In defence of the poster, I think it depends on how much experience you have solving a certain class of problems, and perhaps more importantly, how you were taught.
I don't recall having studying the A* algorithm before, so I first looked it up on Wikipedia for a few minutes, and then came back to the comments. It's worth me saying that 30 seconds of reading your comment and linked gist did more to aid my comprehension than a few minutes looking through the Wikipedia article.
This is the case because you described the algorithm in terms of building blocks I was already familiar with. It's a depth-first search with a weighted queue; as you say, pretty obvious! But it wasn't immediately obvious from the Wikipedia article, because Wikipedia articles are designed to impart information, not necessarily to teach.
If the poster was taught by a professor who was teaching more like the Wikipedia article (and I've known a lot of lecturers who do this!), I can understand the confusion.
I see where you are coming from. I was arguing more on the side of "familiarity with algorithms" is like "hitting the gym" irrespective of what game/sport you play. It either helps you directly, or gives you general skill.
I was mostly taking exception to conflating "great programmers - no algorithmic knowledge".
It's great that you figured out A* search in under 30 seconds, but that's because you had the foundation. As I said, I am not claiming A* search(or any other algorithm) is the baseline, but there is a baseline(I won't bother enumerating algorithms - there is no agreed upon list), and a good programmer should strive to be above that baseline. Simply putting your hands up and claiming I can develop large systems fine and hence algorithmic knowledge is not really helpful - it is counter productive.
Again, if it's working fine for the poster, that's great. But the poster might be missing "unknown unknowns."
On a side note, Wikipedia is a horrible source to learn a new algorithm.
I can't say for A*, but many other algorithms (like Bayes Probability, FFT) I tried look up made little sense to me on Wikipedia, but learning from some well presented material(be it text, graphics or video) made them easy to comprehend.
I wrote this about 10 years ago and hope to spend some time one day making it simpler, but it seems to have helped a lot of people judging by the emails I get... http://www.heyes-jones.com/astar.html
Not as concise as the explanation given by irahul (which actually made me get the concept), but your article is very interesting and well written, and it is absolutely worth sharing to a couple friends of mine who would probably not get it with the wikipedia article.
I Wikipedia has helped me understand many algorithms. It's probably something like 50% chance I will get it, but the articles so short it's worth a first look.
I agree, but I think there are also multiple foundations, and missing out on one or two doesn't necessarily result in a structural collapse into mediocracy.
It's quite easy to end up with a bad teacher and miss the foundation necessary to understand a class of algorithms. In an ideal world you'd backtrack until you understood that missing foundation, but it's easy to get the impression that a particular area of study is beyond your intellect.
I kinda think that's what's happened here. The poster probably has enough foundations to work with normally, but here and there are missing pillars of knowledge.
That said, search is a pretty huge and important class of algorithms! The fact that he had difficulty understanding A* suggests that a rather large pillar is missing from the foundations of his craft.
Something I have found is that with learning algorithms a sideways approach is sometimes better if you find yourself struggling.
I initially had some trouble visualizing the process of algorithm development if I simply followed a book and did exercises (self-taught), so I would pick a complicated looking algorithm and just read about it, then think about it for a while, and try to find somewhere it had been used in code and how it solved some problem.
After a while I would go back to the book and it would sink in and I was able to better understand the mathematical model and intuition. Several such algorithms later and I am able to pick up new algorithms relatively quicker.
I don't recall having studying the A* algorithm before, so I first looked it up on Wikipedia for a few minutes, and then came back to the comments. It's worth me saying that 30 seconds of reading your comment and linked gist did more to aid my comprehension than a few minutes looking through the Wikipedia article.
This is the case because you described the algorithm in terms of building blocks I was already familiar with. It's a depth-first search with a weighted queue; as you say, pretty obvious! But it wasn't immediately obvious from the Wikipedia article, because Wikipedia articles are designed to impart information, not necessarily to teach.
If the poster was taught by a professor who was teaching more like the Wikipedia article (and I've known a lot of lecturers who do this!), I can understand the confusion.