Hacker News Reader

Comment by vardanator | Hacker News Reader

Interesting notes, thanks. I was struggling with such books that compress a complex idea in a couple of sentences, so I'm always trying to explain as detailed as possible. Though understanding that there is a broad audience who likes the opposite (succinct explanations), I wrote disclaimers to secure me just in case. So, read the disclaimers one more time, that's my only excuse for such long and boring discussions :)

View on HN

I share some of the criticisms already stated in other comments. The biggest offender here in my opinion is that it jumps to too many subjects. Trees as graphs, yes, but they are a whole subject of their own. This post should broken down in at least 4 or 5 posts. Moreover, its mixes vastly different concerns: byte level optimization is an interesting topic too, but it has nothing to do with graph per se.

On the positive side: relevant illustrations. I also liked that the code implementation is one that use explicit edges. Too often in books, it is either the adjacency matrix or the linked list ones. The former is inconvenient when you need to add a node, and the latter doesn't let you put weight on edge. So these "academic" implementations are not well suited for groking concept by playing with code.

I work a lot with graphs and there is a big value in having a general enough graph implementation. Last week I implemented a trie class with it, and this week a transducer. Both in two hours, debugging included. Other things such as serialization came from free from the library. Of course, that not optimal data structure, but it is fine for my use cases.

In your post, you don't use the fact that trees are graph to avoid duplicate code explanations, instead you explain it first for a tree, then details the graph one. It should be the reverse: how to use a graph implementation to create a tree. And then only if this implementation does not fit the memory or performance requirements of the app, rewrite a more specialized data structure.

Thanks!

Detailed explanations are good. As is a conversational style. But consider that all jokes and asides that are meant to be conversational can get in the way of a beginner understanding your point, because they are beside the point.

I recommend reading blog posts by authors you like who have a conversational style with an editors eye. How often do sentences not have any relevant information? How often do they actually try for a joke? It's often less than you might think. The style often comes more from tone and word choice. I think a great model is Julia Evans: https://jvns.ca/

Thank you!

Yes, I have talked to people who much prefer a lighter, lengthier style, so I guess this is a matter of taste. If I may offer a suggestion though, the pictures are very distracting and often don't add to the discussion. Consider either limiting their use, or making them smaller and on the sides of the page.

Sure, will try for the next one. Thanks.

Have you heard of the Pyramid Principle book? I am working through the book now. I find the method very helpful in structuring how you present ideas in your writing.

Thanks for the heads-up on the Pyramid Principle. I've been studying storytelling and the outlined principles jive pretty closely with what I've picked up.

This seem to be a great summary of the Pyramid Principle [0] -- I hear the book is a bit of a slog after the first 3 chapters.

[0] https://medium.com/lessons-from-mckinsey/the-pyramid-princip...

If you consider the book applies the principle itself, it is well worth the read.

One tip I would suggest if you plan to buy the book, get the older version from the 90s at a lower price. The newest version is quite pricey.

Reply on news.ycombinator.com