Once more into the teach, dear friends

Since the beginning of April, David Mazières and I have been back in the saddle teaching CS240H at Stanford again.

If you’re tuning in recently, David and I both love systems programming, and we particularly get a kick out of doing it in Haskell. Let me state this more plainly: Haskell is an excellent systems programming language.

Our aim with this class is to teach both enough advanced Haskell that students really get a feel for how different it is from other programming languages, and to apply this leverage to the kinds of problems that people typically think of as “systemsy”: How do I write solid concurrent software? How do I design it cleanly? What do I do to make it fast? How do I talk to other stuff, like databases and web servers?

As before, we’re making our lecture notes freely available. In my case, the notes are complete rewrites compared to the 2011 notes.

I had a few reasons for rewriting everything. I have changed the way I teach: every class has at least some amount of interactivity, including in-class assignments to give students a chance to absorb what I’m throwing at them. Compared to the first time around, I’ve dialed back the sheer volume of information in each lecture, to make the pace less overwhelming. Everything is simply fresher in my mind if I write the material right before I deliver it.

And finally, sometimes I can throw away plans at the last minute. On the syllabus for today, I was supposed to rehash an old talk about folds and parallel programming, but I found myself unable to get motivated by either subject at 8pm last night, once I’d gotten the kids to bed and settled down to start on the lecture notes. So I hemmed and hawed for a few minutes, decided that talking about lenses was way more important, and went with that.

Some of my favourite parts of the teaching experience are the most humbling. I hold office hours every week; this always feels like a place where I have to bring my “A” game, because there’s no longer a script. Some student will wander in with a problem where I have no idea what the answer is, but I vaguely remember reading a paper four years ago that covered it, so when I’m lucky I get to play glorified librarian and point people at really fun research.

I do get asked why we don’t do this as a MOOC.

It is frankly a pleasure to actually engage with a room full of bright, motivated people, and to try to find ways to help them and encourage them. I don’t know quite how I’d replicate that visceral feedback with an anonymous audience, but it qualitatively matters to me.

And to be honest, I’ve been skeptical of the MOOC phenomenon, because while the hype around them was huge, it’s always been clear that almost nobody knew what they were doing, or what it would even mean for that model to be successful. If the MOOC world converges on a few models that make some sense and don’t take a vast effort to do well, I’m sure we’ll revisit the possibility.

Until then, enjoy the slides, and happy hacking!

Posted in haskell
4 comments on “Once more into the teach, dear friends
  1. Sergei says:

    Great slides!

    But one of them is 404:
    http://www.scs.stanford.edu/14sp-cs240h/slides/parsing-slides.html

    Thank you!

  2. doge says:

    Do you plan to record and publish video of lectures?
    Thank you.

  3. Kostiantyn Rybnikov says:

    Video would be highly appreciated and watched with enthusiasm.

  4. David Young says:

    Regarding the lens slides, it looks like both focusFst and focusSnd have two possible implementations (one of them is not really useful, but it type checks).

Leave a Reply

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>