Profile

Dean Schulze profile image

Dean Schulze

Student

  • 2 reviews
  • 2 completed
Content 
Instructor 
Provider 
I'm reposting my original review of this course since my review and all others were removed. I'm currently half way through the Udacity Android course and I recommend the Udacity course over this one. This course suffers from a fundamentally flawed approach to teaching Android development. It tries to teach Android development through lectures instead of having students write Android applications. While the production quality of these lectures is high, they are nearly useless for learning Android development. Learning to write Android applications is mostly learning the Android API. It is a rich, complex API and just like learning any other API such as JPA, JSF, or Swing you have to write code to learn it. While this course does have programming assignments, Coursera defeats the learning process by distributing Android applications that are more than 90% complete with a few TO DO blocks in them for students fill in with a few lines of code. This approach is useless. What is needed is to have students write entire Android applications. In a MOOC applications from thousands of students could only be graded by an autograder, and this I believe is why Coursera chose to over-emphasize lectures instead. Coursera would not spend the money needed to develop an Android autograder. The Princeton courses on algorithms by professors Sedgewick and Wayne have a very impressive autograder that could serve as a model for Coursera's Android courses. The professors specify an API that the application has to implement and the autograder gives detailed feedback on the implementation. There will be some challenges for an autograder for Android applications. It may not be possible to autograde an app that plays an audio or video file. Someone has to watch and listen to see if it is played correctly. Peer grading would be needed in these cases. Coursera is responsible for this failed approach. During the second course in this specialization (taught by Professor Douglas Schmidt) some of us were lamenting the lack of an autograder. Professor Schmidt responded that there was no way to develop an autograder since they had no budget for the two courses. This course alone provided over $400,000 in revenue to Coursera, but they didn't invest any of it back into the course. Coursera ate all of the seed corn. Professor Porter probably did the best he could with what he had to work with. It is up to Coursera to provide the resources (money) to create the technology needed to teach this course effectively. Startups are supposed to create new technology, not repackage centuries-old techniques and deploy them on the web. Get with it, Coursera.
Content 
Instructor 
Provider 
This course suffers from a fundamentally flawed approach to teaching Android development. It tries to teach Android development by having students listen to lectures and take quizzes instead of writing Android applications. While the production quality of these lectures is high, they are nearly useless for learning Android development. Learning to write Android applications is mostly learning the Android API. It is a rich, complex API and just like learning any other API such as JPA, JSF, or Swing you have to write code to learn it. While this course does have programming assignments, Coursera defeats the learning process by distributing Android applications that are more than 90% complete with a few TO DO blocks in them for students fill in with a few lines of code. The few lines of code in the TO DO blocks are then peer graded. This approach is useless. What is needed is to have students write entire Android applications. In a MOOC applications from thousands of students could only be graded by an autograder, but Coursera would not spend the money needed to develop an Android autograder. Hence the over-emphasis on lectures and quizzes which are much easier to do, but ineffective. The Princeton courses on algorithms by professors Sedgewick and Wayne have a very impressive autograder that could serve as a model for Coursera's Android courses. The professors specify an API that the application has to implement and the autograder gives detailed feedback on the implementation. There will be some challenges for an autograder for Android applications. It may not be possible to autograde an app that plays an audio or video file. Someone would have to watch and listen to see if it is played back correctly. Peer grading would probably be needed in these cases. Coursera is responsible for this failed approach. During the second course in this specialization (taught by Professor Douglas Schmidt) some of us were lamenting the lack of an autograder. Professor Schmidt responded that there was no way to develop an autograder since they had no budget for the two courses. This course alone provided over $400,000 in revenue to Coursera, but they didn't invest any of it back into the course. Coursera ate all of the seed corn. Professor Porter probably did the best he could with what he had to work with. It is up to Coursera to provide the resources (money) to create the technology needed to teach this course effectively. Startups are supposed to create new technology, not repackage centuries-old techniques and deploy them on the web. Get with it, Coursera.