Now the storage will return the current length.
And not the set length anymore. But if there are datasets in the storge it
will be as long as the setted length.
It also can return the length of its datasets.
Now we cat add new data to the old storage, which returns a new storage
containg first the new data and then the old data.
Plus a test for this, because it's not implemented now.
This storage should contain the weighted data, before it it's used in the
output classes. So we can create a timeline.
Plus we have a nice place to store the weighted data.