Hacker News Reader

Comment by smt88 | Hacker News Reader

This isn't too hard in some languages and I prefer it, but there are other languages (like C#) where class names are long and descriptive (e.g. ImmutableSortedDictionary) and it becomes harder to read instead of easier.

View on HN

C# offers a way to mitigate that, somewhat.

I often have something like this among the top level `using` statements.

    using FooMemo = System.Collections.Immutable.ImmutableSortedDictionary<string, Foo>;

You can do similar in C++. The problem is when the number of long types and variable names grows, you end up with classes and functions full of `using`, which gets confusing fast.

Also `var` is such an important mitigatory for this in C# that it continues to astound me that some of the worst offenders also eschew `var` in C#. You really don't need the redundancy of `ImmutableSortedDictionary<string, Foo> localCache = new ImmutableSortedDictionary<string, Foo>()`, it's not some sort of Prolog assertion where the strict tautology makes different code than `var localCache = new ImmutableSortedDictionary<string, Foo>()`.

The problems start when you open code to see

var localcache = cachemaster.cacheFor(foo);

Now you have no idea what it is until you dig into the (hopefully correct) API docs, or use an IDE that can untangle things.

I don't buy it. It's one jump to method definition away, which even an editor without IDE capability should be able to manage. But also, this is what dynamic languages look like, and people still manage to be productive.

(I'll agree it can be a bit tricky with LINQ and anonymous types in C#, but I think `var` usage is worth that trade-off)

^ Your line of code is 85 characters though.. Would you bother to split this line somewhere? (if trying to conform to an 80char limit)

Reply on news.ycombinator.com