This is a 2D puzzle game where you drag-and-drop components and connect them together. I liked it enough to play through the entire game, but the game has many problems that drag it down.
THE GOOD:
- The pipe-and-filter style puzzles are a fresh take on the puzzle genre. They are generally fun, and it can be quite challenging to achieve a gold medal on each puzzle.
- I like cats, catsThis is a 2D puzzle game where you drag-and-drop components and connect them together. I liked it enough to play through the entire game, but the game has many problems that drag it down.
THE GOOD:
- The pipe-and-filter style puzzles are a fresh take on the puzzle genre. They are generally fun, and it can be quite challenging to achieve a gold medal on each puzzle.
- I like cats, cats are a major theme of the game, they are cute
- Cute ending!
THE BAD:
- There is one song on loop for the whole game
- The game's story is cute, but irrelevant. You get rewarded with tidbits of advancement as you solve puzzles, but the puzzles have nothing to do with your goal.
- The game promises coding and machine learning, but delivers neither. What it *does* give you is pipes and filters. You can look up "pipe and filter architecture" if you don't know what I mean.
- The supposed machine learning is not even simplified machine learning. For example, the machine learning components get trained by magic without needing to evaluate their outputs. It makes training them boring and pointless.
- The text introductions to each puzzle are hard to understand *and* pointless. They very rarely have *anything* to do with the puzzle.
- The English tutorials and explanations are *very* hard to understand. Since the puzzles have nothing to do with machine learning, there is no way to figure out how machine learning works through practical experience.
- Startups are very poorly done. They are badly explained and provide very little useful feedback on what you've done right or wrong. The potential solution is so dumbed down from the text description of the problem that it's impossible to intuit solutions. See spoilers below for a bigger explanation.
- The driving simulators are also extremely confusing. See spoilers below.
- Doing puzzles can get repetitive since the pipes and filters are very consistent from one puzzle to the next. I usually didn't play for more than an hour or two at a time.
- You can save your setup in the level, but there is no corresponding load button. It's not clear that you can only load a saved configuration in future levels as a sub-component.
- This game seems to be marketing itself as a STEM game, encouraging children to do science, but really it does nothing to teach machine learning or programming, and so fails utterly in this.
THE UGLY:
- Some of the puzzle introductions cover military or spying applications that seem immoral, which is an odd thing to see in an otherwise very sweet and cute game.
MINOR SPOILERS:
- It took me a *long* time to understand the startups! I didn't even understand what the goal was. I believe you have two main goals: (1) To consume *all* of the inputs, and fill in *all* of the outputs, and (2) prevent any traffic jams, and thus allow a constant stream of input to flow through your solution. Hint: avoid using any load balancers as they cost money and make the startup unprofitable. After releasing your solution, if you've done things right, within a few days the startup should be making money! You've finished! The only number that matters is your daily profit. Once that goes into the negative you sell, and usually make a modest profit.
- The driving simulations don't make a lot of sense. In the early ones you just succeed automatically. In the later ones the only thing you can do is fiddle with the settings, if you pick the wrong settings you fail. Just play with the initial settings until you succeed.… Expand
while True: learn()Your Score0 out of 10User Score
