flight, and I was home by around 10:00 am.

This leaves me wondering about airlines. I'm certainly not impressed with USAir for ditching me and putting me through all of that. I was annoyed at Delta in the beginning, but in the end I can't really fault them for too much - it's more the security setup in Cincinnati. And being annoyed with Air Canada is standard. How many airlines do I have to be annoyed at before concluding that the whole thing is just a bad idea.

I'm finally back home after Smalltalk Solutions and a vacation detour through Los Angeles. Unfortunately, we also managed to pick up a nasty virus in California, and I've been completely useless for the last 3 days.

Many people have talked about various conference sessions and their impressions of them. I'd just like to say that from my point of view it was a great success and I'm extremely happy. I'd like to thank everyone who participated, and hope to see you all again next year.