Rephrase

kgryte · kgryte · commit 0362a2a8756d · 2017-02-02T21:47:01.000-08:00
diff --git a/FAQ.md b/FAQ.md
@@ -60,7 +60,7 @@ Native [add-ons][node-add-ons] have several disadvantages:
 
 1. __Complexity__: compilation presupposes the existence of compilers (e.g., [gfortran][gfortran]) and other tooling in order to successfully compile, thus often requiring out-of-band installation, setup, and configuration. In short, compilation increases complexity and increases the risk that something can and will go wrong.
 
-1. __Development__: native Node.js [add-ons][node-add-ons] require significantly more upfront development costs compared to porting implementations to JavaScript. Creating a native [add-on][node-add-ons] entails more than writing a simple wrapper around an existing C/C++ library; the process involves additional tooling, testing, and development procedures, all requiring time and effort. These costs are acutely apparent during iteration cycles targeting multiple platforms. In comparison, as a higher-level language, JavaScript facilitates faster development, has built-in portability, and has minimized performance costs.
+1. __Development__: native Node.js [add-ons][node-add-ons] require significant upfront development costs compared to porting implementations to JavaScript. Creating a native [add-on][node-add-ons] entails more than writing a simple wrapper around an existing C/C++ library; the process involves additional tooling, testing, and development procedures, all requiring time and effort. These costs are acutely apparent during iteration cycles targeting multiple platforms. In comparison, as a higher-level language, JavaScript facilitates faster development, has built-in portability, and has minimized performance costs.
 
 1. __Bundling__: existing low-level C/C++ and Fortran numeric computing libraries are often large monoliths. Hence, writing a Node.js native [add-on][node-add-ons] which exposes a single function from one of these libraries requires linking to a much larger codebase, resulting in larger bundle sizes, increased network costs, and longer installation times. And while many systems may come with pre-installed numeric computing libraries (e.g., Apple Accelerate Framework), their existence is neither guaranteed nor is detecting their existence trivial.
 
