Slashdot Mirror
Big MOOC On Campus: Georgia Tech's $6,600 MS In CS
theodp writes "Next January, writes the NYT's Tamar Lewin, the Georgia Institute of Technology plans to partner with Udacity and AT&T to offer a master's degree in CS through massive open online courses for a fraction of the on-campus cost. Georgia Tech's Online Master of Science in Computer Science can be had for $6,600 — far less than the $45,000 on-campus price. The courses will be online and free for those not seeking a degree; those in the degree program will take proctored exams and have access to tutoring, online office hours and other support. AT&T, which ponied up a $2 million donation, will use the program to train employees and find potential hires. Initial enrollment will be limited to a few hundred students recruited from AT&T and Georgia Tech corporate affiliates. Zvi Galil, the dean of the university's College of Computing, expects that the program could attract up to 10,000 students annually, many from outside the U.S. 'Online, there's no visa problem,' he said."
Being a decent coder has little to do with CS. It's a very valuable skill in its own right, but quite different.
In my experience, what you describe is a doctorate program. A masters is mostly courses with research as an option.
For the UK version at the Open University the exams are held at a local college and proctored in the normal way. Presumably this could operate in a similar fashion.
Somebody please *please* hear this message before it's too late. Too many bright foreign students who get into top notch schools are denied visas. I've seen this happen first hand multiple times at a good school. Politicians can debate visa allocation as much as they want in general. But when MIT (or some other top notch school) accepts someone can you please just give the kid a visa? Oh, and not kick him out when he graduates? Because if not, then your protectionist strategy creates a market for programs such as this one, which is a hundred times worse than the scenario you are trying to prevent.
When you get a masters' degree, you spend a year or more committing yourself 24 hours a day to learning something, and you're in a community of people who are engaged in the same commitment to learning something. Your eating, sleeping, and social life revolves around an intellectual community. You learn a lot through serendipity. A chance meeting in the hall can give you a direction for your career.
When you take a MOOC, you're not giving it the same commitment and you're not among the same community. That's especially true if you take it free.
You could just read the same textbooks that masters' degree students read. But you'd be missing something.
I could read transcripts of the Feynmann lectures. But that wouldn't be the same as going to school and taking lectures with Feynmann.
UT Austin is 0.1 point above Tech in the rankings for CS Grad Schools. As has been noted, if you're in-state or on a GTA or GRA, the tuition drops precipitously or is basically waived. Whether it's a #10 or #9 school isn't really going to matter during interviews. Both are superb schools with an excellent reputation among hiring managers (and I've hired-a-plenty out of both).
Tuition rates between the two schools are not significantly different. Tech is a bit over $13K/semester and UT Austin is a smidge over $12K/semester.
I get that it's cool to hate the educational establishment and all, but if you're choosing between 20 freshly minted CS master degree holders and 20 hardcore coders, you're not qualified to be hiring anybody. Most likely a mixing of people with education and work experience is going to yield optimal results, not choosing to hire only people with one sort of experience. Especially, if you're wanting to create a product that hasn't been done to death.
There's a shitload of crap code out there written by "hardcore coders", none of which is an example to be emulated. Sure, the masters degree holders might not have experience, but they also don't have much experience writing crappy code. Which, from the comments I see around here from "professional programmers", could very easily justify not hiring people that have decades of wrong experience to retrain.
I've been there and done that. The start-up almost surely won't last forever. Even if it does, you won't want to work 55 hours a week while your baby is waking you up at 3AM. At some point, you'll probably want a nice 8-5 with good insurance and time off. When that time comes, you need letters behind your name.
I had all of the other credentials. I have seventeen years of full professional experience. I'm an Apache contributor. At one interview, the interviewer asked me if I had experience with Debian, as that was their preferred distro. I asked if he'd seen that morning's Debian security update. He seen it and applied the update. My name was on that Debian alert, I discovered the security issue all Debian users were alerted to that morning. I didn't get the job. Put letters at the end of your name while you can.
Ok, so data points - at 900k that was a decade ago - say ~'03? (I think @ 300k I joined somewhere in '99?) Maybe we can create a chart, I think it would be interesting... (and I'm sure the /. mods have all that info but am wondering if they would find it as interesting to divulge...)
There are a lot of junk online courses out there. A lot of them are simply videos of lectures, repurposed as "online courses". Stanford does a lot of that. Their original machine learning class was like that, and it is painful. Especially since the instructor's blackboard writing (yes, it's video of a real chalk blackboard) is messy. This in a field which has its own unique (and not very good) notation.
Khan Academy has courses which consist of a color etch-a-sketch display of the instructor's writing plus a voice-over. I viewed the lectures for forces and torques recently. The instructor had clockwise and counterclockwise reversed, used a multiply symbol where he needed an add, and went from talking about a body in free space to one pinned at a pivot point without mentioning that he'd shifted. Not only is the production value very low, nobody is reviewing that stuff, or even proofreading it.
MIT's course on rotating electrical machinery is basically the class notes from a course. There are a few drawings, then endless math derivations. You don't get the labs online.
I've seen some good online courses, but most of this stuff is a low-budget conversion of old lecture and notes.