@@ -45,11 +45,6 @@ page. Below is a snapshot of the diagnostic page when dask is calculating an ima
 .. image:: partial.png
       :width: 1024px
 
-Below we demonstrate the utility of this view. This shows ICAL for 7 visibility ingests. Note the global
-synchronisation point (at 1:40) where the selfcalibration across the full bandwidth occurs.
-
-.. image:: dask-global.png
-      :width: 1024px
 
 Using ARL and dask on Darwin
 ****************************
 
@@ -0,0 +1,105 @@
+.. toctree::
+   :maxdepth: 2
+
+ARL directory
+*************
+
+A simple directory to help those familiar with other calibration and imaging packages navigate the Algorithm Reference
+Library.
+
+See also the :ref:`genindex` and :ref:`modindex`
+
+
+Data containers used by ARL
+===========================
+
+See :py:mod:`arl.data.data_models.py` for the following definitions:`
+
+* Telescope Configuration: :py:class:`arl.data.data_models.Configuration`
+* GainTable for gain solutions (as (e.g. output from solve_gaintable): :py:class:`arl.data.data_models.GainTable`
+* Image (data and WCS header): :py:class:`arl.data.data_models.Image`
+* Skycomponent ((data for a point source or a Gaussian source): :py:class:`arl.data.data_models.Skycomponent`
+* Baseline based visibility tables shape (npol,), length nvis) :py:class:`arl.data.data_models.Visibility`
+* Antenna-based visibility table, shape (nants, nants, nchan, npol), length ntimes): :py:class:`arl.data.data_models.BlockVisibility`
+
+Create empty data set for observation (a-la makeMS)
+===================================================
+
+* For Visibility operations: :py:mod:`arl.visibility.base.create_visibility`
+* For BlockVisibility: :py:mod:`arl.visibility.base.create_blockvisibility`
+
+Gridding, degridding and weighting
+==================================
+
+* Convolutional gridding: :py:mod:`arl.fourier_transforms.convolutional_gridding.convolutional_grid`
+* Convolutional degridding: :py:mod:`arl.fourier_transforms.convolutional_gridding.convolutional_degrid`
+* Weighting: :py:mod:`arl.fourier_transforms.convolutional_gridding.weight_gridding`
+* Tapering: TBD
+
+Visibility Predict
+==================
+
+* Predict BlockVisibility for Skycomponent :py:mod:`arl.imaging.base.predict_skycomponent_blockvisibility`
+* Predict Visibility for Skycomponent: :py:mod:`arl.imaging.base.predict_skycomponent_visibility`
+* Predict by de-gridding visibilities:
+
+ * :py:mod:`arl.imaging.base.predict_2d_base`
+ * :py:mod:`arl.imaging.base.predict_2d`
+ * :py:mod:`arl.imaging.base.predict_wstack`
+ * :py:mod:`arl.imaging.base.predict_wprojection`
+ * :py:mod:`arl.imaging.base.predict_timeslice`
+ * :py:mod:`arl.imaging.base.predict_facets`
+ * :py:mod:`arl.imaging.base.predict_facets_wprojection`
+ * :py:mod:`arl.imaging.base.predict_facets_wstack`
+ * :py:mod:`arl.imaging.base.predict_wprojection_wstack`
+
+
+Visibility Invert
+=================
+
+* Generate gridding kernels (standard prolate spheroidal or W-projection): :py:mod:`arl.imaging.params.get_kernel_list`
+* Invert by gridding visibilities: 
+
+* Predict by de-gridding visibilities:
+
+ * :py:mod:`arl.imaging.base.invert_2d_base`
+ * :py:mod:`arl.imaging.base.invert_2d`
+ * :py:mod:`arl.imaging.base.invert_wstack`
+ * :py:mod:`arl.imaging.base.invert_wprojection`
+ * :py:mod:`arl.imaging.base.invert_timeslice`
+ * :py:mod:`arl.imaging.base.invert_facets`
+ * :py:mod:`arl.imaging.base.invert_facets_wprojection`
+ * :py:mod:`arl.imaging.base.invert_facets_wstack`
+ * :py:mod:`arl.imaging.base.invert_wprojection_wstack`
+
+Deconvolution
+=============
+
+* Hogbom Clean: :py:mod:`arl.image.cleaners.hogbom`
+* Multi-scale Clean: :py:mod:`arl.image.cleaners.msclean`
+* Multi-scale multi-frequency Clean: :py:mod:`arl.image.cleaners.msmfsclean`
+* Entry point for above: :py:mod:`arl.image.deconvolution.deconvolve_cube`
+
+Calibration
+===========
+
+* Create empty gain table: :py:mod:`arl.calibration.operations.create_gaintable_from_blockvisibility`
+* Solve for complex gains: :py:mod:`arl.calibration.solvers.solve_gaintable`
+* Apply complex gains: :py:mod:`arl.calibration.operations.apply_gaintable`
+* Peel compact sources: :py:mod:`arl.calibration.peeling.peel_skycomponent_blockvisibility`
+
+Coordinate transforms
+=====================
+
+* Station/baseline (XYZ <-> UVW): :py:mod:`arl.util.coordinate_support`
+* Source (spherical -> tangent plane): :py:mod:`arl.util.coordinate_support`
+* Phase rotation: :py:mod:`arl.visibility.base.phaserotate_visibility`
+* Polarisation basis conversion (Linear <-> Stokes <-> Circular): :py:mod:`arl.data.polarisation`
+
+Graphs:
+=======
+
+* Perform various types of prediction and inversion of visibility data: :py:mod:`arl.graphs.graphs`
+* Perform generic image or visibility unary operations: :py:mod:`arl.graphs.generic_graphs`
+* Support testing and simulations: :py:mod:`arl.util.graph_support`
+* The canonical pipelines:py:mod:`arl.pipelines.graphs`:
@@ -13,14 +13,14 @@ structures such as image, visibility table, gaintable, etc.
    :maxdepth: 3
 
    ARL_quickstart
+   ARL_directory
    ARL_definition
    ARL_background
    ARL_developing
    ARL_Fourier_processing
    ARL_calibration
    ARL_dask
    ARL_references
-   ARL_glossary
 
 Indices and tables
 ==================
 
@@ -173,7 +173,6 @@ def test_predict_facet_wstack_graph(self):
         predicted_vis_graph_list = create_predict_facet_wstack_graph(zero_vis_graph_list, flux_model_graph,
                                                                      facets=2, vis_slices=self.vis_slices)
         residual_vis_graph_list = create_subtract_vis_graph_list(self.vis_graph_list, predicted_vis_graph_list)
-        self.compute=True
         if self.compute:
             qa = qa_visibility(self.vis_graph_list[0].compute())
             numpy.testing.assert_almost_equal(qa.data['maxabs'], 1600.0, 0)