7 ms·

My CS degree was 100% C based; Java shows up in a class about OOP, and our class on Programming Language theory exposes us to Lisp, Java, and Prolog. They used

by dljsjr 13y ago

My CS degree was 100% C based; Java shows up in a class about OOP, and our class on Programming Language theory exposes us to Lisp, Java, and Prolog. They used to be Java based and chose to switch back to a C based curriculum around 2007, as well as focusing more heavily on algorithms and data structures while pushing the more "software engineering" type courses in to elective and offering Software Engineering as a specialization on a CS degree. Joel Spolsky wrote a great article[1] in 2005 that basically says using a massively high level language like Python or Java as the core of a CS curriculum is damaging because the languages aren't "hard" enough. As curmudgeonly as it might sound, I'd be inclined to agree. Even though I rarely write C code anymore (we do everything in Java in our lab), the underlying understanding of memory management that is acquired from having to learn to handle it manually has been invaluable to me as a programmer. Additionally, C/C++ is the language of UNIX and still the lingua franca of high-performance. Most high-level languages worth their salt will have a foreign interface to dynamically load C libraries; we're in robotics and have to work heavily with ROS/Gazebo as participants in the DARPA Robotics Challenge, and even outside of this a lot of really great tools for mathematical optimization and high-performance numerical analysis are all written in C. I'm one of only two or three people in our lab with the requisite knowledge to be able to effectively hand-roll JNI wrappers, and I don't consider this to be a good thing at all. It's pretty clear to me that C/C++ aren't going anywhere any time soon, and it's always good to see a CS curriculum that still focuses on C/C++. 1: http://www.joelonsoftware.com/articles/ThePerilsofJavaSchools.html http://www.joelonsoftware.com/articles/ThePerilsofJavaSchool...

dragonwriter 13y ago
> Joel Spolsky wrote a great article[1] in 2005 that basically says using a massively high level language like Python or Java as the core of a CS curriculum is damaging because the languages aren't "hard" enough. There's certainly a value in getting closer to the machine (as C does), OTOH, there's a value in not having a lot of boilerplate cruft between you and the algorithm, which makes something like Python valuable. Really, I think a CS program that exclusively uses any one language is probably robbing students of valuable insight that comes from the different perspectives that working through different languages (particularly, languages from different design paradigms) provides.
twrkit 13y ago
I studied Computer Science at USC and was a student in 2 of Professor Redekopp's classes. Suffice to say that he was a top notch educator and the favorite professor of most students in the Engineering school. At USC, at least while I was there, 101 and 102 (the class from which these slides are linked) were taught in C++. The Algorithms class used prolog, and the majority of the remaining core CS classes were in Java. All in all, I'm very grateful for the diverse foundation we were provided, it has helped me immensely in my career as a developer.
- zenbowman 13y ago
  Sadly I never took Redekopp as the AP CS course had me place out of CS101. CS102 was taught by Crowley when I took it, and four years later when I was TAing it.
- michaelrbock 13y ago
  Professor Redekopp was one of the best teachers I've ever had. I highly recommend any incoming CS students take him for any classes you can.
cube13 13y ago
>Joel Spolsky wrote a great article[1] in 2005 that basically says using a massively high level language like Python or Java as the core of a CS curriculum is damaging because the languages aren't "hard" enough. As curmudgeonly as it might sound, I'd be inclined to agree. Even though I rarely write C code anymore (we do everything in Java in our lab), the underlying understanding of memory management that is acquired from having to learn to handle it manually has been invaluable to me as a programmer. Python and Java are great starting languages, especially for the first introductory course. You're really only learning the high level concepts like basic language syntax(Why are there so many semicolons???), program flow control(functions, if, for, while statements), working with computer memory structures, and basic boolean logic. At that point, an easy to understand language actually makes more sense, because the students are working more with the concepts than with the language. After that, when you're teaching data structures, it's important to know how the memory is allocated and how that can effect your implementation of the structure. C/C++ are good languages because they force students to deal with memory allocation and pointers. They also don't have pre-rolled data structures like python or java, which is important because the goal is to learn how the structures work, not how to use the standard language libraries.
- plg 13y ago
  My opinion is that one can easily teach these "high level concepts" (loops, syntax, control flow, etc) using C. Furthermore my (admittedly curmudgeonly) take on the C vs "high level languages" debate is that learning something like Python "first" because it's "easier" is actually sort of damaging, in that the student will learn these "high level" concepts without any clue about the underlying implementation. They will then have to learn a bunch of "bag of tricks" that will seem arbitrary to them, and a bunch of "gotchas" to avoid, again, that will seem arbitrary to them, because they have no concept of things like the cost of memory copying (passing-by-value giant arrays as arguments to functions, etc). I actually think that someone who is incapable or even worse, unwilling to learn programming using C, has no business programming anything at all of any import, beyond the most toy-example tiny scripts here and there.
  - cube13 13y ago
rayiner 13y ago
My CS class in HS was in C++ (Turbo C++ in DOS mode with the Borland graphics). I'm undecided as to whether it was the best approach. On one hand I appreciate learning it. On the other hand, it makes you think too much about the computer to appreciate the algorithm.
- pjscott 13y ago
  I learned algorithms from Segdewick's "Algorithms in C", and had a different experience: I really liked understanding the algorithms and data structures right down to the level of bytes in memory. It felt... less magical. In a good way.

