The source content for blog.juliobiason.me
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

2.2 KiB

+++ title = "Hands-On Microservices with Rust - Denis Kolodin" date = 2019-03-03

[taxonomies] tags = ["reviews", "books", "rust", "microservices", "denis kolodin", "it"] +++

GoodReads Summary: A comprehensive guide in developing and deploying high performance microservices with Rust.

{{ stars(stars=2) }}

The shortest and simpler way to describe this book is this:

let mut a = 1;
a = a + 1;
// increase a

Sure, increasing a may seem interesting, but where the heck is the explanation on why you're increasing a?

The long version is this: this is an exact representation of what a Packt book is. There are a bunch of grammatical errors from the half of the book to the end, which should've picked by the editor; there are a bunch of weird explanations and the incredible tendency of use a different package on each project (one project uses Hyper, the other Rocket, the other Actix-Web) which should've been picked by the reviewer; the whole book have a bunch of what the code is doing, but not why it was designed that way.

So the whole thing feels rushed and without thought.

The "change package on each discussion" is also damming: you never really got deep into a package, understanding its ins and outs, because it keeps changing all the time (logs is another thing that keeps changing all the time in the code, which seems really weird when you want to have a fleet of microservices -- why would you use different packages on all of them; yes, you can do this due the nature of microservices, but does it make sense?).

Also, no code is shown in completion. The book uses snippets all around instead of showing everything. The full code is available on Github, but that means you need to keep two sources open at the same time: one in the book, to follow whatever the author is showing and the source code to understand where each thing fall in the big picture.

Those two last points could be easily solved by starting with a simple code (say, the Hyper that says "Hello world") and then, slowly, refactor it to reach a fully asynchronous code, showing the full code with each refactor highlighted.

Maybe this will be a good book in the 3rd edition.