If the tests are grouped according to what class / method / function
they're testing, then it's a lot easier to see what is being tested and
what isn't, and it's easier to find the tests for `foo()` or to know where
to put new tests for `foo()`.

On the other hand when you names tests like
`describe('channel event handlers',` `describe('event handlers',`
`describe('the "loadAnnotations" event',`,
`describe('the "deleteAnnotation" event',` etc then it can be a lot
harder for a reader coming along later to understand how the tests are
organized and how this organization relates to the code itself.

Most of the examples on https://mochajs.org/ use this style of
organizing tests:

    describe('Array', function() {
      describe('#indexOf()', function() {
        it(...
        ...
      });

      describe('#concat()', function () {
        it(...
        ...
      });

      describe('#slice()', function () {
        it(...
        ...
      });

    describe('User', function() {
      describe('#save()', function() {
        it(...

    describe('Connection', function() {
      describe('#find()', function() {
        it(...

This commit reorganizes the AnnotationSync tests along the same lines.

I've also made use of Mocha's `context()` function to group tests
**within a `describe()` for a method** by context. `context()` is just
an alias for `describe()`, but it has different semantics. (This is
also in line with how `context()` is used in the mochajs.org examples).