I’ve been an Emacs user for many years, and a coder for just as long. Somewhere in the middle, something weird happened: while a text editor was initially considered a perfectly appropriate way of writing code (in C, Lisp, Perl, Shell, …), that ceased to be true for a while. With the gain in popularity of complex languages like Java, or C++, coders started to crave “fancy” features like code completion, refactoring, …

And quite unfortunately, these advanced features relied to complex tooling living outside of the languages themselves. Worse, the languages were developed in complete ignorance of these tools, which made them somewhat hostile to those goals (anybody who tried to implement a correct C++ parser knows what I mean, life before clang was just painful). As a result, very complex tools making heavy use of specialized partial parsers, static analysis, and crazy heuristics emerged. And they emerged as part of even more complex development suites to combine them all in a coherent form: the IDE was born.

To tell the truth, I’ve never been a fan of the concept, and even though I contributed to KDevelop for a while, I did so from the comfort of my beloved Emacs… Still, it’d be dishonest to pretend IDEs like Visual Studio, Eclipse and the likes don’t have a great impact, even today, on how people ultimately develop code.

I guess my main gripe with this abundant tooling is that it’s an all-or-nothing deal: there’s just no (easy) way I can take some of that experience and integrate it somewhere else. At a more fundamental level, it seemed simply “odd”: why would we introduce “deficiencies” at the language level, just to hope somebody else would work around them the best they can in tools? For those reasons, I never actually used an IDE. But then, I also stayed away from the languages that were “requiring” them as much as possible (with mixed results, as I ended up writing a ton of code in those as well :)). Not everybody has that luxury though, so text editors declined.

But more recently, things took a different turn (for the best I think): a new language emerged that was promoting a different paradigm: Go. Instead of making the tooling an afterthought, it’s been pretty much there at some level since inception. It even shows in the language grammar itself, which is designed to enable fast compilation, partial parsing, and a whole bunch of analysis tools. So instead of designing a language to give programmers arbitrary bells and whistles they’re used to (completely disregarding the resulting complexity in the larger ecosystem), the Go creators have made careful choices to create the best overall experience, which includes not only code writing, but also code reading, code maintaining and so on. So while some “features” that people might expect are noticeably missing (macros, generics, …), it’s not just because those guys are lazy (although all great engineers are :)) but also because the balance to maintain is very subtle.

I personally find the result quite brilliant (some people obviously disagree), and it’s a real pleasure to write code in Go, but more importantly to me, I can take the tooling, and integrate it pretty much anywhere. So a complex IDE is not even particularly useful: all you need from your editor is to be able to display the information the tools give, and to act on it.

And then, just like that, Emacs is a perfect environment to develop in Go. It’s not exactly free of course, in the sense that some integration code has to be developed. But it’s definitely several orders of magnitude easier than getting a proper environment for Java or C++, and even Python, despite Elpy’s greatness.

If anybody’s curious, here’s my Go configuration for Emacs. Except for some GOPATH-related gymnastics, it’s pretty straighforward.

I believe this new ability to embrace whatever environment the developer is most familiar with, and with minimal effort, is part of Go’s appeal. I, for one, am most excited about the prospect of many developer-friendly tools to come, and I really hope Go is just the first of many languages to follow the same approach.

comments powered by Disqus