Chemistry, Computing, and Whatever

Chem 111 for January 15

One ChemActivity down, a bunch to go.

Today we developed what we need for the structure of the atom, but just as important, we continued to discuss how we connect our conclusions to underlying data and models. We thereby develop explanations that are, we hope, convincing to others. While it might seem nice for scientific conclusions to be so obvious that all we have to do is pass them on from generation to generation, that would also be pretty boring. We choose then to work on both the process for developing scientific knowledge as well as the knowledge itself.

CMIT 140 for January 15

So today was mostly Chris's day, as he began class with a discussion of yesterday's sorting exercise. I thought every student would think it was easier to do the first part of the exercise, in which they could write their own directions for sorting the cards. But one student noted that the constraints of having the instructions in the second part of the exercise actually made it easier for him. With hindsight, that doesn't surprise me--sometimes constrained choices allow us to find a workable answer faster. Because we easily get hung up on looking for optimal answers, it helps to make the search space smaller before we start looking.

CMIT 140 for January 14

Today, we started by noting that just as student schedules change, so can faculty schedules. My colleague Chris Johnson, who pointed me toward the Guzdial/Ericson media computation, and I are now working on this course together. Chris's involvement will benefit both me and the students.

In teaching this class before (admittedly a few years ago now), I've noticed students having problems focusing on the general concepts underlying programming and instead losing themselves in getting the syntax just right. Part of my interest in using the Guzdial/Ericson book was that just putting off dealing with a main method could allow us to concentrate on what single statements and methods could really accomplish. (DrJava helps with that as well, although Chris Johnson and I were surprised to discover that the interactions pane did not force variables to have the type declared if they are initialized when first used. In other words, a statement like cnumber = 140 is permissible even if cnumber hasn't been declared.)

But I wanted to take a step further away from Java at the start, so today we just sorted cards. Given the goal of sorting 20 randomly selected, unsorted cards, could a group of students write a set of instructions for sorting the cards that another group could implement? How well could each group critique another group's instructions? And when forced to live with a small set of instructions, how well could they implement the method they developed (or develop a revised method that works)?

And while I learned where I needed to make the instructions a little more precise, I think students saw the distinction between developing a general process (we're still not using the word "algorithm" yet, although Chris invoked his software engineering self to remind me that process has a specific and different meaning in his world) and implementing that process with a particular instruction set.

The most interesting discussion was students talking about accuracy and efficiency as if there was a tradeoff. I noted that accuracy couldn't be compromised--the cards had to be sorted. What seems like a whole lot of extra steps to us when each card has to be looked at in sequence (rather than seeing the whole deck at once) may not be an issue to a system that can repeat those individual steps very quickly (but has no concept of how to look at the whole deck).

Next class, we go back to the computer: operators, variables, and types.

Chem 400: First class

So it seemed appropriate that for the capstone seminar for the chemistry major and minor, we should explore what kinds of chemistry there are and what are the common themes that connect these fields to what they've seen in their classes.

First, the kinds of chemistry the students and faculty came up with:

Physical
Organic
Medicinal
Inorganic
Polymer
Materials
Bio
Nuclear
Forensic
Geo
Analytical
Food
Guest-host
Atmospheric
Neuro
Water/Oceanic/Marine
Immuno
Pulp & Paper
Textile
Nano
Computational
Theoretical
Pharmaceutical
Bio-Inorganic
Bio-Physical
Bio-Analytical
Astro
Chemical Biology
Chemical Physics
Clinical
Environmental
Electro
Solid state

We know this list isn't complete, but we needed to move on with the class.

Then, in thinking of the common themes that run through all these kinds of chemistry, we developed some broad categories. First, there's what we consider at the atomic and molecular level (what I typically call the sub-microscopic):

Atomic interactions
Bonding
Reactions
Electron Movement
Orbitals

Then, we talked about analysis and characterization, which led to the following (incomplete) set of techniques (that covers much of what we have at Guilford):

NMR
IR
Mass Spec
HPLC
GC

We went back to basics and discussed fundamental measurements--volumetric and mass--that everyone's done since their first chemistry lab. (Which led to the best moment of the night--if you can't remember the word "buret," just refer to "that drippy thing.")

Finally, a bunch of more macroscopic physical chemistry/properties that we study

Thermodynamics/energy
Equilibrium
Kinetics
Magnetic/electrical properties

All in all, a pretty good list illustrating that the kinds of chemistry in the first list have multiple connections to all the themes below.

The rest of the class was an introduction to SciFinder Scholar to allow everyone to start working on their assignments, including doing the research for their first presentation. Coming up next week, presentation skills.

Chem 111 for January 13

The first day of Chemical Principles I always seems to be a lesson in how much paper I can generate (and have to talk about): syllabi, release forms, instructions, ChemActivities. So it was a while before we got out of the discussion about how the course runs and actually did content

At the beginning of this course, the hard part is getting students to use the given data to explain their answers rather than relying on past knowledge. That's especially true when the past knowledge doesn't give a complete answer to the question.

For example, one of the questions asks to describe the difference between a neutral atom and an ion. I asked four groups for their answers. All were different, none were fully wrong, and yet none were a complete description. So that's where we'll start on Thursday.

On a different note, I told my class about yesterday's NY times article that described the amazing things they're doing with science education in physics at MIT. While I'm glad that they quoted Mazur and Wieman and actually made some reference to the research that supports these changes, I'm disappointed that the substantial number of not-so-well-known, often smaller, schools that have been pioneers in these methods (and not just physicists, either). So a big shout-out to the POGIL people (since that's the pedagogy I'm using) and all the others who use similar methods and see how and why they work.

Although the statement in the article that "...attendance counts toward the final grade, and attendance is up to about 80 percent" is on the lines of "News flash: water still wet, sun still rising in the east, Panthers still not going to the Super Bowl."

And directly related to the article, there's no problem in academia that a $10 million(!) hammer can't make a dent in. ($1.25 million per classroom?!)

AI-AL for January 12

While tonight's class was mainly covering the syllabus and the assignments, we tried to do some actual content as well.

I showed both the Robosapiens and the Spykee robots during class. As one member of the class noted, we could have explored the sensors on the Robosapiens more to get a sense of how it interacts with the environment. In both cases, however, we have limited autonomy. It was clear to everyone how I was controlling the actions of each robot. But we did note two behaviors that perhaps give us a sense that there's something more. First, I noted the ability to program the Robosapiens. While the initial version that I have has limited memory, it's not hard to imagine that a system that can hold larger programs could also display behavior that might seem more autonomous. For the Spykee, we noted its ability to find its charging station when told to do so. While there are big limitations on that (it has to be close and it has to be told), the actual way in which it parks itself is compelling.

(A digression: I like how the first feature on the Robosapiens site is "6 different kung fu moves" and that the word reflexes is in quotes in '...programmable "reflexes" to sound and touch stimuli.' I have nothing good to say about the Spykee site. There is nothing here that requires it to be completely in Flash.)

We looked quickly at ASIMO, my go-to example for bipedal robots. We watched the video of ASIMO conducting the Detroit Symphony Orchestra with particular attention to how the members of the orchestra interacted with ASIMO and how we in the class reacted to ASIMO's behavior. We didn't have time to get into how our reaction to the behaviors of man-made objects likely affect how we would attribute life or intelligence to them, but we'll certainly return to that subject as the semester progresses. (I also recommend the shorter video of ASIMO demonstrating avoidance behaviors.)

We then tried to create lists of objects/systems we considered obviously alive, objects/systems that are obviously not alive, and objects/systems that we just couldn't tell. Here's what we came up with:

Obviously alive

Humans
Animals
Plants
Bacteria
Insects

Obviously not alive

Some computer viruses
Rock
Gases
Building
Synthetic material (clothing, concrete, plastics)
Robots
Political systems
Cars, Buses
Laptop
Chatbot
Star

Can’t tell one way or the other

Viruses
Some other computer viruses
Earth
Economic network (capitalism)
Automated voice response/recognition systems
Game AI (e.g. Left 4 Dead)
Automated recommendation systems (e.g. Amazon.com, Pandora)
Massively multiplayer online (such as Second Life, World of Warcraft)
Programs that try to repair themselves (such as in Cisco network routers, servers, etc.)
Social networks of animals, insects, humans, etc.

Lots of interesting points for discussion here, and as you read Bedau, you'll start adding to these lists.

That led to the beginning of the discussion about what criteria something must meet to be considered alive, and our short discussion came up with the following (with no classwide agreement on very many of these criteria):

Reproduction
    But robots can reproduce themselves
Sense of self-preservation
    Self-awareness (might ultimately lead to consciousness)
    But is self-awareness necessary
Learning
Ability to die
    Can make the distinction between the state of being alive and the state of not being alive
Processing energy (metabolic processes)
Ability to act or respond to stimuli
Ability to feel or emote
Ability to perceive the environment
Self-modification by directive/choice rather than by chance

Two main meta-points were raised. First, we might wish to discuss that some collection of these criteria might be necessary for something to be considered alive. Second, there was concern that several of these criteria were more defining of sentience or intelligence than of life.

Having had this discussion before the class has a chance to do any reading is interesting in that it's clear how we're anticipating many ideas that we'll be encountering throughout the semester. There won't be many clear answers, but we will gain greater perspective on how one can advance the discussion.

Coming up...Bedau and the problem of defining life, Kelly and the wonders of complex systems, and Farmer, Belin and how artificial life might fit into our conception of life.

I didn't get a chance to talk about Venter's work toward synthetic life. I'm interested in the distinction between "synthetic" and "artificial" life. The words have definitely been chosen carefully.

CMIT 140 for January 12

First day, first post.

So what happened today?

We looked at some of my current favorite programs as a quick overview of the wide variety of applications that can be developed. Those programs included:

1) Some interesting programs

Monster Milktruck: http://www.google.com/earth/plugin/examples/milktruck/ (Note that you might have to install a plug-in to make this work)

GIMP: http://www.gimp.org
There are versions for Mac OS X, Linux, and Windows.

Protein Data Bank: http://www.rcsb.org
Lots of fun stuff if you like big molecules.

Pandora: http://www.pandora.com
If you just want to find some new music to listen to.

As you play with these and other programs, start considering what's required to make them work. How many different computers and networking components are needed to get the information to you? Each of those must be running some kind of computer program.

2) The syllabus

Remember that the syllabus is always available in the Moodle course. We'll review elements of the syllabus more as we go through the semester.

3) Today's hands-on exercise

The goal of today's exercise was to get DrJava working for everyone and for all of you to demonstrate that you could get Moodle and use your Guilford email.

And that mostly worked. We learned that the first step in debugging a problem should always be to restart the application and see if the problem persists. So I've updated the document in Moodle to add the part about restarting DrJava after adding the extra classpath and correcting the problem with the String folder.

If you have problems getting this to work on any other computer, let me know.

Coming Wednesday: Algorithms vs. programs