Hacker News new | past | comments | ask | show | jobs | submit login
MIT 6.006: Algorithms in Python (mit.edu)
260 points by helwr on July 5, 2011 | hide | past | web | favorite | 27 comments

It's a bit misleading to describe this as "algorithms in python". You'll find that the course is almost entirely theoretical (as it should be) and that there is a small implementation component for each problem set that just happens to be in Python.

does anyone know of a directory of all MIT CS lectures which may not be available on OCW? would be great to have such a list. extra brownie points for Python... :)

Erik Demaine great lecturer.

This is an understatement.

Erik Demaine is absolutely fascinating. He does brilliant work, and has a great style. I saw that he and Vi Hart once co-authored a paper together-- can you imagine the offspring they could produce?

He is a great lecturer to be sure, but other aspects of the class were not ideal. The course staff was unable to return any assignments promptly, and students were eventually allowed to late drop the class due to the complete lack of feedback on assignments. Although Demaine is brilliant and his lectures were amazing, the class as a whole was a negative experience for many students.

[source: I'm a student at MIT]

Always, Always been a fan of algorithms taught in Python.

Waiting for the video lectures...

We're actually producing a whole new set of 6.006 video lectures this fall for OCW! (There'll also be some amazing new and reorganized material.)

Does the new ones are based on python as supporting language? what all things have changed in this new set? And lastly, any timeline?

Btw. I am Super Excited to hear this :D

Yes, it will still be taught using Python. Almost all of the undergraduate CS courses are moving from Scheme or Java to Python if they haven't already. (One notable exception is the excellent 6.172, about Making Things Fast(tm), which uses C and Cilk.)

The videos will be recorded this fall semester; Erik will be lecturing again, along with the equally awesome Srini Devadas. In the past, when Erik has recorded his own lectures, I recall that they were posted as soon as they were edited. Every time I've worked with OCW, though, one of their staff members comes in and collects materials at the end of the semester.

I'm not sure exactly what the material differences will be; that's the lecturers' call (way above my pay grade!). If you want more than 6.006 offers, though, check out 6.046, which is the follow-on undergraduate algorithms course.

Since we're doing as much of our planning now as possible, I'm curious: what sorts of things do you wish we'd include? Does anyone have suggestions?

I'd love to see the entire course in the format of that first lecture linked above, i.e you start with a naive implementation and gradually improve it both algorithmically and with language-specific hacks until you get orders of magnitude faster version working on your machine, not on paper.

Make every homework a contest on the fastest implementation,e.g. see Tim Bray's Wide Finder benchmark: http://www.tbray.org/ongoing/When/200x/2007/10/30/WF-Results

Python is great for teaching, but I'd let them do the optimizations in any language.

Use problems from your current research as examples/assignments, rather than boring textbook problems.

Let them work with real data: http://www.quora.com/Programming-Challenges-1/What-are-some-... , http://www.quora.com/Data/Where-can-I-get-large-datasets-ope...

Add parallel algorithms to the mix. Ask to parallelize serial algorithms that you explain in the lecture. They should at least start thinking about programming for multicore and clusters, they will thank you for that later.

Hmm...Not sure how useful they would but here are 3 suggestions:-

* Try Adding some more lesser known but useful data structures like Trie,Bloom filters etc.

* Covering recent industry topic - explanation of core concepts that drive bigtable,hadoop, nosql etc.

* Mobile programming from algorithmic prospective.

Here are the videos for the Fall 2005 "Intro to Algorithms" class (http://amps-videoserver2.mit.edu:8080/reports/sma/index5503f...), and you can find videos of his other classes here (http://erikdemaine.org/classes/).

None of the links seem to work on that first page

The links all point to files on the host mit-only.mit.edu, which would suggest a source IP restriction...

Anybody know when they will become available?

Could Khan do each of these in about 7 minutes?

McDonalds can make burgers real quick too. Does not make it a good meal though...

Why the false equivalence?

Are you really comparing the efforts of one man ( for the most part ) who produces free videos for mostly high-school and collage students with the output of a giant global corporation?

Maybe a simple pointer that this does not happen to lie in Salman Khan's area of expertise could probably help.

Your right i guess. My point was that although Khan's videos might be great primers for some topics they tend to miss the boat rather badly when it comes to topics that need a bit more depth like history. That being said algorithms might actually make for comparatively good material to explain in short videos so no need for me to be so grumpy.

Seriously. The thing that always irritated me about university lectures, especially when presented online is that the material covered in 50 minutes could likely be covered in 7. But who would pay $30k for that?

Thanks for the down votes, though.

Too terse. It turns out that some points cannot be made in one Tweet, just as other points cannot be made in seven minutes.

And you need to be a bit more generous to those who still admire fine old art forms like the classic long lecture. It is a genre. It used to be the genre, and now it is just one, but the masters of the big lecture were and are great, and a lot of them aren't going to master the new genre - they were born too late, perhaps - so enjoy them for what they are. Use the high-speed button if you must.

All that said: Yes, you're right, the central awesomeness of Khan is not Khan himself or his lectures - people have rightly pointed out that Khan isn't the teacher to end all teachers - but his relentless pushing of this new genre, a genre that relies on the presence of ubiquitous handheld portable video players. In a world where everyone on the bus and in line at the supermarket has a smartphone, a traditional lecture, with lots of throat-clearing and repetition and class mechanics ("before I tell you anything useful let me talk about the TA assignments") and a long recap at the start of every lecture to let the people who slept through the last lecture catch up... it's tedious. Cut out half of it, deliver the rest in chunks, build a replay system that makes it easy to navigate to and play the individual chunks, and suddenly you're in the 21st century. I've been waiting for this to happen ever since I first used YouTube, and countless YouTube teachers have pioneered the technique; Khan is finally doing the evangelization.

I keep wishing, probably in vain, that TV would go the same way. I put off watching Mythbusters for years because I just couldn't stand the editing: shows are edited for people who tune in and out constantly, so every five minute chunk has one minute of pure review of stuff you just saw ten minutes ago, two minutes of new material intercut with one minute's worth of stuff that you have already seen, and thirty seconds of previews for things that you will see fifteen minutes from now.

Erik Demaine's lectures are plenty dense, and you don't have to pay $30K to watch them. Khan breaks things up into short pieces in a way that is not practical for a university lecturer, because you couldn't ask everyone to show up for 10 minutes at a time.

Sheesh, 3 more downvotes? I get it, HN doesn't like brevity!

Registration is open for Startup School 2019. Classes start July 22nd.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact