Comentários e feedback de alunos de estrutura de dadosestrutura de dados da instituição Universidade da Califórnia, San Diego

The fact that test cases are not available is extremely frustrating and time consuming. I absolutely disagree with the instructors about the reason why test cases are hidden.

Michael Levin

Hi,

Learning new complex things is time consuming, it is essential, and it is ok. Without this time spent you wouldn't have learned even half of what you've learned while trying to find your mistakes yourself. Also, the testing techniques presented here are indispensable in the real life, and many learners of the Algorithmic Toolbox course have already confirmed this

Hello Mr. Levin,

First please forgive me for communicating with you in this place. I am not sure if there is a direct way to answer you, but I believe you will have access to my response.

I entirely agree with what you said that this needs time and I do not mind spending time learning new things. That is why I am here in the first place. But I think that hiding the test sets misplaces the balance where this time is being spend.

Let me give you an example: week 3 – hash chains - assignment no. 2. I wrote the app, submitted this and it failed on test no. 5. Having generated tons of test cases I was nowhere closer as all looked good, but still failed your engine. As it turned out the problem was not with the algorithm implemented but with the presentation layer. ‘Check x’ command when x is not found in the hash table should return blank line and my program returned nothing. I could not find this in the specification hence lots of hours trying to figure out the solution.

It was not a bug with the program, it was the problem how the data is presented. I do not mind spending time developing solutions to the assignments. I would love if there were more advanced (even optional) problems to solve but I want to spend this time on the algorithm and solving the task rather than trying to figure out how to display the data.

Please, do not get me wrong, I absolutely adore the course, lectures are brilliant and my opinion should be regarded as a constructive criticism even if you do not agree with me. Thank you again for what you are doing.

Regards,

Marcin

Data Structures was really interesting over all, also assignments are quite challenging. It's important to consult the external references & discussion forums if you want to get the best of it.

Algorithms in lectures were useless to programming assignment

The lectures and the reading material were great. The assignments are challenging and require thought before attempting. The forums were really useful when I got stuck with the assignments

The parts about hashing and balanced trees are bad:

1. The part about Balanced trees is completely terrible:

1.1 All the materials about rb-trees is one little article with links, no lectures and no tasks about rb-trees. But there's a detailed explanation of AVL trees. That's strange because rb-tree is a popular data structure: data bases, programming languages. And I can't give any real area where AVL trees are used in.

1.2 Quality of material about AVL trees is very bad. Author hasn't given implementation of RebalanceLeft(N) in video "AVL Tree Implementation". Explanation of tree rotation is worse than wikipedia article about AVL trees.

2. The part about Hashing is worse than I expected:

2.1 Authors described only simple rehashing. Nothing about more advanced rehashing, for example rehashing in real-time systems.

2.2 Nothin about more advanced methods of collision resolution. For example in HashMap of Java a chain with length more than 8 becomes a red-black tree.

2.3. Proof about probability of collision was confusing, Cormen's book helped me to understand the proof.

Great continuation of the Algorithmic toolbox course. There's a lot of new, interesting material here. The videos are very good, the slides are of high quality, and you will also find some good references to online university materials and interactive visualizations.

The homeworks are challenging but enjoyable, and you will also find some really challenging optional assignments here.

The only downside of the course is that unless your friends are computer scientists, you're going to lose them if you talk too much to them about AVL trees, distributed hash tables and splay trees! ☺

Even some programmers/coders scoff at these things, but as a programmer I'm certain that you will become a better at your job if you learn these.

Can't wait to start the next course in the specialization!

I found the course a little tough, but it's worth the effort. It takes more time than mentioned. Apart from that, it is actually good and covers most of the topics required for interviews.

we need more of kulikov!!!

The lecturer Daniel Kane does not explain things clearly. I constantly had to switch to other material listed in references to understand what he was talking about. I know that usually lectures are supposed to give you only general understanding of the problem and you still have to read additional material, but with Daniel Kane it is practically useless to watch lectures. I am sorry if I offend him, but lectures were a real problem for me so I think I have to speak out.

Another issue with this course was poor design of home assignments. For example, after a huge Week 5 where we cover: search trees, binary search trees, AVL trees and all operations on these trees - there was no home assignment! Why? In consequent Week 6 we cover Splay trees and get 5 problems as home assignment. The first three problems have practically the same solution, you only need a few adjustments and these problems are on binary search tree properties (not AVL or any in particular). Then, there are remaining two problems that only cover Splay trees (as far as I understand). Such assignment design makes it hard to sort out the topic (at least for me). Moreover, the rest two problems are huge and as the result you cannot check only a fraction of your alogrithm instead you have to check it entirely. And since these problems are not necessary to complete it is really hard to motivate yourself to keep trying to submit them considering that it may take hours to find a bug in a huge (comparing to other assignments) chunk of code.

Overall, I find this course very useful, but comparing to Algorithm Toolbox the self-study section really suffered and two MAJOR topics were given to one not so good (in my opinion) lecturer. As the result I struggled not because of the course complexity, but because of inconviniences.

And also grading system acts strangely. In python3 graders don't usually accept recursion because of RecursionDepthLimit error, however, all stress tests on my computer were completed successfully, but I still had to rewrite everything in loops. I guess it is usefull to know how to implement an algorithm both in loops and recursion, but I'd prefer if they specified the appropriate method for the problem beforehand.

one of the best course i have ever taken on any platform.

i love to learn on coursera platform.

coursera makes one to think fro solution.

after completion of course one feel satisfied.

The overall course was good. The instructor Daniel Kane, is the worst part of this course. He was rushing the whole time and wasn't explaining properly. No proper examples, no pseudocode, It was really annoying to complete all his lectures. In the end, I had to drop his lectures and find another source to study splay trees(which just kills the motive of taking this course) and solve the questions. Extremely unsatisfied and annoyed from the last part of this course. All the other parts were really great and all deserve a 5-star rating.

I liked this course alot.

If you are a student and come from the previous course, you may only read the cons, since this course has the same spirit as the previous.

Pros:

1. The lectures contain quite a good material which is somehow difficult and they made me to pay attention. The lectures are based on dasgupta's book and MIT course book.

2. There are links to additional materials, I found the dijkstra's book and MIT course book, so I was able to gain extra information for topics which interested me.

Also slides are very useful

3. The forum! This course has a life forum where you can find help or share your ideas.

4. The teaching stuff! They are answering student questions and taking part in discussions

Cons:

Having compared this course with the previous one from the set, this course suffers from luck of interesting problems. The previous course has more than ~25 problems and for each module it has advanced problems,even more they added extra problems during the course running.

This course has ~12 problems and only one advanced for the whole course!. Only this advanced problem made me take a piece of paper and a pen and draw trees, and play with toy examples. Only for this problem I wrote a stress test. That is the most fun for me of studying!

If this course didn't have this advanced problem, I would barely give it 3 stars.

P.S.

Theaching stuff, please conider to add extra problems, the first course is awesome and it is way too good. This course is good, but it think you may develop it not only the first one. Otherwise students may get dissapointed if they come from the previous course.

I hoped that you would have added extra problems, so I slowed down, my expectation didnt come true :(

The quality of the material and the level of challenges proposed in this series organized by the University of California San Diego are really surprising. In this module, each week details different data structures and the computational complexity of its operations. Then the concepts are verified in several algorithmic challenges, nothing trivial, automatically evaluated on its own platform through a battery of black box tests.

Just a few striking points:

* Very interesting to understand amortized analysis in dynamic arrays or hash tables and, even with the restructuring, how it remains weightless, considering that the additional cost is diluted in consecutive operations.

* Hash families and the guarantee of low collision, use of hashing in textual search (Rabin-Karp) and distributed hash table applications in cloud storage (Dropbox) and Big Data.

* Manipulation of large strings using Rope structures based on Splay Tree and how this is applicable in text editors.

* Take stack overflow even in delete/destroy functions, due to the high amount of data, and realize that all recursion can be restructured in terms of a Stack.

Once again I give five stars and strongly recommend.

This is an amazing course with great programming assignment - I have really enjoyed it, thank you very much!

Good Course for Knowledge

very very good

Great course

Excellent intermediate course with many challenging enough problems and questions. For brushing up the DS and for preparation for coding interviews - thumbs up!

It was difficult in using of pseudocode in my code

Needs more description & more practise problem

I think the course content and assignments were great. A suggestion though, it will be more helpful if there are more and varied corner cases that would save time spent in thinking and making cases.

The best data structures course that I have taken!

The complex topics are made simpler at the expense of teaching style that allowed me to make it applicable in a real world situations.

Lots of fun problems to solve. To get the most of this course, solve every single problem. Good content, not enough sample problem on tree.

Good explanation and problems to practice. Assignments will drive you to complete them, rather than just theory.

Excellent course. Great coverage of data structures. Good practice questions and explanatory videos.