topics suggestions

[decarlin]

Here are a few extra topics that I think could be covered in 200 that instead of learning in this class I picked up from other graduate students:

Unix utilities (e.g. cut, awk, sed)
Unix filesystems, and the purpose of, for instance, bin versus etc versus usr versus local
Unix pipes
Makefiles
CVS/git/subversion
Advantages/disadvantages of different programming languages

I also think that there is a place for an assignment that makes students find, download, install, and use code from some source outside the department. I think that we should emphasize the fact that most bioinformatics tasks already have code written for them, and its just a matter of learning to find it and apply it yourself.

[KevinKarplus]

That's a lot of topics. I can get some of that into 200, but probably not all. Could you prioritize?
Perhaps some could be in 205 as well.

I'm thinking of adding gnuplot either to 200 or 205 this year. Probably 205, as there is one assignment there (the null-model assignment) that after revamping will definitely benefit from use of gnuplot.

[AndrewUzilov]

I second Dan. I've had to explain this stuff to several incoming students and I have had a lot of "wow, they should be teaching this!" reactions. This is why I keep pushing for a lab-like course teaching some basic hands-on *nix stuff. I keep getting told that this kind of thing belongs in an undergrad curriculum (and yes, in my undergrad computer science curriculum we indeed had TAs holding our hand), but I think that for grad students, a few MORE sessions (since Kevin does share some wisdom in BME 205 on this) of doing command-line basics, build systems, practical knowledge, etc. would be tremendously useful.

[AndrewUzilov]

By the way, here are some resources for Dan's points that one can explore on their own:

Unix utilities (e.g. cut, awk, sed)

http://en.wikipedia.org/wiki/Template:Unix_commands

Unix filesystems, and the purpose of, for instance, bin versus etc versus usr versus local

http://www.pathname.com/fhs/

But the rest of the stuff requires a bit of an expert to give a proper tutorial... yes, there is the GNU make manual and so forth, but yeesh does it take a few months of shooting yourself in the foot to understand that build system. And I won't even get into compiling your own source...

[dearl]

I'm going to cast a glance askew here and suggest that instead of taking up valuable BME200 and BME205 class time, and I'm sure we can all agree that in the quarter system every minute counts, we hold a one-to-two day pre-fall quarter CS skills boot camp for incoming grads. It could be jointly taught by a faculty member and one or two graduate students and would cover all of the basic *nix topics everyone always wanted to know but was too afraid to ask, i.e. everything Dan mentioned.

This way you get the biologists (or other non-CS students) "up to speed" as best you can (full disclosure: I came in a biologist), while not boring the CS students to death. I'll also take this opportunity to ride one of my favorite hobby-horses -- To Become Proficient You'll Need to Work on it on Your Own.

What I mean by this, and in this context, is that if students come into the department lacking much experience in the basics of living in the land of *nix (as I did), then they aren't going to get that experience in a BME200 lecture, or even in my CS bootcamp idea. These proposals are really only going to, at best, lay out the landscape that the student will need to master on their own (csh/bash, sed, awk, cut, pipes, Makefile, VCSes, emacs/vim/nano, etc individually take anywhere from hours to years to master), and at worst waste precious class time.
I would have really benefitted from a bootcamp and if the department decided it was a good idea, I would happily contribute to making it happen.

But if the choice is between spending class time on the basics of *nix or expecting the grad students to learn it on their own, I'm for asking students to learn it on their own.

d

[KevinKarplus]

A boot camp is a good idea, but many of our students first appear on orientation day or (worse) on the first day of class.
If it could be done on the first weekend after classes start, it might work.

Of course, for the information being proposed, almost any time in the first quarter could work.

[AndrewUzilov]

A boot camp sounds like a good idea. It will depend solely on volunteer labor to run, though, so have to be careful not to lose steam... but I'd be glad to volunteer if I'm not too busy.