[Lesson 0] Is Lesson 0 basic enough?

[ahmadia]

I'm worried about a couple of different things that students might encounter when getting started with Lesson 0.

Installation instructions.

How do they get their IPython Notebook environment going? How do they launch IPython Notebook?

I can start working on these as a web page with lots of screen shots using Anaconda's installer.

The students will either use lab machines or have instructors help them install the environment on their computers during office hours.

IPython overview/introduction.

How do they make cells "go". What are the different types of cells? What are those crazy symbols creating the equations? HOW MANY LANGUAGES DOES THIS THING SPEAK?

There are some really great IPython tutorial/overview lectures out there. We should link to them, and consider putting together a web page that gives a really gentle introduction to the IPython notebook as well.

We are going to add a little more help at the beginning of the notebook to distinguish the cells.

Cochrane-Orcutt Procedure

Is the dive into the Cochrane-Orcutt procedure too quick-and-steep?

Its a useful statistical procedure, to be sure, but it involves a convolution as well as some fairly well-developed statistical intuition to understand. I think we should move it to a "Dive Deeper" section for students to explore on their own, and instead have them just consider the simpler problem, which is calculating the Durbin-Watson statistic to decide if the data is serially correlated.

[gforsyth]

Regarding installation procedures, I think the best option (which may be too much to ask) is to have a centralized ipython server (ipython.gwu.edu or something) which students can log into using their credentials. Does GWU use Kerberos for authenticating student users? Stanford set up something similar and was kind enough to put up their code.

https://github.com/cni/ipython-hydra

[labarba]

Types of Cells ---For the first few lessons, I envisage students touching only code, in two ways: (1) they shift-enter in place (agree this instruction needs to be added), change parameters, do it again; (2) they reproduce the bits of code into their interactive window and/or a separate .py script.

(When using the AeroPython lessons in class, I instructed students to actually type the bits of code into their script—on the Canopy editor, in that case— and to avoid copy-paste. The muscle memory of typing the code and the errors that come from typos or overlooked details are pedagogical.)

Thus, I don't see the need to explain the different types of cells early on. In fact, I think it's a distraction. The goal is to have them solving problems with computing as fast as possible, not to teach IPython Notebooks. So, crazy symbols creating equations? I wouldn't even mention it.

I would delay discussing the IPython Notebooks until students have had a good taste for the fun part of computing: getting a solution, a plot, an insight … The principle of instant gratification.

That still leaves the problem of how to get the IPython environment installed on their own machines. We are going to have the students come in to a computer lab for the first lesson. They can either start working right away on the lab machines, which will be ready to go, or bring their laptop and get their environment set up. I agree that a web page with instructions will be needed. But we'll also have trained learning assistants in the lab at certain hours.

[ahmadia]

The idea is that this course should be completely self-contained. If I find the directory containing this course, there are instructions for the instructors, students, etc... No decisions we make now are set in stone, but I think a little bit of extra work in making these modules accessible to others interested in using them will pay huge dividends in the long run.

[ahmadia]

Regarding installation procedures, I think the best option (which may be too much to ask) is to have a centralized ipython server (ipython.gwu.edu or something) which students can log into using their credentials. Does GWU use Kerberos for authenticating student users? Stanford set up something similar and was kind enough to put up their code.

There are a couple of options for cloud IPython solutions. Wakari would be more than suitable for this purpose, and SageMathCloud will likely be stable enough for testing when the lessons enter beta.

[agolding718]

I'm with Dr. Barba on the topic of types of cells. We aren't expecting the students to need to really look into the other cells.

I am, however, also a bit concerned about the Cochrane-Orcutt procedure. As I noted with my commits, I'm not particularly familiar with it, and it will be pretty scary-looking, I think, for new students(this module is meant for Freshmen/Sophomores, correct?) Without having taken a statistics course, especially one that covers that topic in particular, it looks completely alien.

[labarba]

Responding to The idea is that this course should be completely self-contained. … At the stage of this pilot—i.e., the Spring 2014 effort making just three modules and evaluating them with focus groups—I don't want to worry too much about making this material easy to use for others that just happen to find it (if it involves extra work to do so).

After Spring 2014, and with the initial evaluation done (hopefully showing the effectiveness of our approach!), then I am happy to put time and effort into increased usability outside our GW sandbox.

[agolding718]

Sorry for the close. still not used to Github. didn't realize I was closing the topic

[ahmadia]

Okay, I think we've resolved point 1 and 2. What's the consensus on Cochrane-Orcutt?

[labarba]

About the the Cochrane-Orcutt procedure—OK, let's cut this out. This material comes from a lab in Philippe Bardet's course. In fully deployed stage, he can still have a lab with this, only assuming that the first half of the material has been covered in the modules.

[labarba]

And on Gilbert's point … I think a local Wakari server would be awesome! Shall we talk to Andy T. and Engineering IT people at GW and explore this?

[gforsyth]

I think it's worth exploring and possibly laying groundwork for when/(if) this expands with GWU. I just took a look at sage math cloud (very cool) and checked up on my wakari account and they both support git(!).

This is more an administration thing, but in terms of "assignments" we can have the presentation notebooks and also assignment notebooks, with questions, hints, examples, etc... but with code blocks to be filled in/completed. Then students can just git clone the "assignments" repo to their own account (wherever it is) and complete it.

On Thu, Feb 20, 2014 at 9:17 AM, Lorena A. Barba notifications@github.comwrote:

And on Gilbert's point ... I think a local Wakari server would be awesome! Shall we talk to Andy T. and Engineering IT people at GW and explore this?

Reply to this email directly or view it on GitHubhttps://github.com/barbagroup/JITcode-MechE/issues/1#issuecomment-35624916 .

[ahmadia]

@labarba - It would be easiest to just get a dedicated node or two from either Wakari or Continuum without having to manage it ourselves.

@gforsyth - Excellent idea. You tell me what you guys need and I will set up the code for deploying it. For now, worry about getting content in whatever form you'd like to see it.

[ahmadia]

I'm going to close this issue, as all three questions have been resolved (two actions items, one no-go). If you want to keep discussing anything raised here, please start a new issue so we can keep our discussions organized by topic.