Robots Coming to Intro Computer Science Classes

BlueCup writes "Two colleges are hoping to make computer science courses more attractive by including personal robots with the textbooks. Looking to boost enrollment in introductory computer science classes, Microsoft Corp. is working with Bryn Mawr College and Georgia Tech on developing new ways to bring robotics technology into the classroom. Douglas Blank, a computer science professor at Bryn Mawr, said the goal will be to start incorporating the robots in introductory courses at the suburban Philadelphia college next spring. Georgia Tech hopes to start during that term as well. The idea behind the program, Blank said, is to make computer science more hands-on and practical, rather than simply about debugging programs." Update: 07/13 15:52 GMT by T :Professor Blank wrote in with some clarification on one of his statements — read on below.

dougblank writes

"Note to self: when talking to the press, don't use complicated technical jargon, like 'debugging' :) I think what I actually said was 'rather than debug a program to make it give the right answer, the students must debug the program to make the robot behave the way they want it to.'

I think many of you will actually like the hardware, software, and curriculum that we are designing. Check out roboteducation.org/ and pyrorobotics.org. The new version of the software will be based on Pyro, Python Robotics. We think of the hardware as something like an iPod on wheels. The software is also being developed with an open source license. This project is not what many of you guess it might be.

The CS1 and CS2 that we are developing won't be watered down, but also won't be just the standard 'intro to programming, using robots.' It's a complete rethinking of the intro courses."

Comments

The Japanese are doing the same

[Anonymous Coward]

... with sex ed classes.

Re:The Japanese are doing the same

[Das Modell]

Sign me up.

Re:The Japanese are doing the same

[kalirion]

Just goes to show how far ahead they are of us.

Re:The Japanese are doing the same

[Anonymous Coward]

This is gonna be the best prom evar.

Re:The Japanese are doing the same

[worthlessturd]

Don't date robots!
-Crocodylus Pontifex

Great

[jpaz]

As if books aren't already expensive enough. I wonder how much a used robot/textbook will cost, as well.

Re:Great

[jackb_guppy]

Well it is BillGatesbot, so the cost is $50 Billion.

Why you ask, that was what it costed in 1990 and MS does not sell any for less than they sold it before.

Re:Great

[Alamoth]

The Sony AIBO (before it was discontinued) cost $2,000. However, schools are allowed to buy them for educational purposes at a discounted price of ~$1,700. This is a robust platform, and not everyone needs one. As far as personal robots goe, the B.O.E. Educational robots go for $100 ~ $500 depending on quality. The lower end is a little more pricey than your average new textbook.

Re:Great

[Rethcir]

Insightful! I have more than one CS book that I paid $100+ for and opened maybe twice, due to the fact that my professor was pretty much speaking off the cuff the whole time. Oh well, they look pretty good on my shelf.

Re:Great

[hal2814]

Most of my CS professors just assigned books that would be helpful to have on your professional bookshelf and taught us what they pleased. I have plenty of $80+ books (never a $100 though) that I've never needed for class but have helped out a lot since then. The most important IMHO being my Algorithms textbook.

Re:Great

[The Snowman]

Most of my computer science professors understood the racket that is the textbook industry, or had written so many of their own textbooks they just handed out notes (apparently it's unethical to teach from your own textbook, but not from the notes you used to write the textbook). So they had us buy cheap books that serve as good reference material, if any at all. And actually we did use them quite a bit anyway.

Maybe the market will find a good robot or two that are common across universities, that or an open standard that they all use. Imagine purchasing a robot your freshman year that is useful your entire college career. Studying data structures? Well you can use those to queue instructions to the robot. Studying operating systems? You could write a simple OS to control your robot. Pretty much any computer science topic has some application to robotics, and with a robot, you get visual confirmation that your program works. I think this could be awesome. Could be. Whether this idea will succeed or even get off the ground at all remains for us to see.

Re:Great

[SatanicPuppy]

Just sounds like a gimmic to me.

I flirted with Comp Sci a long time before I actually got physical, and took a lot of classes at a lot of different places. I had classes that were too heavy on theory, I had classes that were too heavy on "practical" skills, which usually amounted to "how to use this language/program to do this thing".

I think, in the long run, a lot of places really don't have the faintest idea what it takes to make a good CS person. It doesn't help that CS covers way too much ground anyway. I'm unusual in that I don't really have a specialization within the field. I'm a programmer/administrator with java/c#/perl/php/python experience and unix/linux/windows/mac experience...Not claiming to be godlike at any of those things, but I'm competent and flexible. The thing is, as fricking crazy as I am, I still haven't used all the classes I had to take for my degree.

I can't help but think we need better "gut" courses, and more flexibility with higher level classes, so that you get a more elegant toolkit to take with you into the workforce. If you give people a better range of courses that appeal more to their interests, you don't need a fricking robot to make them excited about their subject. It's a hell of a lot better than pushing students to design another "airline reservation system" (I had to do this in Scheme and Java), or "Pokemon" (Java and C), or "IM client" (Java, C#).

Blah blah. I'm rambling. Summary: Robot == lame...Cooler courses > Robot.

Re:Great

[Anonymous Coward]

Unfortunately I haven't made an account so don't dismiss me for being an AC. I have actually been working with a small college called Canisius in Buffalo and we have already introduced robotics into our lower level classes. Believe it or not, it does help. Students were going nuts with it. I even think we gave Doug the idea for this. If you check out AAAI last year, there was a paper from canisius college about robotics in CS classes. Of course you don't use the Aibo (which I love programming on) for lower level classes. It is far too advanced and delicate. It costs way over 2K after you get all the bells and whistles so having a CSC-1 kid play with it is stupid. We used Lego mindstorms for the lowest levels. They are robust and easily programmable in java (which is our primary language). We also have some handy boards flying around, but that is more for the adventurous older students. In addition, Doug is actually waiting for the work I am doing through Canisius with their Pyro program,the Aibo, and OpenCV. It kinda sucks that we weren't mentioned at all, given that we've worked so closely with Doug. Hell, I don't know why I'm posting I doubt anyone will believe me anyway.

Re:Great

[Anonymous Coward]

Allan Turing = Computer Scientist. programmer/administrator with java/c#/perl/php/python experience and unix/linux/windows/mac experience = code monkey

Re:Great

[SatanicPuppy]

Ah, CS snobbery from an AC...I mod ACs -4, but I still check 'em when they're under my threshold, which is stupid. Anyway, Turing was deeply interested in practical computing you know, and in his day, there really wasn't enough computer science that it was in any way difficult for one person to study it all. Thank you for making my point though, because, for what I do, there was no other choice for a degree program.

I know the CS purists sneer at the guys in the trenches, like we should all have gone to ITT or something, but I get a hell of a lot of use out of my programming, dbms, and network architecture classes. Every time I have to do memory management, packet shaping, or funky data conversion between networks, I rely on theory I never could have gotten from some piddly tech college. But there was a lot of crap in there I didn't need: a whole class on binary number theory, a class on machine code, another on circuit design. I'm never going to work with computers on that level, and I'm hardly unique among CS majors in that.

At the same time, majoring in programming was too fluffy and useless for me. There wasn't even a networking degree, because most four year universities view that as akin to plumbing. You have to actually get out in the field to get anything other than vague theory, and that stuff is seriously interesting and difficult once you move past the "plug this machine into dat machine" crap.

I guess I kind of think of it like medicine...Nobody majors in "Doctor" but we're all expected to major in "Computer Scientist" even though that describes us about as well as "Doctor". I'm not looking for more practical classes, because practical becomes obsolete too quickly, I'm looking for theory that is focused and useful to whatever specialty you're going into within the set of specialties that is comp sci.

Re:Great

[StarvingSE]

But there was a lot of crap in there I didn't need: a whole class on binary number theory, a class on machine code, another on circuit design. I'm never going to work with computers on that level, and I'm hardly unique among CS majors in that.

Even if you don't use those classes, its useful to know on a basic level certain topics. Agreed, you are never going to program in machine code, but I know I like knowing what my C++ will compile to, and it also helps in making you a more efficient high-level programmer. And basic circuit design is just logic, which any programmer should have a solid grasp on. Correct me if I'm wrong, but I'm pretty sure that doctors all take the same classes for the first 4 or 5 years, and then they specialize. If you want to specialize, get a MS or PhD in computer science and then you can focus at becoming an expert in a particular topic.

Re:Great

[The Snowman]

programmer/administrator with java/c#/perl/php/python experience and unix/linux/windows/mac experience = code monkey

Not really. A good programmer is also a computer scientist in the academic sense, since he must understand all the theory. I have the unfortunate task of cleaning up a code monkey's program. If he had understood computer science he would still be a programmer with Java and Windows experience, but would have wrote better code.

I consider myself a computer scientist and I can check off most of the list you provided. Sure, the degree is part of it, but I actually care about the programs I write. I think before I code, and make sure it works correctly.

Re:Great

[SatanicPuppy]

Shrug. I got to Comp Sci by way of Cognitive Science, which means I had all the logic I could handle long before I had to take the fricking circuit design course...CS required two logic classes, which, compared to the five I'd already taken was like having to go back to Algebra I after finishing Calc IV. I'd be fine with having to take a course on binary numbers, machine code, and circuit design, but having to take one of each is absurd for a non-engineer. And you don't have to take all the same courses to get into a Medical program, not even close...In fact different medical schools require different courses, and some medical schools look favorably on things like Psychology (which blows my mind, though my doctor friends assure me its true).

