I recently came across two interesting posts about the Phoenix framework: A benchmark comparing the performance against Rails and a how-to for creating a simple JSON API. I was interested by the performance characteristics described in both articles. What really got my attention though was the syntax: The feel for it was very ruby-like.
Probably one of the most well-known books among rubyists, “The Ruby Way” by Hal Fulton with André Arko, has now been updated and released in its third edition. The first part of the book is dedicated to the language itself and covers syntax, semantics, some comparison to other languages and specific issues, like garbage collection, that developers are well served to know when writing ruby.
I started using fuzzy matching when I switched to Sublime Text 2 a few years ago (I currently use Atom, which also has the same feature built-in). The seemingly little improvement increased my productively greatly. It saves a few moments while opening files, but more importantly it prevents context switching. It lets me start working with a file (usually prompted by knowing or wanting to know the contents of it) without needing to think about the location of that file.
Gustave Dutra at Plataformatec, discusses a concept called Discrete Integration. The process described is very similar to what I use at work at the moment. A sort of dialed-back version of Continuous Integration. It reaps a lot of the benefits of rapid feedback, just not as extreme. Feature branches are worked-on until they are at a stable state before merging into master. Good developer communications is encouraged at all stages.
Janko Marohnić presents a well-thought out argument against requiring all code in gems up-front. His concerns are around what this communicates to other developers. The intention of each file or library required is lost as are the dependencies of each individual class. Finally, it makes it harder to clean-up dependencies once the are no longer used. I intend to adopt this style in the next gem I work on. I have actually done a similar thing in some Rails project, with a smaller scope: For example, if a class uses an external service wrapper, I require it on the top of the class even though it’s already required by bundler, as a notice to other developers on the project.
I previously wrote about my thoughts on the Software Craftsmanship metaphor. My main objection to the metaphor is that it evokes imagery that is not helpful to software development and completely unnecessary. In this article Dan North’s tells us that the risk behind the craftsmanship movement lies in putting software ahead of the benefits it provides. My understanding is that so called ‘craftsmen’ pride themselves in delivering value first, but the article did provoke in my a deep sense that there is a lot of navel-gazing going on. Let’s just focus on professionalism and excellence. Keep calm and code on.
In his usual clear way Pat Shaughnessy writes about when to mark methods as private. Essentially, methods that have tests are part of what is considered the public API of the object. Untested methods and those that are private are not. Ruby performs the most basically check against calling private methods, that can be easily circumvented by using
send. However, marking methods as private is more about communicating to other developers what part is internal to the class and should not be relied upon. This is very much in vein with Sandi Metz take on testing
A few months ago, the RSpec team announced the new expectation syntax. In the simplest form, the old way of saying: