suggestion - consider more detail in level5 problem description?

[tomhood]

As a cascading newbie, I went through each level and learned a lot. Thank you.

For level5, I did get some differences that are probably okay to ignore. Many look like rounding errors (e.g. you had 0.16507055512144617 for "primary" while my result was 0.1650705551214462). Other differences look like the logic for extracting & normalizing words is a little different (e.g. I have the word "2" in the "intro" document and the letter "c" in the "filesystem" document). My only suggestion is to consider changing the problem description to specify how many digits of precision to use as well as more detail on what is expected for extracting & normalizing words.

Thanks again for the great tutorial. It was very helpful.

[BertrandDechoux]

I would say the root cause is the (absent) definition of what is considered a word. If the list of words is not strictly the same then the "numbers" won't be the same.

I supposed the definition was not important in itself. I still doubt the specific logic to implement is important. The main thing is to understand how to split.

However, your comment on the rounding is spot on. The numbers should have a fixed number of digits of precision. That should solve most problems and it will make reading the results easier.

I will probably redo the exercises myself with the release of cascading 2.2 (I didn't keep my previous answers) and fix all open issues at that time.

Thanks for the feedback!