Anyway, Comp Sci is a practical degree. You can have a perfectly fulfilling career without going past your BS. All I'm saying is, use the 100 and 200 level courses to give you a nice grounding, and then loosen up the requirments on everything else, so individuals who are interesting in programming can go one way, people who are interesting network protocols can go another way, and people who are interested in hardware design can go their own way as well. I still posess the ability to do binary long division with a pencil and paper, and it irritates the hell out of me.


        His ignorance was as remarkable as his knowledge. Of contemporary literature, philosophy and politics he appeared to know next to nothing. Upon my quoting Thomas Carlyle, he inquired in the naivest way who he might be and what he had done. My surprise reached a climax, however, when I found incidentally that he was ignorant of the Copernican Theory and of the composition of the Solar System. That any civilized human being in this nineteenth century should not be aware that the earth travelled round the sun appeared to me to be such an extraordinary fact that I could hardly realize it.

        "You appear to be astonished," he said, smiling at my expression of surprise. "Now that I do know it I shall do my best to forget it."

        "To forget it!"

        "You see," he explained, "I consider that a man's brain originally is like a little empty attic, and you have to stock it with such furniture as you choose. A fool takes in all the lumber of every sort that he comes across, so that the knowledge which might be useful to him gets crowded out, or at best is jumbled up with a lot of other things, so that he has a difficulty in laying his hands upon it. Now the skillful workman is very careful indeed as to what he takes into his brain-attic. He will have nothing but the tools which may help him in doing his work, but of these he has a large assortment, and all in the most perfect order. It is a mistake to think that that little room has elastic walls and can distend to any extent. Depend upon it there comes a time when for every addition of knowledge you forget something that you knew before. It is of the highest importance, therefore, not to have useless facts elbowing out the useful ones."

        "But the Solar System!" I protested.

        "What the deuce is it to me?" he interrupted impatiently: "you say that we go round the sun. If we went round the moon it would not make a pennyworth of difference to me or to my work."

--Sir Arthur Conan Doyle, A Study in Scarlet

Re:Great

[andrewman327]

If it uses Java, then I'm in. I don't want to start a flame war, but Java is a decent first or second language to learn. Personally, I learned BASIC, C, C++, then Java.

Re:Great

[DJStealth]

In my 3rd year robotics course we built robots from scratch. Avg price was $50-100. Basically did the same things TFA was talking about. Following lines, etc. However, most iterations were analog; the final project went digital.

Re:Great

[a55clown]

The hardest part about programming on a mobile robot platform is what happens to the bot after so many iterations of compiled code being executed on it. My robotics class in college wasn't so much sending commands to the damn thing as much as it was determining whether sensors were acting correctly (our bots were ice-cream tub-sized machines with two independent drive motors on the sides and a caster wheel for balance).

On a computer, 99.9% of the time, code gets executed the same way every time. In the actual world, you don't have a controlled environment, and tiny variations in the environment can butterfly-effect a scripted series of commands sending the bot head on into a wall, further damaging it.

On the other hand, it put our programming (and thinking) paradigm more along the lines of Google's early research -- design the system around the idea that sub-systems could (and will) fail without warning. Not a bad learning subject for advanced programmers, but it's not something I'd teach to an introductory course.

Re:Great

[fozzy1015]

Two colleges are hoping to make computer science courses more attractive by including personal robots with the textbooks. Looking to boost enrollment in introductory computer science classes...

How about figuring out a way to get more girls to enroll into computer science classes?

Although that might be harder to figure out then robots.

Why?

[1u3hr]

Robots Coming to Intro Computer Science Classes

As teachers or students?

Re:Why?

[Money for Nothin']

Robots Coming to Intro Computer Science Classes

As teachers or students?

You mean there's a difference between robots and the people in the CS dept?

Re:Why?

[Amphiaurus]

Neither, just Microsoft employees.

Da Cheatbot

[Average_Joe_Sixpack]

If only I would've had this lesson before deciding on a career in technology

repair

[zotkop]

I wonder who's going to repair those things every thime they crash (then seem to run Micro$oft)

Nothing all that new

[technoextreme]

At Northeastern University I took a course similar to the one in the article except it was related to a program called CenSSIS. It was pretty interesting because it combined ultrasonic technology and programming to work on different projects. The most impressive of which was mapping an object found in jello without cutting into the jello. Though that course was an engineering course and not a computer science course.

Re:Nothing all that new

[DaBigEnchilada]

At Oregon State we used robots (tekbots) in many EE and CS classes. In my experience, it was terrible as a freshman, as those of us without prior knowledge in electronics were completely clueless as to how to debug them. The first version we built was very complex, yet was required for the course. The longer your robot didn't work properly, the more your grade would inevitably suffer. The TAs were thrilled with having to debug everyone's robots all the time, too (of course, some 90% of this was bad soldering). Later classes in college, when we started diving into Assembly, we got an updated robot kit. This updated kit included addons to the original robot. If you didn't have an original robot, or it didn't work, welcome back to square one where you had to rebuild it. Oh, and robots are cheap in neither price nor labor. At least at this point in college, we were using the robots to execute assembly; there was an immediately practical use for them. I think robots can be useful as an aid, but really not much moreso than a simulator. They have the coolness factor, but that's about it...at least for basic undergrad education.

A bit deceptive, isn't it?

[tygerstripes]

...to make computer science more hands-on and practical, rather than simply about debugging programs.

Kinda like discussing Bronte during Maths to make it "less about numbers", isn't it?

Re:A bit deceptive, isn't it?

[ThinkingInBinary]

No, it's kinda like discussing economics or physics during calculus to make it more practical and show people the real-world applications. Robots are an application of computer science; Bronte is not an application of math, but physics and economics are.

It's a matter of giving people more practical work, which is both more interesting and easier to learn for some people. I usually find that I learn a language better when I can play with it, and doubly so if I can write something real with it. Having a real piece of hardware that responds to your program is more exciting than just printing messages on a console.

Re:A bit deceptive, isn't it?

[welshwaterloo]

Kinda like discussing Bronte during Maths to make it "less about numbers", isn't it?

...or - more like discussing the aesthetic properties of the Golden Ratio during Maths to make it "less about numbers"

Re:A bit deceptive, isn't it?

[alas_anon]

>> No, it's kinda like discussing economics or physics during calculus to make it more
>>practical and show people the real-world applications.

One problem with this is that a person with a CS undergrad degree isn't qualified to go into automation programming, so it becomes a course which does not lead to employment. It sure would be an exciting course to most, but not a very useful one.

When a course like this is put into the curriculum, some other course must be taken out to make room for it. You can't just tack more requirements onto the degree (or throw out the wonderful liberal arts stuff). Throwing this course in the degree might have the effect of diluting the quality of the CS curriculum, depending on if there is room or not for it. Something's gotta give...

Re:A bit deceptive, isn't it?

[Anonymous Coward]

Jane Eyre's missionary cousin John has a 1% chance of contracting malaria during each day that he spends in India. Once he has contracted malaria, his survival time is Poisson distributed, with a mean time to failure of 200 days. What is the probably that John survives for at least one year?

Re:A bit deceptive, isn't it?

[AcidLacedPenguiN]

or you could change the first year assembly course from your typical Motorolla HC* microprossor to the roboticsillything microprocessor. That way you actually get taught the assembly stuff and you get the cool robot that will eventually be relegated to holding your beer on last class. . . I don't know about other colleges and universities, but at mine the first year sees over half of their enrolled dissappear after christmas, most of them either hate the first year content, or just plain suck. I could imagine a program like that to be pretty popular around here though. The only problem I could foresee is maybe some of the background theory might be neglected, If memory serves me correctly it's that background theory that saved my ass when it came down to midterms and exams, I wasn't about to cram opcodes into my brain afterall.

Re:A bit deceptive, isn't it?

[Ailicec]

Its quite a shift for Georgia Tech's CS intro; 10 years ago when I was there the Intro to CS class was entirely taught in a pseudo language somewhere between Pascal and Java, the rationale being that denying the feedback of a compiler would force students to understand their logic better. You'd implement programs, algorithms, classes, etc in multi-hundred line pseudocode programs.

While the CS majors would usually figure it out well enough to get through, the non-CS people who had to take this class were usually quite lost. They went through the entire class having never seen the product of this mystery code they were writing, often entirely missing the point of what a computer program is and not gaining the broad overview of what computer science is about because everything they did was obscured by a contrived language.

I can see some logic for the psuedo language approach, but I think the biggest reason they went with it is because it made the class hard and produced the desired high dropout rate.

Re:A bit deceptive, isn't it?

[CastrTroy]

If you're programming the robot in assembly then you aren't going to get very far in getting it to do much. Unless your assembly instructions includes MOV A,50, Which moves the arm 50 degrees, then you're going to be pretty limited in what you can get the robot to do. Also, if that's the kind of assembly programming you're doing, then you're not really learning assembly. Despite the fact that we all hated the assembly course, because we had already learned C and Java, and in assembly it takes an hour to write a function that iterates an array and does some simple operation on the numbers (at least when you first start), I still think that assembly is a very important skill. They tell you how complex certain things are in algorithms class, but you don't really get a feeling for how complex certain things are until you program them in assembly. Also, I took a robotics course in 4th year. It was a 4th year course. It was hard and there was lots of math. That's the way robotics is. Unless you're programming the robot to do a specific task (such as follow a specific path, or move from a home position and pick up a ball in a predetermined position), which is extemely boring and simplistic, then robotics gets very complicated, and is not something for first year introductory students.

