Totally agree on the need for tuple syntax (and pattern matching to make it really useful):
return Tuple.Create(foo,bar);
Should be:
return (foo,bar);
The worst part is using the Item1..ItemN properties. One way I've gotten around it is to create a Tuple extension method called Apply [1]. I allows you to do this:
ParseName(fullName).Apply((first, last) => ... )
That allows for giving names to the tuple ItemX fields and uses them 'in scope'.
At least with C#6 we can now open static classes like namespaces and do something this:
using Language.Helpers [2]
...
return tuple(x,y);
Which already makes it start to feel like it's part of the language.
Apply creates new scope, which is often undesirable. Consider the "Declaration expressions" feature in C# 6:
ParseName(fullName).Unpack(out string first, out string last);
However, this makes for essentially left-to-right assignment, which is contrary to the usual order. So this creates discontinuity when moving from one return value to two return values.
My personal opinion is 'out' is a total abomination. I will do everything I can to avoid using it. I prefer expression based programming where there's a balance between the two sides of an operator. Declaring named values on the RHS is just horrible (again, IMHO). The only reason for it to exist is because C# hasn't got proper tuple support.
By the way, creating a new variable half-way down a method is also creating a new scope (the variable isn't in scope above the declaration, and it is in scope below it, until the end of the method).
For example:
public int GetSurnameLength()
{
string fullName = ReadFullNameFromDb();
Tuple<string,string> result = ParseName(fullName);
return result.Item2.Length;
}
Has the same effect as:
public int GetSurnameLength()
{
string fullName = ReadFullNameFromDb();
return ParseName(fullName).Apply( (first,last) => last.Length );
}
The only real downside to the Apply method is the syntax clutter from the closure. But there's no real issue with scope as far as I can tell (because the rest of the method can be within the closure if necessary). This is very similar to how 'let' works in F#.
At least with C#6 we can now open static classes like namespaces and do something this:
Which already makes it start to feel like it's part of the language.[1] https://gist.github.com/louthy/f4cee2524d0d049b4378
[2] https://gist.github.com/louthy/dc06868c446ed76f0d7a