Add aliases for encodings

[tfausak]

Fixes #103.

Still to do:

• Add documentation. Some basic usage examples would be nice.
• Bike shed naming. How should lazy and strict be differentiated? What about little endian and big endian?
• Consider deprecating ISO_8859_1 in favor of LATIN_1, which is both shorter and what the text package calls it anyway.

[tfausak]

How should lazy and strict be differentiated?

Options:

• As it is now, with *Strict and *Lazy suffixes.
• Make lazy variants have no suffix and strict variants have a *' suffix.
• Make strict variants have no suffix and lazy variants have *Lazy suffixes.
• As above, but with prefixes instead of suffixes.
• As above, but with abbreviations (S & L) instead.

[tfausak]

What about little endian and big endian?

Options:

• As it is now, with *le and *be suffixes.
• Same as above, but with uppercase suffixes instead.
• Pick a default to have no suffix and only apply the suffix to the other one.
• Provide an option without the suffix that attempts to figure out the correct endianness.

https://en.wikipedia.org/wiki/UTF-16#Byte-order_encoding_schemes

[tfausak]

It may be good to avoid using "strict" since it could mean either "not lazy" or "not lenient", both of which are applicable to encodings.

[tfausak]

Another alternative: Only provide aliases for strict byte strings.

type Utf8 = UTF_8 ByteString
via @Utf8 "..." :: ByteString
via @(UTF_8 LazyByteString) "..." :: LazyByteString

I have a hunch that conversions for strict byte strings are more common, but I don't know how to prove it. And if I later want to add aliases for lazy byte strings, I'll be right back here.