Re:A bit deceptive, isn't it?

[1iar_parad0x]

So students get to buy an expensive and fairly worthless robot so they can get a more 'hands-on' feel for a course. I'm taking a shot in the dark most students taking an 'intro to CS' course aren't ready to deal with the mathematics of Robotics/AI or are ready to implement complex path-finding algorithms. If anything some students (admittedly the not so good ones) are still struggling with syntax and coding style. Sure, we should incorporate more AI and robotics education on the undergraduate level. Let's try to avoid making it into a 'hook'. Secondly, any physics or economics taught in an introductory calculus course is pretty superficial at best. They might use it to illustrate mathematical concepts and create a few word problems, but students are even really ready to tackle any real problems in mechanics. In other words, it seems like a lot more trouble than it's worth.

Re:A bit deceptive, isn't it?

[Anonymous Coward]

So students get to buy an expensive and fairly worthless robot so they can get a more 'hands-on' feel for a course.

Or they could use one provided to them in the lab, which would be pretty cost-effective for the college if the things laster more than 2-3 years (most college-class-friendly robots I've seen definitely do). As long as your code is abstract enough to run on more than one robot type and there's a decent simulator (the Python software linked from Blank's addendum to the article meets these criteria), students can also work on getting the simulated bot to do stuff and then run their programs on the real one. You want to do this anyway since being able to put the robot in an arbitrary world helps you quickly test cases that you can't in real life.

I'm taking a shot in the dark most students taking an 'intro to CS' course aren't ready to deal with the mathematics of Robotics/AI or are ready to implement complex path-finding algorithms.

Teach them more "New AI" type stuff instead (no, not neural nets). You could certainly do something cool with, say, a simple subsumption architecture, wherein you define a heirarchy of simple behaviors and the robot changes which one (/ones) it's doing depending on the input. It also might help students learn about coding something other than single-threated imperative/object-oriented programs in Java or C (which sometimes screws them later when they have to write real threaded or event-driven code).

Braitenberg vehicles are another good example of something simple but could you can do with robots: show students how wiring together a ridiculously simple robot brain with little more than inputs connected directly to outputs can create a robot that will do fairly complex tasks.

Robots *can* be made to work in intro classes without feeding the students B.S. or forcing them all to buy robots.

Re:A bit deceptive, isn't it?

[Anonymous Coward]

why? why are robots an application of computer science? computer science engineering maybe, but computer science no.
robots are more about electronics and mechnanics then they are computer science.

You don't even need a computer, per say, to get robots working. You could use PLCs and ladder logic.
computers might not be a part of it at all.

Robotics is just one more thing that Microsoft won't do well but due to their hugeness they will get a market share.

Hardly Revolutionary

[Anonymous Coward]

That's nice, too bad the Peter Kiewit Institute has been doing this for a few years now.

It's a Trap!

[ObsessiveMathsFreak]

It case anyone hadn't noticed, computer science has very little to do with computers, and nothing whatsoever to do with hardware. I can just imagine the course instructors cackling as the naive students skip inside expecting arrays of sophisticated robots waiting to be programmed:

"Fools!! Did you really think it would be that interesting? You're mathematicians now!! Now get back to computing runtime complexities for applications you will never have call to write, or understand! *Wwwuu-ttisshh* Bwahahahhahahaaa !!"

Re:It's a Trap!

[sesshomaru]

This is why computer science professors get all the chicks, "Then I made them prove why one travelling salesman problem was not NP complete, Baby..."

I have the perfect robot for a computer science class. It doesn't do anything, it's just a plastic toy. But it is called "The Big O."

Re:It's a Trap!

[jimbolauski]

I'm having flashbacks to all the naps and crosswords I did while my professor discussed P v NP. Even he said the class was boring and had to teach it because he was the new guy.

Re:It's a Trap!

[StarvingSE]

It all depends on the professor. Admittedly, comp theory can be very boring and dry, but the prof I had for it made it one of my favorite classes.

Re:It's a Trap!

[drawfour]

I never knew how to write the sound that a whip makes. Now I know. Thanks!

Argh

[RandUser]

I really wish I weren't two years through the CS program here at Georgia Tech already, because the hands on robotics stuff tfa talks about sounds really slick and it would definitely interest me even now. There really isn't an "intro robotics" course here, mainly an AI intro course and then a hodge podge of specialized areas. I did skip my intro CS course but if I had had this kind of stuff as a choice, I would have taken it. Oh well, that is life.

I hope this program does well and encourages students to get into the robotics/AI field, an area I think hasn't quite broken out into the consumer products market yet and has a ton of potential for everyday applications. Or maybe I'm just dreaming, who knows.

Re:Argh

[alas_anon]

>>There really isn't an "intro robotics" course here,

Yes there is, it is over in the college of EE. They have a very nice robotics program at Ga Tech. Get a dual degree (CS and EE) and you will be unstoppable.

Really, I don't think robotics belongs in the CS undergrad curriculum. The graduates of that type of program are going into desktop and server programming, not automation. Automation programming requires a lot of hardware knowledge (microprocessors, bit banging, actuators, sensors, signal conditioning, etc.) and that is too much info to pack into one course to make the person capable of getting a job in automation programming. Maybe at the graduate level it would be a good specialty (with cross-cooperation with the EE school).

That said, the reason they might want to have a robotics course is to excite the students into pursuing CS. When students first come into their programming classes they are all cranked up about it and the intense study of syntax, OO, and algorithms knocks the excitement out of most of them. As salesmen say, it would provide a "Wow Effect" for the student to continue studying the other "dull" aspects of programming.

Another opinion is that if the student considers the standard fare of programming (syntax, OO, and algorithms) to be dull, maybe they are in the wrong field anyway.

It is unlikely that a CS student with one course in "Robotics" will get employment in the automation programming field. It just ain't enough.

Re:Argh

[hal2814]

Shouldn't CS be exciting enough as it is if you're planning on persuing it as a degree? I went through the CS program at UGA back in the late 90's when everbody and their brother was looking to get a CS degree because of the dot com boom. There were a lot of people who just didn't belong in the program. They were folks who just weren't excited about computer science. They would moan and complain when they had to do assignments that required any real though and they'd never try out anything being taught in class that wasn't specifically assigned. They just shouldn't have been there. They would've been happier in some other major. And these were folks who were swayed by the prospect of a high paying career. Imagine the quality of folks you'd sway by offering a shiny robot.

Re:Argh

[MajinBlayze]

Contrarily, I think that's kind of the point. Most people view robots as complex, or outside of their ability to understand. This would more than likely keep out the "I can type on MS Word, so i bet i can be a programmer" kind of people (which I've seen a few of try to go through CS classes).

On the other hand, I'd like to see more "tinkerers" go through CS. These are the people who I think would be swayed by the robot.

Re:Argh

[Anonymous Coward]

http://borg.cc.gatech.edu/ipr/

held every Spring, and you don't need to take intro to AI.

Re:Argh

[Anonymous Coward]

>> And these were folks who were swayed by the prospect of a high paying career.

  >> Imagine the quality of folks you'd sway by offering a shiny robot.

That's the idea, boost the enrollment with something exciting. You are right
that this will attract a lot of people who will drop out later when they hit
the real curriculum.

I'm teaching a course right now that was called "Robotics". It was actually a classic
automation course and it involved PLC programming. It had nothing to do with robots,
but the administration said "Robotics" sounded more exciting.

now the course is called "PLC Programming", which is also off-the-mark because
the course is automation, involving all the concepts of hardware AND software, not
just programming. Again, administration decided "PLC Programming" sounded more
exciting than "Automation".

I also teach beginning Java and C++. Many students loose interest when they find
out there is so much to learn. They want to program games, mostly. Many of them
believe they are going to click buttons to magically make games and then they
loose interest in the courses.

I try to give them some fun programs to do, but the beginners are often
disappointed with the large amount of work and knowledge it takes to make simple
things. They are only focused on the result and are not interested in the process
of programming.

A lot of money and a shiny robot might keep those people in the program, but
it's hell to try to teach those students who don't have an intrinsic interest in
the art and science of programming.

CS isn't all about debugging programs

[Ritchie70]

I love the comment about "all about debugging programs" in the summary and article. If that's what these colleges are teaching as CS - programming and debugging - then CS has changed a lot in the last 16 years since I graduated. I had classes in algorithms, operating systems, various sorts of math I don't even remember, AI, networking theory, and more.

And freakin' robots? Kids were playing with robots in grade school 20 years go (Logo anyone?) - this sounds way too similar to me to belong in college level classes.

Real CS is hard. WTF would you want to increase enrollment in intro classes unless the whole program is a joke? There aren't going to be (I hope) any fun robots in the next class when the people who can't hack it hopefully get weeded out. Why not just weed them out up front?

