Honor Principle vis-a-vis COSC 31, W25

Dartmouth's Academic Honor Principle applies to this course. Please be sure to read the principle, which you can find here. Below, I am writing the policies pertinent to this course offering.

• Weekly Problem Sets.  Working with other people on the problem sets is allowed and indeed highly encouraged. Indeed, if done right, working with others (learning from them and teaching them) will make you learn efficiently. However, what you hand in, must be in your own words. You also must include a "collaboration/credit statement" for each problem you hand in listing everyone (name names) you talked to about this problem (instructor, TAs, classmates, pet dog, etc). You must include this even if you worked alone stating that you collaborated with no one.
• Exams. The exams are timed, in-class and closed-book unless specified by the instructor should be taken with access to nothing at all (no internet, notes, etc).
• Drill Quizzes. These should be done individually. If stuck, you can and should ask for clarifications from instructor/TA/friends, etc. Remember: this is extra credit.
• Coding Assignments. This should be done individually, You should neither share your code with some one else, nor obtain code from someone else. If you access the web for something basic (eg, "how to delete an element from a list in python" or "how to read a file in google co-lab") and then use snippet of code, then you must CITE the source. If you don't use the snippet but just learn and type the code yourself, then no need to cite. Please don't "search" for the problems on the web. Remember: this is extra credit.

Generative AI and CS31

Can we use ChatGPT? Short answer: I won't recommend it. But if you do end up using it, then you (a) must cite and (b) tell exactly how you used it. Using GenAI and not stating these is a violation of the Honor principle.

You could put in any of the weekly problems into ChatGPT and it's likely it will give you something sensible. It doesn't necessarily give something correct or as efficient (especially for the latter half of the course). My impression for CS31 questions is that a lot of "prompt engineering" is needed to get a "near perfect" answer from ChatGPT (one that'll get a 5/5); the first output is often so imprecise that it doesn't make sense (but to a lay-person it may read perfectly okay). It is also quite easy to see which answers are GenAI generated. So beware. For coding assignments, GenAI is much better (as is well documented by now) but still sometimes makes errors. 

However, ChatGPT can be useful for auxiliary stuff. For example, you may be using LaTeX and the TeX file isn't compiling well: ChatGPT can solve your problem faster than googling it. Or, you want to know how to plot a graph in Python 3: ChatGPT teaches you this much more efficiently. 

So, if you use it, then use it for the right reasons, and cite properly.