Irreal

The Verge has a story that is at once hilarious, infuriating, and deeply disturbing. In this time of COVID-19 much education has moved online. Except for scale, this isn’t really new. It’s been possible for many years for children to get their education online. Some students—especially high school students—prefer that option to dealing with all the nonsense a traditional high school education entails.

Now, of course, we’ve had to scale up to include almost everyone and those providing that education are new to dealing with online learning. It’s hard for teachers used to the classroom but not impossible. I have a family member who is a middle school teacher and she has mostly tailored her classes to run through Zoom. Mike Zamansky, of Using Emacs fame, is a university teacher and he, too, has adapted to running his classes online. Neither he nor my family member thinks it’s as good as in-person teaching but they’ve adjusted their teaching to provide the best pedagogy they can.

Sadly, not everyone has adapted as well. The Los Angeles Unified School District, for one, has attempted to automate the teaching and grading of several courses. That’s not necessarily bad but, as the article relates, the “AI” grading of short answer tests involved merely scanning for keywords. Of course, the students immediately discovered this and started gaming the system by simply answering the question with a word salad of any keywords that might be applicable to the question. Sometimes, even just cutting and pasting the question into the answer would work.

The school district has, of course, tried to frame this behavior as cheating but I think it merely demonstrates that the students are way smarter than the school district. If you’re going to test students, then you’re morally obligated to carefully and rigorously evaluate the answers to those tests. When you outsource that to a simple minded algorithm that simply scans for keywords, it’s you who is cheating and you should shut up about students who found you out and responded appropriately.

As I say, it hilarious that the students discovered this and reacted suitably. But it’s also infuriating and disturbing that those charged with educating our children would do so with such an inadequate and immoral system. There are plenty of teachers who are proving daily that we can do much better.

Apologies to those of you who aren’t members of the cult but this is another post on Apple products. For the last year or so I’ve been using a wireless charger for my iPhone. I like it because it’s easy to pick up the phone if it rings at night or when I’ve set an alarm. When I used a wired charger I always got tangled up with the cord and ended up flexing the cable.

A couple of weeks ago, the wireless charger stopped working. It was an inexpensive Anker that had worked perfectly for me so I just ordered a new one. Sadly, the wireless charging still didn’t work. My next thought was that it must be the power supply but before I changed that out, I put my wife’s iPhone on the charger. It worked fine.

I spent a day or two wondering what to do when it occurred to me to do what I should have done in the first place: I asked DuckDuckGo. It turns out it’s a known problem as explained here. The TL;DR is that you need to do a hard reset on your iPhone. Apparently, just restarting the iPhone sometimes helps but it didn’t in my case. As soon as I performed the hard reset, everything starting working again.

The only problem is now I have two wireless chargers. I’m sure I’ll find a good use for it. This post will probably be of no interest to most folks but if you were wirelessly charging your iPhone and it stopped working, try a hard reset. Also take a look at the link above.

One of my favorite development strategies is using finite state machines. They’re not right for every type of problem but they have greater applicability than you might think. I’ve written several FSM-based applications almost always in C. C isn’t the only possibility, of course. You can write an FSM in nearly any language.

Arpit Bhayani has an interesting post that describes writing FSMs in Python. That language is fairly powerful and offers several possible implementation strategies but Bhayani suggests a particularly nice one: coroutines.

I think of coroutines as being the inverse of generators. Rather than having yield return a value and suspend executions as generators do, in coroutines, yield suspends the execution and waits for other code to send it a value. It’s very much like a blocking read.

Each time the FSM gets a new event, it sends it to the coroutine representing the current state through the send/yield mechanism. The coroutine uses the event to determine the next state and do whatever processing is necessary for the state transition. Bhayani offers several examples illustrating the technique. Take a look at the post for some actual code.

A surprising number of people hate Amazon. I say “surprising” because I don’t understand their animus. Sure, they send us a huge amount of what amounts to spam, have an unfriendly attitude towards organized labor—at least as concerns their own labor force—and have effectively cornered the ebook market. But consider: if you’re a customer is their advertising a price you’re willing to pay to have reasonably-priced, needed staples delivered to you door in this time of COVID-19? If you’re an employee, are you happy to have a job to go to, even if—despite Amazon’s precautions—there might be some risk involved? Before you dismiss that question, ask a restaurant worker what they think. My life during the last few months would have been considerably more complicated without them.

Many people are advocating for the DOJ to take antitrust action against Amazon. Byrne Hobart has an interesting article that considers the case for this in his newsletter, The Diff. Such things are not really in Irreal’s purview but I think the article will be interesting to many Irreal readers. Hobart begins with a long section on how Amazon works and makes its money. It’s a lot more complicated than you probably thought. Then he discusses antitrust and why, according to current antitrust standards, Amazon is not a viable target. But, he argues, even though Amazon is completely innocent on paper they nevertheless feel as if they’re violating antitrust laws. The TL;DR is that for Justice to make their case they would have to press a novel interpretation of antitrust and that it’s by no means sure that they’d prevail. Still, Hobart believes that there’s a good chance that the courts might accept the novel antitrust standards and that Amazon might be found guilty of them.

There are two things about this article that I found particularly interesting and that make it worth reading. First is the explanation of how Amazon works. It’s really fascinating and probably not at all what you thought. The section on antitrust mostly discusses predatory pricing and that it’s a lot more difficult to get away with than you think. It’s really interesting even if you don’t care about economics or antitrust theory. Definitely worth a read.