Re:CS isn't all about debugging programs

[engagebot]

Yep. I'm really the last of the "real" CS students from LSU. Midway through my time in college, they started changing the classes over. Its more software development than anything else. Except starting out with .NET is not great in my opinion

They got rid of all the architecture classes, especially the good one where you learn about *how* memory works, threading, processor scheduling, all that stuff. They also got rid of the OS class. I mean, they still have an OS class, but its now a touchy-feely class where you don't actually *learn* anything. I feel bad for the kids who are going through right behind me...

We used to have a mandatory class on assembly too. Granted, its somewhat useless as a programming language in real life, but it still helps teach alot about what's going on at the low, low level.

Re:CS isn't all about debugging programs

[tecnopa]

Having just recently graduated from the CS program at GT I would assume that this robot enrollment thing is aimed more at non-CS majors... I know they're currently re-doing the curriculum so that there are different "threads" one can follow through the curriculum... who knows... I will say however, that the intro CS courses at GT do not assume that you've ever programmed before. When I took the course it was basically an intro to algorithms done in "pseudocode". I enjoyed it that way and thought it helped me focus on the concepts rather than having to deal with compilers, learning syntax, etc. which at that time was foreign to me. Basically it was a good, soft intro into Computer Science that was more focused on theory than specific languages/implementations (which we had to figure out for ourselves). Then they switched to Scheme, created new Intro CS courses just for non-engineering students, etc. so who knows where its at these days...

Re:CS isn't all about debugging programs

[icebrain]

What really sucked is that they used to require everyone to take Scheme (godawful useless language for most of the engineers). Then, the semester after I took the Scheme class (got put in the "advanced" section somehow... didn't really belong there since it was over my head and I'm an aero engineer, but was worth it anyways), they start up their new "Computing for Engineers" course (using Matlab). And now all of my professors assume we know matlab very well because that course is offered... but we never took it!

It is good that they're branching out with the program though. They went from the standard "_everyone_ must take Scheme!" to having:

Intro CS, for CS and (I assume) CompE, EE, and the like
CoE (the Matlab course) for other engineering majors
CS 1315, an intro course for the liberal-arts majors which uses Java, Python, and others, and does stuff like image manipulation, sound generation, etc.

Also, they have the "threaded" CS now, and a new program called "Computational Media." From what I understand from my friends in the program, it combines CS with visual/arts stuff and the literature/culture group, and is intended for those who want to focus on, say, video-game development, graphical media, animation, and the like. They also have a good relationship with Cartoon Network, which sits just north of campus on 10th street.

Re:CS isn't all about debugging programs

[coreb]

LSU is teaching .NET now? I've gone through ULM and it is Java-Heavy. But I've noticed a water-down of the curriculum there, too. They did some changing around with their courses and caught a class of students without a Senior Level Architecuture class and combining assembly language and computer organization along with teaching computer organization to that group twice. It's sad that institutions are watering down the curriculum for an increase in students that should really be somewhere else.

Re:CS isn't all about debugging programs

[Anonymous Coward]

various sorts of math I don't even remember

Taking two years of Calculus for nothing sure is fun.

Re:CS isn't all about debugging programs

[hevenor]

You're right...there's too much emphasis on making things practical. It _is_ university after all. If you want something practical then go to college. If you're in a pure math program are you going to complain that it's not practical? In my first year of CS we already had robots. It was named Karel and it taught us all about objects...it didn't have to fetch me my morning newspaper.

Re:CS isn't all about debugging programs

[engagebot]

As far as the programming languages go, in CSC, they now start the first two basic classes with regular old C (gcc and pico) and then they go object oriented after that. They've dropped java altogether from what I gather. We had an awesome java instructer (props to Dr Gwee), but he left and I think they just didn't get anyone to replace him.

Until recently, it used to be several classes of C, and then the required electives you choose from were all the different object oriented languages. That way you had a really strong base in C and Assembly before you started OO.

Now in ISDS (the business college 'IT' curriculum, basically the CSC dropouts), they start with VB.NET, then have one class on basic (read: basic) html, then they're done with technical stuff. From then on out, its business classes about how technology empowers business...

Re:CS isn't all about debugging programs

[jamiethehutt]

We used to have a mandatory class on assembly too. Granted, its somewhat useless as a programming language in real life, but it still helps teach alot about what's going on at the low, low level.

I found the same thing with my computing course at college, but found that they hadn't stopped the Assembly module they'd only switched it to one of their Electronics courses. It seems if you want technical low level computing stuff you need to do electronics these days. (Or go to university, which I'm doing now and getting all the low level 'fun').

Re:CS isn't all about debugging programs

[StarvingSE]

If you care to read the update, the media misquoted the professor on that debugging statement.

And as to CS degrees, we still take algorithms, theory, and OS as main core requirements, but we also have options to take elective high level courses such as AI, networking, etc.

Re:CS isn't all about debugging programs

[Anonymous Coward]

And now all of my professors assume we know matlab very well because that course is offered... but we never took it!

Then learn it on your own. Probably most of the engineering classes at my school assumed you knew Matlab, the stats assume you know Matlab and R, oen of the intro cs classes gives you a week to learn javae mostly on your own, and the advanced cs classes din't care what language you knew as long as a TA knew the language (one I took prefered python and another c/c++). I don't think there are any real classes for Matlab, R or python (not counting the 2 hour TA intro most classes have) however that simply means they expect us to be smart enough to learn them on our own. Heck in my current internship I had to learn (as in I could use them in practice) perl and bash in 2 days (barely knew either), and in 2 weeks I became decent with them. If you can't learn on your feet and organize your time to do so than I feel bad for you given what I've seen in the economy (you adapt or get fired when you're 50 and replaced with a guy half yoru age because they cost less, of course no one else will hire you anymore).

Re:CS isn't all about debugging programs

[chengmi]

I feel bad for the kids who are going through right behind me...

Don't feel sorry for them, feel sorry for yourself. Employers will start to see these kids without any low-level knowledge coming out of LSU, and think the same of you. Good luck.

We built and programmed football-playing robots

[Bromskloss]

...where I study in Stockholm, Sweden. Loads of fun!

Re:We built and programmed football-playing robots

[Anonymous Coward]

KTH (http://www.nada.kth.se/kurser/kth/2D1426/robots05 .html)? So it's soccer now? When I took the course a few years ago it was ice-hockey :-)

However, this course is not an intro to CS, but rather an intro to robotics, which you can choose in your final year.

Re:We built and programmed football-playing robots

[fbjon]

They do that at a lot of unis, actually, if it's RoboCup you're talking about?

Re:We built and programmed football-playing robots

[Bromskloss]

They do that at a lot of unis, actually, if it's RoboCup you're talking about?

No, this wasn't RoboCup. Our competition involved only those who took the course. Pretty much the same sort of game, though. I suppose other universities too are holding their own competitions, for that matter.

Practical CS.

[Anonymous Coward]

"The idea behind the program, Blank said, is to make computer science more hands-on and practical, rather than simply about debugging programs."

Welcome to DeVry.

Re:Practical CS.

[Anonymous Coward]

Hey man dont diss DeVry, they are on the verge of a breakthrough answering the longstanding P vs NP problem....

Unless ITT-Tech beats them to it.

Sounds like chaos to me

[eighty4]

Pintsize, anyone? Imagine 50 of them running around a computer lab...

Re:Sounds like chaos to me

[ZSpade]

Questionable content is too indy to talk about here.

Coming?

[Chris Pimlott]

This is hardly cutting edge;
Case Western Reserver University started a program like this 5 years ago using Lego Mindstorms kits, and I'm sure they weren't the first. This is seperate from the higher-level Autonomous Robotics (aka Lego Lab) course that's been going on since 1995 and is based largely on MIT's 6.270 Autonomous Robot course that created the Handy Board.

Re:Coming?

[Duncan3]

We had the little turtle robot back when I learned LOGO.

In 1981...

Coming soon to people outside Minnesota? Guess some places are just ahead of the rest of the world ;)
-

Keep it low level.

[oyenstikker]

With only a CS1 and CS2 under my belt, and having programmed in only Java and VBA, I did a project in school that had me programming a self-navigating robot in C. We had a small processor* with a C compiler and a debugger. I soon augmented the debugger with a row of LEDs wired to one of the registers, so I could debug while the thing was driving around and not hooked up to the computer with the debugger. It should just be kept low level enough that students have to solve their own problems. Its a great way to learn.

* TI MSP 430

Re:Keep it low level.

[murraj2]

We have a course almost identical to this at RPI. Intro to Embedded Control, where we use an Intel 8051 microprocessor to program a robot to follow a white strip of tape around a course using IR LEDs and sensors. Hell, I remember doing stuff like this 10 years ago with a set of legos at an after school program at a nearby tech HS.

My intro to CS (I'm actually EE/CSE dual not CS) classes focused on the basic of programming with C++ including how the compiler worked etc, and then got into more complicated algorithms. The only debugging that was taught was the fact that you had to debug your programs when you did them wrong.

Then it's not computer science

[s1axter]

See the problem is, robotics is not computer science... it's electrical and computer engineering. Just because you want to bring more people to the dicipline doesn't meen you redefine what the dicipline is

Re:Then it's not computer science

[denim]

Programming is involved in computer science, at least on the lower levels. I don't personally believe that incorporating robotics is sensible with freshmen though.