by dljsjr 13y ago

dragonwriter 13y ago

> Joel Spolsky wrote a great article[1] in 2005 that basically says using a massively high level language like Python or Java as the core of a CS curriculum is damaging because the languages aren't "hard" enough. There's certainly a value in getting closer to the machine (as C does), OTOH, there's a value in not having a lot of boilerplate cruft between you and the algorithm, which makes something like Python valuable. Really, I think a CS program that exclusively uses any one language is probably robbing students of valuable insight that comes from the different perspectives that working through different languages (particularly, languages from different design paradigms) provides.

twrkit 13y ago

I studied Computer Science at USC and was a student in 2 of Professor Redekopp's classes. Suffice to say that he was a top notch educator and the favorite professor of most students in the Engineering school. At USC, at least while I was there, 101 and 102 (the class from which these slides are linked) were taught in C++. The Algorithms class used prolog, and the majority of the remaining core CS classes were in Java. All in all, I'm very grateful for the diverse foundation we were provided, it has helped me immensely in my career as a developer.

zenbowman 13y ago
Sadly I never took Redekopp as the AP CS course had me place out of CS101. CS102 was taught by Crowley when I took it, and four years later when I was TAing it.
michaelrbock 13y ago
Professor Redekopp was one of the best teachers I've ever had. I highly recommend any incoming CS students take him for any classes you can.

cube13 13y ago

>Joel Spolsky wrote a great article[1] in 2005 that basically says using a massively high level language like Python or Java as the core of a CS curriculum is damaging because the languages aren't "hard" enough. As curmudgeonly as it might sound, I'd be inclined to agree. Even though I rarely write C code anymore (we do everything in Java in our lab), the underlying understanding of memory management that is acquired from having to learn to handle it manually has been invaluable to me as a programmer. Python and Java are great starting languages, especially for the first introductory course. You're really only learning the high level concepts like basic language syntax(Why are there so many semicolons???), program flow control(functions, if, for, while statements), working with computer memory structures, and basic boolean logic. At that point, an easy to understand language actually makes more sense, because the students are working more with the concepts than with the language. After that, when you're teaching data structures, it's important to know how the memory is allocated and how that can effect your implementation of the structure. C/C++ are good languages because they force students to deal with memory allocation and pointers. They also don't have pre-rolled data structures like python or java, which is important because the goal is to learn how the structures work, not how to use the standard language libraries.

plg 13y ago
My opinion is that one can easily teach these "high level concepts" (loops, syntax, control flow, etc) using C. Furthermore my (admittedly curmudgeonly) take on the C vs "high level languages" debate is that learning something like Python "first" because it's "easier" is actually sort of damaging, in that the student will learn these "high level" concepts without any clue about the underlying implementation. They will then have to learn a bunch of "bag of tricks" that will seem arbitrary to them, and a bunch of "gotchas" to avoid, again, that will seem arbitrary to them, because they have no concept of things like the cost of memory copying (passing-by-value giant arrays as arguments to functions, etc). I actually think that someone who is incapable or even worse, unwilling to learn programming using C, has no business programming anything at all of any import, beyond the most toy-example tiny scripts here and there.
- cube13 13y ago

rayiner 13y ago

My CS class in HS was in C++ (Turbo C++ in DOS mode with the Borland graphics). I'm undecided as to whether it was the best approach. On one hand I appreciate learning it. On the other hand, it makes you think too much about the computer to appreciate the algorithm.

pjscott 13y ago
I learned algorithms from Segdewick's "Algorithms in C", and had a different experience: I really liked understanding the algorithms and data structures right down to the level of bytes in memory. It felt... less magical. In a good way.