if you have any idea in advance of what you might have to teach from, it's definitely helpful to think about how you will teach a book early, and not figure it out week-to-week as a lot of people do the first time (and as some people do every time). or maybe at one level of remove, to think about what one actually needs to know from a subject.

an example that comes up a lot is rudin's analysis book, specifically his chapter 2. it looms large in a lot of classes because it's where some programs try to fit in general metric space topology, and maybe even general topology, without having to devote a separate class to it. but from the point of view of the rest of the book, much of that material is never used again.

so you might really go hard on it, or skip it, depending on where it fits into a program (or, if you have the ability to shape it, where you think it ought to fit in)

That’s interesting, when I think of lecture notes, I think of this: cs.cmu.edu/afs/cs.cmu.edu/academic/class/15859-f11/www/notes/…

sorry, I do everything on this site with an Ipad, and the return on this invokes a send instead of new line, so forgive me while i type a response to the above

@leslietownes So my plan is to teach a “Dynamic Programming” Graduate Seminar to a few professors at my college since the college doesn’t offer it as a course, and I’ve independent studied quite a few subjects to get to that point such that I can (hopefully) create this as a course via the seminar. My goal is to produce something akin to the lecture notes above from multiple books, homework, and a few projects. Though, the above is a lot of work to do and prep for

and I was wondering if there exists guides on how to create the above from an instructor’s perspective because I cannot find any

to follow up on the “plan” point - the professors agreed and such, and are excited for me to do this, but it will in the upcoming Fall, so my own plan is to get much-much ahead of it now and over the summer because then I can address fault points before it all starts bc there’s soo many points the lectures can go wrong if i’m not prepared enough

last point: these notes would be something I would give to students while also teaching from in class

ah, ok. i don't know of any general guides for that kind of teaching. when the audience is professors or graduate students, there are fewer constraints - more opportunities to personalize the material to the audience, and not as much stuff you can 'do wrong' as a lecturer.

there are aspects of the structure of the notes that i like a lot. it is mostly definitions and statements of results. (that isn't to minimize the work that goes in to preparing something like this - the selection and sequencing is really hard to get right, especially if you haven't done it before.) that's a good level of detail, even for a less sophisticated audience.

it's tempting to want to include proofs of everything, 'worked out examples,' etc. (i see some of this in the notes) but a lot of that stuff is better left out. it's something i might put in a set of notes that you use in lecture, but maybe not share.

you have to ask yourself how realistic it is that someone will learn something from an example that is worked out in a piece of paper, when (at this level of sophistication) it might be more likely that they would just learn by doing it themselves, or (at a lower level of sophistication) they would need to see the examples and proofs worked out in real time, as part of a discussion, in order to follow it.

from a time management point of view, it's also where typos are more likely to sneak in, and be harder to notice/fix, and confuse the unlucky person who does rely on your notes. or mislead a person who has an old, unfixed copy of the notes and not the fixed copy of the notes. this is less likely to happen or matter with definitions and statements of results.

so yeah, if i were doing what you were doing, i'd probably start by distilling whatever i wanted the course to be about into a list of definitions and results. not with an understanding that the list "is" the course, but that the list is going to determine the course, and maybe be most of what the public-facing lecture notes focus on.

it's cool when you can design something like this from a fairly blank slate. the compromises involved with not being able to choose your starting materials (e.g. being forced to use one textbook or one approach) are sometimes the worst part of teaching.

having to cover X material in Y way, or deliberately avoiding Z because you can't assume that the entire class will have exposure to a prerequisite, even though Z is in the book that you're forced to use and you have to snip it out of the middle of a chapter that you otherwise do need to cover.... ugh.

@leslietownes I got a glimpse of this notion from one of my professor's notes before, he had a lot of worked out examples for cases and sub cases of problems of what has been asked before in class such that if/when he messed up he had something to help him in the class-runtime-crunch, but it wasn’t something he shared with students

another good reason to leave examples and proofs out is that it makes that version of the notes easier to re-use, if you change the underlying materials. :D

@leslietownes that’s actually a really good point, though while I would be teaching this course to three professors, the goal should be grad students, so I may need to focus the audience more towards and that and pretend the professors are kids who just finished non-linear optimization and such - it would be a better structure for the class when i hand over everything to the professors at the end of the semester so it’s immediately reusable and deployable

this is a lot to think on, so I’m gonna stew on it for a while, but thank you for all the advice!