Here's the important bit. Whatever you get the kids hooked on up front is what they'll stick with for a long time. Microsoft is doing what many other companies have done before. I'm not sure why they're bothering with robotics now, honestly. When I started in college (1982), some people were thought to be thinking "video game programming!" but it didn't seem to be the case. I wonder what's changed.

Personally, I'd like to see certain other things become involved in CS degrees before robotics. Say, some clues relating to security, and maintainability of code. My school required a course in tech writing, which would be a Good Thing for all, I say.

Re:Then it's not computer science

[Alamoth]

Anyone who thinks that Robotics is not Computer Science has never actually worked with a robot. Robots encompass Mechanical Engineering, Computer Engineering, and Computer Science.

MechE provides the muscles, the bones, the skin, and the structure of the robot.

CompE provides the nervous system, the veins and arteries, the heart, and the hormones.

CompSci provides the brain.

Take any one of these disciplines away and the robot fails.

Re:Then it's not computer science

[RShizzle]

Brilliant. Good quick synopsis. Much of what being taught today in college doesn't belong in only one of the the subfields, whether it be EE, CS, or ME. Instead, a lot of the focus is trying to get students to work across fields, and play with systems integration. This is one of the exciting parts of the RoboCup competition, especially in the small sized league.

Future work, for both money and glory, will likely require you to be much more than a code monkey, and instead require knowledge of how to create components of larger systems, and then integrate them seamlessly into a larger solution. By providing a way to see how a practical end is possible from vague beginnings (which CS most definately is for the uninitiated), this is definately a cool project.

Re:Then it's not computer science

[dim5]

No, you misunderstood... they're actually robotic travelling salesmen. You program them with a turing machine.

Re:Then it's not computer science

[organgtool]

So is it the electrical engineers or computer engineers who program the complex, multi-threaded artificial intelligence applications? I mentored a high school robotics program that participates in the FIRST Robotics competition and I can tell you that the hardware students didn't know or care how to program the robot. They focused on building the robot and the software students focused on programming the behavior of the robot. I'm not saying that there aren't people interested in the hardware and software aspects of robotics but robots have become complex enough where there are few people who know enough to build the robot and create the software to drive the autonomous behavior of the robot.

Re:Then it's not computer science

[Langley]

They are doing it, simply, because of this: http://msdn.microsoft.com/robotics/

Re:Then it's not computer science

[denim]

Okay. That expands on it a bit. Getting it into the hands of frosh expands on what I was saying. It's nothing so much about CS as about money.

Microsoft doesn't want it to be about...

[Anonymous Coward]

just debugging programs? Maybe they should talk to the colleges about developing courses specifically for debugging software and put their people through it.

The idea...

[pesc]

The idea behind the program, Blank said, is to make computer science more hands-on and practical, rather than simply about debugging programs.

Or maybe the idea is to make sure that the students have to use windows in order to use the robots. MS wants its OS to be used more for embedded and controller applications and have to do something to stop the students from using those small, open, inexpensive Linux systems.

Or am I wrong? Could the students use the robots and textbooks without MS tech?

Re:The idea...

[ch-chuck]

to make sure that the students have to use windows in order to use the robot

or even worse, it's the foundation of Microsoft's future robot clone army built to enslave the human race. Bill Gates isn't getting any younger you known.

HouseBot: "Sir, it looks like you're trying to polish your shoes, may I be of assistance?"

Joe: "Sure. Wow Mable, I don't know how we got along before we got these household robots."

HouseBot: "Robot to HQ: I have gained control of his shoes. Proceeding with plan."

BalmerBot: "Excellent. Excellent!".

Throwing chairs

[cerberusss]

Microsoft Corp. is working with Bryn Mawr College and Georgia Tech on developing new ways to bring robotics technology into the classroom. Douglas Blank, a computer science professor at Bryn Mawr, said the goal will be to throw chairs with superhuman accuracy.

Didn't Seymour Papert do something like this?

[CurtMonash]

Wasn't LOGO some kind of primitive version of this?

Cornell's Doing it Too

[sous_rature]

As the subject of this reply suggests, Cornell is also getting in the robot game, although I don't think they're collaborating with MSFT in the effort. In fact, I'm signed up for the course right now. The idea had the full endorsement of the campus's top computer science pedagogue, and here's how my advisor explained it to me (I'm a math major): The point of an intro computer science class is to teach you how to write clean programs, independent of what language you're working in. Languages are relatively easy to learn. How to not write "spaghetti code" is not. It doesn't really matter what you're programming, be it java or aibo (Sony's robot dogs), so long as it has a computer's logical structure and you learn how to use it effectively. Also, many people underestimate just how much of a challenge many robotics programming tasks can be, and how relevant they are to emerging computer scientists. I've seen talks on uses of de Bruijn (forgive spelling) sequences for position recognition, and lectures on genetic algorithms for getting robots to perform complicated task. Someone tell me that these ideas are too simple or irrelevant, and I'll show you hordes of computer science professors who disagree.

Since when is computer science about programming?

[Anonymous Coward]

If I recall correctly, computer science degrees are more about algorithms and learning how to think like a coder, and not really focused on programming. Won't this reinforce common undergraduate misconceptions?

Perhaps it's high time that we had an official concentration split - something like choosing a focus between theoretical computer science and applied computer science. Hell, for all I know such choices at a single university already exist, but I haven't seen them....

Manoi

[jlebrech]

http://www.maxconsole.net/?mode=news&newsid=7135

Who will this attract?

[Anonymous Coward]

It's all well and good to try and attract more students to computer science classes, but are people who are wow'd by a simple robot really going to stick around for the tougher subjects or even make good programmers? I help educate high school students on degrees at my university, and the overwhelming majority of students I talk to only want to "make games" because they think it is an easy or fun process, but neglect to realise it takes a good understanding of program theory and design before even considering such a career.

I understand graduate numbers are down, it is a problem we are facing at my university as well, but you have to look at getting potential CS students, not just any student, and this seems to be a waste of time and resources.

All The Cool Kids Are Doing It

[Alamoth]

At Lehigh University where I just finished up my B.S. in Computer Engineering I was able to take part in the creation and infusion of a robotics curriculum into our CompSci department. The response was incredibly positive. When we opened up our course catalogues one semester to find that "Real-Time Vision Processing for Autonomous Robots" would be a course offered along with "Mobile Robotics" and "Robocup" we were ecstatic. Artificial Intelligence has always been a big seller in CompSci departments but it has been theoretical. Imagine taking an entry level course on C++ and not being able to write code on a computer. Theory without application has its limits. Robotics brings practical, observable results to the realm of A.I. Every Monday, Wednesday, and Friday the Engineering School invites prospective students to tour the labs. Part of their tour is the CompSci robotics lab. They are privy to demonstrations of work being done with the Sony AIBO and several other robots that were all made in the labs. Needless to say that the biggest thrill for almost all the prospective students (and especially their parents) are the robots. They are simply enthralled by the thought that at our university we have computers that can (to an extent) think for themselves. Computer Science as a college discipline has come to a point where departments that don't incorporate robotics soon will find their enrollment dwindling!

Re:All The Cool Kids Are Doing It

[Xiroth]

Eh, AI can be applied relatively easily with a little bit of work. When I studied AI, my lecturer created a 2D virtual environment, and we had to write agents in LISP which could explore, gather food (for utiles), stave off predators and do it faster than the agents that other people had written, in a variety of scenarios. Good fun.

Re:All The Cool Kids Are Doing It

[Alamoth]

Good fun indeed. I've done similar projects. However I've always felt that AI that is rooted in an environment that is completely deterministic, such as another computer program, is not as challenging as putting a piece of software into the real world.

The "predators" you're staving off are also A.I. and therefore have the same limitations and adhere to the same rules. Try programming a robot to go out into the Amazon Jungle and evade true predators that have no hardware limitations on their thinking.

This is the application of A.I. that belongs in high level CompSci courses and in graduate work.

Re:Oh Great

[Adelbert]

And so it all begins, the downfall of computers and humanity.

What are you talking about? Robots are our friends.

Re:WTF!!! (o/t)

[Anonymous Coward]

sounds like it's time you installed; Linux (fedora rules!), Firefox, Adblock, filterset-G... I would have hoped that someone with a karma bonus would have already had these things...

Re:Let's get all the cliches out at once

[de_smudger]

I think they're running out of funny those jokes - even combining beowulf, overlords and does it run linux into one-great-big-meta-joke isn't really working any more...

*wipes a tear*

Kudos to Microsoft

[Illserve]

I hate Microsoft as much as the next card carrying slashdot reader but I'm glad they're doing this. I'm sure they have a profit motive of some kind, but this funding scheme can't help but to improve the state of education.

I have to wonder what kind of robots these are that cost so much money however. Robots like this should cost about $100 -$300 tops.

Re:Kudos to Microsoft

[bit01]

... can't help but to improve the state of education.

Depends on whether it's general education or M$ vocational training.

And don't forget that most of the "funding" is likely to be M$ licenses, pseudo-money that costs M$ nothing. Hardly kudos for that.

---

Keep your options open!

Persocoms?

[Lord of Hyphens]

I, for one, welcome our Chobits overlords... and I'll take a Chii model.

Re:Persocoms?

[AndyboyH]

leaves more Yumi for the rest of us ;)

Good old Computer Engineering memories...

