Column without a name produces an invalid serialization

```python
import asdf
import asdf.table
asdf.dumps({"c": astropy.table.Column(data=[1, 2])})
```
produces a ValidationError:
```
ValidationError: None is not of type 'string'

Failed validating 'type' in schema['properties']['name']:
{'description': 'The name of the column. Each name in a\n'
'[table](http://stsci.edu/schemas/asdf/table/table-1.2.0) '
'must be\n'
'unique.\n',
'pattern': '[A-Za-z_][A-Za-z0-9_]*',
'type': 'string'}

On instance['name']:
None
```

I propose that we introduce a new astropy column tag/schema and stop writing files with the core column tag/schema (but continue to read them).

Reproduce the failure with astropy.table.Column(data=[1, 2]) and asdf.dumps. Trace how the core column tag/schema handles a column whose name is None, then review the proposed astropy column tag/schema direction. Done means unnamed columns serialize validly while existing files using the core column tag/schema remain readable.

Written by the indexing model from the issue text.