update installation instruction (#3849)

* installing.rst: update instructions

* whats-new

* explicit link and anchor

* :doc: -> :ref:

doc/installing.rst

@@ -11,6 +11,8 @@ Required dependencies
 - `numpy <http://www.numpy.org/>`__ (1.15 or later)
 - `pandas <http://pandas.pydata.org/>`__ (0.25 or later)
 
+.. _optional-dependencies:
+
 Optional dependencies
 ---------------------
 
@@ -24,7 +26,7 @@ For netCDF and IO
 - `h5netcdf <https://github.com/shoyer/h5netcdf>`__: an alternative library for
   reading and writing netCDF4 files that does not use the netCDF-C libraries
 - `pynio <https://www.pyngl.ucar.edu/Nio.shtml>`__: for reading GRIB and other
-  geoscience specific file formats
+  geoscience specific file formats. Note that pynio is not available for Windows.
 - `zarr <http://zarr.readthedocs.io/>`__: for chunked, compressed, N-dimensional arrays.
 - `cftime <https://unidata.github.io/cftime>`__: recommended if you
   want to encode/decode datetimes for non-standard calendars or dates before
@@ -121,16 +123,15 @@ xarray itself is a pure Python package, but its dependencies are not. The
 easiest way to get everything installed is to use conda_. To install xarray
 with its recommended dependencies using the conda command line tool::
 
-    $ conda install xarray dask netCDF4 bottleneck
+    $ conda install -c conda-forge xarray dask netCDF4 bottleneck
 
 .. _conda: http://conda.io/
 
-We recommend using the community maintained `conda-forge <https://conda-forge.github.io/>`__ channel if you need difficult\-to\-build dependencies such as cartopy, pynio or PseudoNetCDF::
-
-    $ conda install -c conda-forge xarray cartopy pynio pseudonetcdf
+If you require other :ref:`optional-dependencies` add them to the line above.
 
-New releases may also appear in conda-forge before being updated in the default
-channel.
+We recommend using the community maintained `conda-forge <https://conda-forge.github.io/>`__ channel,
+as some of the dependencies are difficult to build. New releases may also appear in conda-forge before
+being updated in the default channel.
 
 If you don't use conda, be sure you have the required dependencies (numpy and
 pandas) installed first. Then, install xarray with pip::

doc/whats-new.rst

@@ -81,6 +81,9 @@ Documentation
 - Fix documentation of :py:class:`DataArray` removing the deprecated mention
   that when omitted, `dims` are inferred from a `coords`-dict. (:pull:`3821`)
   By `Sander van Rijn <https://github.com/sjvrijn>`_.
+- Update the installation instructions: only explicitly list recommended dependencies
+  (:issue:`3756`).
+  By `Mathias Hauser <https://github.com/mathause>`_.
 
 Internal Changes
 ~~~~~~~~~~~~~~~~