[Lord_Slepnir]

We had those in my freshman computer enginnering classes. We had these little Rug Warrior robots that we got to program to do crazy things like navigate a maze, measure the area of a room, etc. They were a great introduction not only to the field, but also to a lot of other concepts such as the C language, pointer, and working with registers and drivers.

sadly, our Computer Science department is moving in the opposite direction. They recently changed the first language they teach freshman from C++ to Java. I can't think of a worse (learning) language. They're not getting basic concepts such as memory management, pointers, improving performance, or debugging without some fischer-price gui there to hold thier hands.

Re:Good old Computer Engineering memories...

[Anonymous Coward]

I agree. They had just moved from C++ to JAVA when I enrolled in my CS101 class at Vanderbilt and it was fine, easy, great. I loved that class, good prof, too. But then I enrolled in CS201, a C++ course (and second on the list of programming courses you need for a CompSci minor), and was told that I would be programming for at least 5 hours a week outside of class (this is a class I was taking mostly for kicks and because I like computers, mind you) and that we'd "transition from JAVA to C++, since most of you don't seem to have had any C++ background". The "transition" was mainly him (new professor, straight from China) occasionally saying, "This is similar to JAVA!" and continuing right along rambling about data structures and pointers. I still don't have a totally clear idea of what a pointer is/does or why anyone would want to use one, and I spent 5 weeks in that damn class before dropping it so I could focus on Thermo and Electrical Science (I'm a MechE). JAVA is a good language (IMExtremelyHO), but please, teach us the basics on the most complicated language you can find.

Oh, and engineers were mostly encouraged to take CS103 instead of 101. 101 was JAVA; guess what 103 was?

MATLAB.

*shoots self in head*

I know the real problem

[kukickface]

If their professors think computer science is just about debugging programs then the solution is quite obvious: get new professors. Maybe it's the way the article is written. Maybe it is something being jammed down the faculty's throat by higher ups. Either way, computer science can be taught with absolutely no emphasis on programming at all. Programming itself was the practical application introduced into the mix, and now they are taking that one step further by introducing something that maybe 1 in 100 computer professionals (mostly former/current electrical engineers by the way) actually deal with on a daily basis? How practical is that?

Lack of Interest

[jchawk]

The problem with Computer Science right now is it's not the "hot" field. Most kids going to college are going to college so they can get out and earn a better living then if they didn't go to school...

The job market for computer science folks is flat right now with respect to new grads... If you don't have 5 years or more experience you are likely to have a difficult time finding a jump off point in the business.

Honestly I can say I don't help much... It's hard for me to hire grads out of college. They tend to be relatively worthless. They have 0 business experience and can't function without constant supervision. It's easier for me to just go out and hire someone with more experience... Until the job market heats up again and IT people are in demand I think most companies will continue to snipe the best people rather then someone new.

More Details - MS Press Release

[Anonymous Coward]

More Details - MS Press Release

http://www.microsoft.com/presspass/press/2006/jul0 6/07-12PersonalRobotsEducationPR.mspx

http://www.roboteducation.org

[Anonymous Coward]

There's more information on their http://www.roboteducation.org/>> website

another nice distraction

[stibrian]

how bout they teach debugging programs? Looking at the CS grads I graduated with and a lot of the new grads I interview, they don't need a robot, they need to learn how to write software...

A new dev that can't follow a stack trace isn't a dev at all... if they had a cool robot, that does us no good at all.

Oh dear

[spellraiser]

Can't you just hear it?

"Algorithms. Don't talk to me about algorithms."

A more appropriate way

[gr8_phk]

I always said teaching game programming would be a good way to get students interested and to maintain their interest beyond the classroom. Simple board games can be used to teach data structures and search algorithms. Simple 70's or 80's style arcade games teach real-time methods and basic cooperative multitasking. OOP anyone? The best part is that when the class is over, students are more likely to continue on their own. With a little thought, you can cover most of the CS spectrum using various games.

It's all a plot....

[hanshotfirst]

...Microsoft must be resurrecting Actimates Barney!

By getting entry-level programmers writing robot code they will be pre-disposed to the Actimates API, and will therefore build micro-borg robots instead of open-source robots.

Profit is not evil..

[Tominva1045]

There are many of us here who love Microsoft. We just get drowned out by all the one's who do not.

Just because Microsoft is involved doesn't make this a "scheme" - as if they are up to something evil.

So Microsoft makes profit, so what. The company that made the components inside the computer you are using made a profit. This morning you got up and ate breakfast - the company that sold the food made a profit. And sometime today you will go potty. The company that made the toilet paper also made a profit.

Unlike the food and toilet paper, you DON'T have to buy the Microsoft software. So, the food and TP guys are preying on your need for their product.

I don't hear you slamming them. Why is that?

Bryn Mawr should fire Dr. Blank

[the_tsi]

"The idea behind the program, Blank said, is to make computer science more hands-on and practical, rather than simply about debugging programs."

If his classes and the department curriculum is "simply about debugging programs", I never want to hire (or work with) a Bryn Mawr graduate. If his statement is not accurate, then he's maligning the school and the department (and, to a great extent, the entire field).

Although, when I read TFA, I notice there aren't quotation marks around the words attributed to Blank, so it could be that the reporter just wasn't listening when doing the interview.

Let me guess, robots provided by the NSA?

[elrous0]

Sure, it's all fun and games until someone loses an eye.

-Eric

DANGER DANGER!

[roman_mir]

Don't fall for this trick, what are their REAL motives? What will happen if an entire generation of future computer scientists fall pray to human eating/destroying robots? Everyone knows that default programming of any robot includes these very very simple steps:
1. Find humans.
2. Kill them all.
3. Define moment as 3000 milliseconds.
3. Collect some pretty flowers and enjoy the moment.
4. Go to 1.

Re:DANGER DANGER!

[roman_mir]

Of-course maybe the robots shouldn't have killed ALL the human programmers, because obviously it was a human sympathizer who wrote the above program hoping to CONFUSE the default robot language compiler by introducing an ERROR ERROR ERROR ERROR ERR...

But another programmer who is not a robot decided to fix the bug in the program, and thus this was created:

1. Find humans.
2. Kill them all.
3. Define moment as 3000 milliseconds.
4. Eat some brains.
5. echo Muhahahahahahaha
6. Collect a large properly formated data file OR a puppy that is electronic, you know, the good kind.
7. Go to 1.

I didn't quite say that...

[dougblank]

Note to self: when talking to the press, don't use complicated technical jargon, like "debugging" :)

I think what I actually said was "rather than debug a program to make it give the right answer, the students must debug the program to make the robot behave the way they want it to."

I think many of you will actually like the hardware, software, and curriculum that we are designing. Checkout http://www.roboteducation.org/ and http://pyrorobotics.org/ The new version of the software will be based on Pyro, Python Robotics. We think of the hardware as something like an ipod on wheels. The software is also being developed with an open source license. This project is not what many of you guess it might be.

The CS1 and CS2 that we are developing won't be watered down, but also won't be just the standard "intro to programming, using robots". It's a complete rethinking of the intro courses.

-Doug Blank

Re:I didn't quite say that...

[Anonymous Coward]

I'm curious why you decided to introduce robotics in your intro classes rather than as elective courses, special programs, or as part of your upper division course. With all the "stuff" being introduced in those early classes -- algorithms, languages, compilers, logic, etc -- it seems like trying to add something as complex as robotics into the mix would be too much for most students to absorb. Obviously if you've decided to try it in the intro classes rather than through some other approach there much have been some thinking as to why that was the right route to go...

Call me when...

[dsfox]

Call me when they open enrollment to robots.

O.G. Programmer: Back in my day

[Anonymous Coward]

> "The idea behind the program, Blank said, is to make computer science more hands-on and practical, rather than simply about debugging programs."

Yeah, we don't want students to actually know how to debug programs. After all, they aren't going to be doing much of that in the real world. WTF, the typical CS graduate doesn't know crap about debugging or writing bug-free code. And colleges want to refocus them on robots, which 99.99% of programmers will never program in their life. I guess that must mean that toy robots are becoming cheap. Colleges never teach expensive skills, but will go for something flashy if it's cheap. I'm thinking Lego Mindstorms.

Oh well, only a fool would major in C.S. today anyway. Unless you plan on living in China or India, there aren't going to be any jobs for you anyway, so you might as well study something useless.

Robots are DEFINITELY not any fun...

[Mister Jimm]

...unless they're Turing Complete.

Who needs computer science for that?

[bobs666]

I was programing robots in artificial classes in the 1970's

The real meat of this story is this.

Microsoft Corp. Looking to boost enrollment in computer science classes.

Its MS that made computer science obsolete in the first place.
When every 12 year old is an innate computer expert who needs
computer science classes. And if you buy into the MS paradigm
any research you do will be simply more padding for MS's pockets,
as if there was any need for that. A few dollars in advertising
is not going to fix the problem.

Why would any one write a computer program when
all the typewriter programs have been written by the
people in Redmond. That what people mostly use computers
for to type stuff. Who needs computer science for that?

Georgia Tech's savior[ette]

[electrosoccertux]

Finally! Georgia Tech has listened to our begging and is finally going to improve the Tech Ratio(TM). Not by admitting more girls (after all, the problem isn't the number of girls at Tech, it is the _quality_), but by making them! How much better can you get?! Discussions were under way to implement a send-your-picture-with-application requirement for the female applicants. After a girl met the academic requirements, a comittee of 15 or so guys would determine if she met the "attractiveness" standards. This was all fine and good until we realized we could never get away with it. Womens rights activists would nuke our school.

