If you don't have access to the buttons "Add a new column" and "Edit" (in the drop-down list, it can only be explained by access rights. You should retry when logged as an administrator to troubleshoot if the problem is an access right configuration. You can also test this functionality on http://runbot.openerp.com because it works there.

Hey man! I ran into the same issue.. there seem to be a couple of things at play here. 

If you look at how Notes and the CRM lead is designed youw ill notice 2 ways of grouping the "Columns"

one is by overriding the read_group method as seen in Notes and the other by using

_group_by_full = {

'grouping name': _read_group_stage_ids

}

Now, OpenERP magically relies on the definition of these two separate ways of gathering the needed data for kanban views.

Explicitly what you need to have is the field you want to group by exactly match the model field in the xml view def:

<record model="ir.ui.view" id="view_sodancasync_files_kanban">
      <field name="name">sodancasync.files.kanban</field>
      <field name="model">sodancasync.files</field>
      <field name="arch" type="xml">
        <kanban default_group_by="folder" >*******this is where you specify what to group by******* this relies on the mentioned above.. if there are discrepancies the "Add new column" will not show up nor will the header names.

         
          <field name="name"/>
          <field name="folder"/>
          <templates>
            <t t-name="kanban-box">

              
            </t>
          </templates>
        </kanban>
      </field>
    </record>

Furthermore the data returned by read_group or _read_group_stage_ids is very specific. I cant say there were any direct instructions nor charts to follow, but i was able to accomplish what you seek by reverse engineering Notes and CRM Opportunities aka crm_lead. Good luck!