diff options
Diffstat (limited to 'doc/complement-lang-faq.md')
| -rw-r--r-- | doc/complement-lang-faq.md | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/doc/complement-lang-faq.md b/doc/complement-lang-faq.md index 6a96a7728de..3572a411e0a 100644 --- a/doc/complement-lang-faq.md +++ b/doc/complement-lang-faq.md @@ -236,7 +236,7 @@ They start small (ideally in the hundreds of bytes) and expand dynamically by ca * Requiring explicit bounds means that the compiler can type-check the code at the point where the type-parametric item is *defined*, rather than delaying to when its type parameters are instantiated. You know that *any* set of type parameters fulfilling the bounds listed in the API will compile. It's an enforced minimal level of documentation, and results in very clean error messages. -* Scoping of methods is also a problem. C++ needs [Koenig (argument dependent) lookup](http://en.wikipedia.org/wiki/Argument-dependent_name_lookup), which comes with its own host of problems. Explicit bounds avoid this issue: traits are explicitly imported and then used as bounds on type parameters, so there is a clear mapping from the method to its implementation (via the trait and the instantiated type). +* Scoping of methods is also a problem. C++ needs [Koenig (argument dependent) lookup](http://en.wikipedia.org/wiki/Argument-dependent_name_lookup), which comes with its own host of problems. Explicit bounds avoid this issue: traits are explicitly imported and then used as bounds on type parameters, so there is a clear mapping from the method to its implementation (via the trait and the instantiated type). * Related to the above point: since a parameter explicitly names its trait bounds, a single type is able to implement traits whose sets of method names overlap, cleanly and unambiguously. |