So this solution is much better. Not only will the fembots improve the ratio, they will help ensure every student learns all the information in their books through certain...erm..*persuasive* means.

Pity the poor bots during tests

[sponglish]

"Yo, threepeeoh, the answer to number 7 or it's the hammer!"

mod parent funny

[Anonymous Coward]

Mod parent
+1 funny
-1 dammit the bastard beat me to it

Lego Mindstorms

[neonprimetime]

My undergraduate cs department purchases some Lego Mindstorms off eBay and used them in the intro courses. They don't cost much (couple hundred max), so our tuition didn't go up anything. You got to write programs for them in Java. It was very exciting and sparked lots of interest (everybody wanted to take the class). Although it's not as cool as each student getting an individual robot, it is as close as some of the smaller campuses can get, and it's a great idea!

Re:Lego Mindstorms

[Anonymous Coward]

Tennessee Tech has been using Mindstorm robots for two years in the intro to cs course. The robots are not programmed directly in Java, but in a language developed by a group of graduate students that is translated to Java. Since this course it not only taken by first year cs students, but also students from many other majors, it provides a more friendly introduction to programming. I do not see robots being useful though later on in the program though.

Re:Lego Mindstorms

[neonprimetime]

I do not see robots being useful though later on in the program though.

Agreed. Most of this stuff was kids play, this recognizing whether or not your hit a wall and which direction you were going. It could become very interesting though if it were in an artificial intelligence course. There are several additions you can get such as a device that determines what color surface you're on, what temperature it is, etc. Programming, for example, a bot that could complete a maze (remembering where it has already been, where the exit is, etc.) can be very interesting.

Re:Lego Mindstorms

[sugarmotor]

Which college?

Stephan

Re:Lego Mindstorms

[Starji]

Oregon Institute of Technology. Graduated about a month ago.

Re:Lego Mindstorms

[voorko02]

Along those lines, the next version of Robosapian will be compatable with Lego Mindstorms.
          http://www.engadget.com/2006/07/08/robosapien-rs-m edia-gets-new-lego-mindstorm-nxt-brain/

So thats a cheap option for colleges looking to sex up the CS major.

Re:Lego Mindstorms

[neonprimetime]

So thats a cheap option for colleges looking to sex up the CS major.

I would still prefer more female cs majors instead. Wouldn't it be awesome if part of a college's recruiting slogan is that they have a large number of sexy female cs majors?!?!?

Re:Lego Mindstorms

[Drooling Iguana]

What if you wanted it to kill all the humans?

Re:Let's get all the cliches out at once

[Anonymous Coward]

The really pathetic thing is that not only did someone waste mod points on the GP, they wasted mod points moddind down yours as well. Oh, well, at least they won't have to worry about this post since I posted AC ('course, they probably will anyway...)

Fresh, new ideas

[Syberghost]

You mean there are schools left that didn't do this 20 years ago? Huh.

You know theres a problem when....

[zerosix]

Last time I check programming was about creation not just debugging. Of course as with anything in life, if your first aproach doesn't work you must step back, look at the problem again, and try to get it right, ie debug. I'm seriously begining to think that if students are having such a difficult time with debugging, then maybe more time should be spent on actually teaching the students proper programming tactics.

I mean, if you are trying to get people to join computer science classes simply because there is something cool involved I don't know that is the best aproach. It's like giving them an ipod just because. And last time I checked there are robots at the school I attend although not until higher level courses.

Plus I think you need to be just a little bit of a perfectionist to really succeed in programming. If you have a problem debugging code because you are too lazy or whatever, you shouldn't be programming. And on a side note, when is computer science not hands on??? Only when you are taking theory classes and even in those I have always had practical programming assignments. To me the issue seems much bigger than what the article is letting out....

Great idea b/c it works

[ayounge]

i am currently teaching a summer course called 'programming and robotics'. its designed for highschool kids who are interested in both sections. from personal experience, the kids enjoy a lot of the programming, however u really see their interests spark when the robotics kits comes out. its a great because it makes them more interested in programming and also gives them something more concrete to put into their hands, which is very important.

CS != Debugging, but Programming = Debugging

[Anonymous Coward]

CS is about learning the meta-stuff that will help you design programs, subprograms (functions), and systems, with an awareness of practical constraints ("good" vs "bad" algorithms, time efficiency, space efficiency, time vs space tradeoffs)

While learning about the hardware is not really CS, it can provide very useful *background* info for CS, to explain why going to disk is a million times slower than a memory access, etc.

Even though CS is about programming, it is definately NOT programming. You can learn all of CS without ever writing a single line of code (though I would argue that it would be hard to really grok parts of it until you've been coding some). And certainly, you don't have to be able to debug at all, to learn CS.

Programming, on the other hand, is mostly about debugging. Once a system exists, you quickly find out that it doesn't do what the client wants it to (even if it does everything they ASKED), so it must be changed. Even if the system was perfect (ha!), changing it will require debugging.

If you program a system, and continue to support it, you will very likely need to spend far more time debugging than on adding new features. Though, I am including the time spent studying existing code, to puzzle out how it works, so that you know how to safely add changes, as a sort of 'pre-bug' debugging.

Anyways, my point is that you should not make light of debugging. The lead developers need to have a good grounding in CS and design and all that. And its never bad for everyone else to know all that, but day-to-day, it's more important for most programmers to be good debuggers.

And this is new?

[Richard Frost]

Drexel had a robotics course a few years ago when I was still in college, using a Lego-like system. The focus was on mechanics and locomotion (gear ratios, etc.), AI (how to deal with uncertain input from sensors), and concurrent programming (how to get the sensors and motors to work with each other).

My University is during this

[Com2Kid]

Not the actual CS intro classes, but instead what amount to "feeder" classes that are designed to get students interested in CS who might not otherwise think about the field.

We have two such classes, Intro to Game Programming, and Intro to Robotics.

I know that the CS Department managed to get at least one of them designated as a GUR (General University Requirements) course for Mathematics and Logic, hoping that students will find Health = Health - ShotVelocity*ShotPower more interesting than y = x^2 - x.

Personally I find both interesting. :)

In what could be the second CS course you take in our program, the first Functional Programming course (there are 2 series of courses taken at the same time, one traditional Procedural/OO and the second is Functional), you end up programming a 3D Raytracer. This either inspires students, or causes them to change major. More of the latter I am afraid.

Exploring Robotics @ Brooklyn College

[Prof.Phreak]

I'd also like to point out that Brooklyn College is introducing ``Exploring Robotics'' as a ``core'' (specifically for non-CS-majors; more like an upper level basic computers class for everyone). I believe it's planned to use Lego stuff---and it's being offered starting this Fall.

For majors, there are other options (as in, taking an AI class with a professor who uses robots, or joining a group and programming AIBOs, etc.)

Re:Exploring Robotics @ Brooklyn College

[dougblank]

Prof Phreak,

We'd like to pick your brain. The goal of the institute is to try to make a shrink-wrapped robot so that every student will have their own, it will be more advanced than the typical Lego-based robot, and yet inexpensive. The intro course will be for majors, but not be just the same ol thing. If you have suggestions, please feel free to drop us a note over at http://www.roboteducation.org/ Thanks!

Lego Mindstorms

[Starji]

One of my first computer science/engineering courses used Lego Mindstorms for us to get in the habit of problem solving. My favorite project was one where we needed to make a line follower, and then we raced our creations against each other for the best times. That combined both hardware and software engineering concepts because we had to build the car to go fast, but also needed the software to tell the car what to do and not to screw up on us. We used that 'block language' that comes with the mindstorms for the programming. All in all it was one of the most enjoyable classes I had in college.

Pfah!

[mark-t]

Computing science courses could be made more attractive by there being more jobs out there for new CS grads.

Uhmm... sorry... do I sound bitter?
Maybe just a little.

But do the robots have a CANNON?

[Anonymous Coward]

When I wrote for C robots, mine did! (too bad it was virtual... would have solved certain angst problems) =)

sweet! i get to paly with 'bots at college...

[alphastryk]

so yeah...second semester at tech, we get bots...and they look awesome...

Worst idea yet.

[Penguin Programmer]

I sit on the computing science curriculum committee (as the student representative) here, so I've been actively involved with lots of ideas and planning regarding how to increase enrollment and continuance in our CS program. I think introducing robots into the first-year courses is the worst idea yet.

Why?

1) Not everyone who's into CS is interested in robots. I'm as hardcore as CS students come, and I'm not into robots in the slightest.

2) Robots are fiddly and frustrating. They teach robotics courses here with AIBO dogs and RoboCup soccer robots, and everyone I know who's taken the courses says they're annoying as fuck to work with. They break a lot, the sensors are garbage and you spend as much time compensating for the deficiencies of the robot as you do actually making it work properly. Which is fine in a fourth-year course about robotics, of course, that's what it's supposed to teach: here's a field where things don't work right and you have to compensate. But in a first-year course, this is a terrible idea.

