Irreal

Well, not really a poll but at least some random opinions. A while ago, I mentioned a discussion I had with Perry Metzger about pair programming. Metzger, who is a serious person and well worth listening to, is a huge fan and says that he’s never been so productive as when he’s pair programming. I, on the other hand, imagine that it would be much like hell.

I have always thought that most programmers—for reasons that I couldn’t imagine—like pair programming. Certainly there is a lot of positive commentary about it on the Web. But perhaps I was wrong.

abdullah_ayyash over at the Programming subreddit asks what people think of pair programming and why they like it if they do. The answers are what any reasonable person would expect: some like it, others don’t. Those who don’t like it appear more emphatic in their opinions but the opinions seem evenly split.

Part of the problem, it seems, is exactly what you mean by “pair programming”. I think of it as a formal and planned sitting down together to write code. Others have a loser definition. Their definitions include such things as sitting down to debug a problem together or as a simple mentoring. I’ve done both those things but wouldn’t describe them as pair programming.

Even among those supporting it, my sense is that they see it as an occasional way of mentoring or sharing ideas, not as an everyday way of working.

Sorry, but I’m recalcitrant. I’m not going to do anything that resembles the strict definition of pair programming. Not now, not ever. Still, lots of people feel otherwise and I have no objection to them indulging themselves. Perhaps that’s the nub of the matter: if you pair program because you feel it helps you, go for it. My objection is to mandating its use.

According to an Ars Technica report, ISPs say that their customer service is so good that customers don’t seek other providers. That claim was made by the NCTA-The Internet & Television Association, a cable industry trade association in response to an FCC filing proposing new rules to improve customer service.

The situation is, of course, rife with politics but there are some points worth making:

• Not all providers are the same. I’m more or less satisfied with my provider, Spectrum, although they are far from perfect. On the other hand I have friends stuck with the provider who shall not be named that thinks customer service is a bad joke.
• The growing trend of using never ending phone trees and, worse yet, AI chat bots will do nothing but increase customer frustration and will likely cause more customer complaints. It’s just silly, and in the end, not really serious to claim that those measures will improve customer satisfaction.
• It’s just not true that dissatisfied customers can simply change providers. In Tampa, the main choices are Spectrum and Verizon: by no means the worst but no one would describe them as a panacea. Many areas aren’t even this lucky. They have a single provider that has absolutely no incentive to provide decent customer service.
• Irreal, as many of you know, is always skeptical of involving bureaucrats in our affairs but ISPs are essentially monopolies and the quid pro quo for that status is agreeing to the regulatory regime.

All of that aside, it’s surely humorous to hear claims of customer service so great that customers wouldn’t leave even if they had a choice. That’s not a serious claim and gives one reason to doubt anything else those making the claim say.

As many of you know, Edsger W. Dijkstra was an important and prolific mathematician and computer science. He was responsible for, among other things, the widely used shortest path algorithm.

One oddity of his work was that a substantial part of it was “published” as handwritten notes that he photocopied and distributed to colleagues. These notes were called EWDs because of his habit of labeling them with his initials and a sequential number.

EWD831 considers the question of which of the following notations should be preferred for referring to the sequence \(2, 3, \dots, 12\).

1. \(2 \le i < 13\)
2. \(1 < i \le 12\)
3. \(2 \le i \le 12\)
4. \(1 < i < 13\)

His conclusion is that 1. is the proper choice. That leads to another question: If we wish to use a sequence to index, say, an array or list, what should the first index be? His answer, of course, is zero. You can read EWD831 for his reasons for these conclusions. It’s short and will take only a couple of minutes. It’s a fascinating look at the thinking a great mind and his striving for simplicity.