3) It's not a good introduction to computing science. It gives the mathy people the incorrect impression that CS isn't mathy at all (which they won't like and will make them switch into math), and it gives the non-mathy people the incorrect impression that CS isn't mathy at all (which they will like, and will cause them to continue and then fail out in second year).

They're experimenting with the introductory courses here this year. The idea behind the experiments is to introduce students to the _real_ guts of computing science to start off with - finite state machines, data structures, algorithms - without going too in depth and scaring people off and while at the same time showing real applications for CS and teaching a language (Perl or C++, depending on the course). I think this is the right approach: it will attract the people who are truly interested in computing science, whether they knew they were before or not. I _wish_ that my first-year courses had included this kind of thing. If I had been one of those students teetering between CS and another field, my first-year Java house-drawing course certainly would have convinced me CS was the wrong way to go. So would robots.

Re:Worst idea yet.

[dougblank]

Penguin, Good points. First, it is exactly this type of student involvement that will help change CS. Before you completely throw the baby out with the bathwater, though, consider these points: 1. This CS1 isn't aimed at you (or others that are interested in CS). It is aimed at those that aren't. The idea is to attract them into the field. 2. The robot will be designed to be as sturdy as an ipod. 3. How do you that the curricular materials won't be a good intro? Maybe a better concern is: "how will we work mathy things into the course?" You just have to remember that we are trying to attract the anti-slashdot student. What is she looking for in an intro course? Help us figure that out at our blog: http://blog.roboteducation.org/

The GT Robots

[Rachel Lucid]

I've seen the robots they plan to use though, and I'll be shocked if they're any bigger than the AC adapter for a laptop.

Oh, and thanks for the webcomic link. *Loves Firefox's Morning Coffee Extension*

Re:The GT Robots

[eighty4]

Bah, humbug. Robots shouldn't be called robots unless they're hell-bent on world domination.

ps. Thanks for the extension tip, I never thought of anything like that. I'll check it out.

Logo Turtles for freshmen.

[bobs666]

So Microsoft is going to spend only $1 million to bring
to boost enrollment in introductory computer science classes.
Why? When all the 12 year olds are all ready computer experts.

The result will be Logo Turtle in hardware. It will be just the thing for there 12 year old minds.

All wrong

[vga_init]

First of all, robots are stupid! I don't mean that they suck (by which I don't mean that they do anything orally), but that they aren't intelligent machines like some people imagine. In fact, robot programming is very tedious and only fun for a select group of individuals.

I think a better idea would be to include computers instead of robots. I mean, it's a computer science course, right? And before you get on my case about affording computers and whatnot, when we're talking about intro to computer science, we should be talking about doing some very basic programming (no pun intended) so that students can learn things like logic, control, and design. A very low power machine can be had very cheaply--hell, there are graphing calculators for less than $100 that are more programmable than the majority of PC's (I don't mean to say that PC's aren't capable of being programmed for, but where are the software tools? On an average Windows system, no one can just sit down and write software unless you're terribly interested in x86 ASM on a virtual machine. You're going to have to install additional software whereas the little calculators have built in development environments).

Capilano College for a while now

[sugarmotor]

Capilano College in North Vancouver has been offering Comp 106, Programming with Robots, for a few years now.

"OBJECTIVES: To introduce students, with little or no previous computer experience, to the basic concepts of hardware, software, and computer programming using Lego Mindstorm robots as a teaching tool. In addition to fundamental programming concepts, students will also be exposed to different topics in Computer Science, including robotics, hardware, operating systems, communication, and social issues. "

http://www.capcollege.bc.ca/programs/computing-sci ence/courses/c106otl.html

http://courses.capcollege.bc.ca/comp/gallery/robot s.htm

Stephan

Already do that...

[Jaysu]

At Georgia Tech, if you are EE you take a digital design lab second year in which your final project is to program a robot. Robot or not, you're still sitting there debugging programs.

Pyrobot not just for Inro CS

[isdale]

The Pyrobotics software has been out for several years. It was originally developed by Doug under a grant from NSF. That grant ran out last year and I am very glad to see MS has picked up the ball. Doug has been using Robotics to teach CS as part of the NSF grant. I got to sit in on one of his classes this spring and it was quite interesting to see sophmores, etc discussing high level AI concepts. Those who are writing to compare Pyrobot with the Lego Mindstorm or AI/engineering classes would do well to go read the papers posted on the pyrobotics.com web site. They specifically address issues with those other approaches: languages unusable outside the one robot, spending too much time getting the b#&$! hardware to work and not enough on CS issues.

The Pyrobot system is also not just for CS teaching. It makes a very nice system for working with robot control, especially when using a simulator instead of a real bot. I have been using Pyrobot in conjunction with the Delta3d OSS simulation system. Delta3d provides realistic rendering on real terrain for multiple robots while Pyrobot lets us experiment with very high level 'Brains' to control the bots (UAVs, UGVs, etc.) Python makes it very nice for the AI folks with lots of support libraries, rapid development, etc.

The downside of using Python for Intro CS, IMHO, is that it sidesteps so many basic Software Engineering and other basic software potty training issues. I look forward to seeing how Doug and his associates at GATech integrate those into their classes... and perhaps picking up an intern to work out here in Malibu some summer.

Programming Robots Not About Code

[Dareth]

Programming robots is not always about code.

I once asked how industrial robots were controlled. I was thinking cool code, scripting languages. Unfortunately the answer was that they use more of a "macro" approach. They have a human who knows how to do the task the robot will be performing manually move the robot thru the motions and they record it like a macro. Then the robot can just repeat these motions to do the task. The macros may be edited for efficiency of motion, but overall not alot of programming going on.

do robots drink tequilla?

[gemcigital]

teaching kids to code using robotics is a great idea with deep roots. I recall the turtle language Logo from WAY BACK WHEN. I also recall burning a hole in the carpet at my rental pad in Bean Blossom working on "carbot" with professor blank.However, I would like to see a robot that automatically conjures up tequilla shots, including sliced lemons when asked. Doug, can you get on that please? gem company http://www.cigital.com/ podcast http://www.cigital.com/silverbullet book http://www.swsec.com/

Iowa State

[Loconut1389]

When I entered Iowa State as a freshman in 2000, I took Computer E 183X and 184X (I'm pretty sure those are the right numbers) and we had little robots as part of our curriculum (I believe they were Rug Warriors). It was a great project, we made them wall-hug and do all kinds of neat tricks like distance finding with ultrasound.

Note that the X part of the course names meant experimental- I'm not sure if those courses went mainstream or not, but point is, ISU was checking this stuff out 6 years ago and as a student, it was fantastic.

Re:Iowa State

[Loconut1389]

Also note, the robots were the property of the college and remained in a lab, but were available for after-hours checkout.

Robot girlfriends

[Yahiko]

Now all of them will have a platform to start building their companionship from.

--CS Graduate

ULM Alumni?

[Dareth]

When did you start your degree in CS at ULM?

I started in the mid 90's. I was one of the last people to take all of the CS courses in C++ before the switch the Java. What do you think about the CS programming merging into the CIS program?

Science is hard

[Anonymous Coward]

Science is hard.
Therefore we don't have enough people going into science.
Let's make science easy and entertaining.
That way we get a lot more people going into science, even if they're stupid.

Whatever you do, don't invest in research programs that give scientists meaningful employment that inspires other science-minded people to go into science. Also, be sure that scientists get no recognition for their work so that children see it as worthless waste of time that it is, unlike socially significant activities like football and basketball.

uh oh

[Weasel474747]

I'm only a few miles away from Bryn Mawr. I guess it's about time I buy an insurance policy that covers robot attacks.

TekBots at OSU

[svnt]

Oregon State University has a similar program under the heading "Platforms for Learning." The TekBot program started in ECE disciplines, but it also branching into ME with mechatronics and is probably headed toward replacing Mindstorms in early CS classes.

http://eecs.oregonstate.edu/tekbots

Freshmen designing, programming, debugging robots

[Plocmstart]

http://feh.osu.edu/Design-Project/Design-project.h tm I've been a part of this since '98. It doesn't get as much publicity as it should though.

robo-no

[filthy_mcnasty]

How about a few more members of the female population instead of robots? That would get more enrollment guaranteed. I don't remember more than 3 girls in any class I ever took.

Doing something similar

[Bilby]

I've been planning on doing something similar in a first year programming course I run. There are a number of reasons, some of which are quite academically oriented, but the two big ones are:

a) Stepwise refinement in algorithm design, which is something I really want to get the students to understand, can be illustrated easily if the action being refined is something that the students are already familiar with. Bubble sorts are ok, but picking up an object, putting it down somewhere else, and waving is better.

b) Back when I started programming, (which was a while ago now), making your computer say "Hello World" was wonderfully exciting. We wren't bored with computers, and weren't disappointed when two semesters of study still didn't mean that we could write our own version of Half Life II. But making a robot do something - even something fairly basic - is exciting, because it isn't something that we are bored with yet. I want programming to be exciting. Bubble sort routines may be worth learning (maybe) but they are never exciting.

Mind you, I wouldn't want to insist that students pick up a $500 robot to do the course, either. At the moment I'm demoing the robots to illustrate concepts, and in the future I'll try to build something that they can play with. But it doesn't need to be complex - it just needs a decent micro controller, 2-4 sensors, and some ability to move within and interact with the environment. Not at all unlike those linked to from the article.