adapted the configuration filenames, highlighting, spelling

docs/advanced.rst

 Advanced and productive use
 ===========================
 
-Using FireWorks with a batch system on HPC clusters
----------------------------------------------------
-
- * QAdapter and FWorker configuration
- * Separate configuration, launches and templates
-
-
-
 
 Use firework categories and fireworker names
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -17,46 +9,75 @@ If fireworks have to be launched on different workers (for example in
 heterogeneous and distributed computing workflows) the keywords ``_category``
 and ``_fworker`` can be used to specify which categories of fireworks can be
 launched by what fireworkers. In the following example, the firework can be 
-launched by fireworker with name *uc1* and configured to process category
+launched by any fireworker with name *uc1* that is configured to process category
 *turbomole*::
 
   _category: turbomole
   _fworker: uc1
 
-Note that the same fireworker may process more than one category or
-non-categorised fireworks. This is specified in the fireworker configuration
-file in **demos/6_advanced/fworker_turbomole.yaml**::
+A matching fireworker is configured in file
+**demos/6_advanced/fworker_uc1_turbomole.yaml**::
 
   name: uc1
   category: [turbomole]
   query: '{}'
 
+In this example, the fireworker may launch only fireworks from category *turbomole*
+and tagged with fireworker name *uc1*. Note that the same fireworker may process
+more than one category or non-categorised fireworks if configured correspondingly.
+
+
 Submit jobs to the batch system using ``qlaunch``
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 If a firework launch takes large resources then it should be submitted to a
 batch system on the HPC cluster. This is done with the ``qlaunch`` command. The
-command is configured with a qadapter configuration file. For MOAB and SLURM
-example configuration files see **demos/6_advanced/qadapter_moab_turbomole.yaml**
-and **demos/6_advanced/qadapter_slurm_turbomole.yaml**.
+command is configured with a qadapter configuration file. For configuration files
+for the queing systems MOAB and SLURM see the examples provided in the folder
+**demos/6_advanced/config**.
+
+In practical use cases, it is generally of advantage to separate the configuration
+files, the workflow templates (i.e. the input) and the launches. This structure
+is used in the demo 6 in folder **demos/6_advanced**.
+
+To submit a single fireworker batch job on bwUniCluster the ``qlaunch`` command
+is called like this::
+
+    cd launches
+    qlaunch -q ../config/qadapter_uc1_vasp.yaml -w ../config/fworker_uc1_vasp.yaml singleshot
+
 
 
 Deal with failures and crashes
 ------------------------------
 
- * FIZZLED: what to do next?
-  - If the error is external (i.e. not in the ``spec``) and fixed then just ``lpad rerun_fws``.
-  - In-place fix with ``lpad update_fws`` and ``lpad rerun_fws``. Advantage: preservation of all independent COMPLETED fireworks is guaranteed
-  - Fix the error in and resubmit the whole workflow.
+For some reason the execution of a firework may fail and the firework gets 
+*FIZZLED* state. Depending on the reason for the error there are different
+approaches to handle the error:
+
+* If the error is external (i.e. not in the ``spec``) and fixed then the firework
+  can be rerun using the command ``lpad rerun_fws``.
+* If the the error is in the *spec* of the firework then this can be in-place
+  fixed with the command ``lpad update_fws`` and then rerun with ``lpad rerun_fws``.
+  Advantage: preservation of all independent *COMPLETED* fireworks is guaranteed.
+* If the error is in the *spec* then it can be fixed in the workflow template
+  and the whole workflow is added again to launchpad. This approach is not
+  practical with increasing number of errors and updates in the same workflow.
+
+Detect lost runs
+----------------
+
+If a job is killed by the batch system its status *RUNNING* gets never changed.
+In order to detect such running fireworks we use the command ``lpad detect_lostruns``
+which will return the IDs of fireworks with lost runs. Optionally, these can be
+rerun set to *FIZZLED*.
 
- * RUNNING forever
-  - Use the command ``lpad detect_lostruns`` to set to FIZZLED or to rerun
 
 Detect duplicates
 -----------------
 
-Fireworks can reuse the data from the launches of identical Fireworks
-(duplicates). To enable detection of duplicates the following key is added to the *spec*::
+Fireworks can reuse the data from the launches of identical Fireworks (duplicates).
+To enable detection of duplicates the following key is added to the *spec*::
 
   _dupefinder:
     _fw_name: DupeFinderExact
@@ -77,7 +98,7 @@ identical section including four fireworks (1-4)::
 
 The second run of ``rlaunch`` detects four duplicate pairs whereas only the last
 firework of the second added workflow is executed. After this both workflows are
-in COMPLETED state which can be checked with ::
+in COMPLETED state which can be checked with::
 
   lpad get_wflows -t -m 2 --rsort created_on
 
@@ -86,7 +107,8 @@ Let us now delete the first workflow for that all fireworks have been executed::
   lpad delete_wflows -i 1
   
 We see *Remove launches []* in the output, i.e. its launches have not been
-removed. With deleting a workflow including duplicated fireworks the shared launcher is removed from launchpad only if all duplicated fireworks are deleted.
+removed. With deleting a workflow including duplicated fireworks the shared
+launcher is removed from launchpad only if all duplicated fireworks are deleted.
 The launches are related now only to the relevant fireworks of the second
 workflow. The launches will be removed if we remove the second workflow::
 
@@ -113,24 +135,15 @@ us test this with re-running a firework that is identical to another firework::
   # the two fireworks are COMPLETED
 
 
-Security best practices
------------------------
-
-Configure security (MongoDB authentication and authorization)
-
-
-
-
 Query and analyse data from fireworks and workflows
 ---------------------------------------------------
 
-TODO
 
 
 Use FilePad to store fireworks file inputs and file outputs
 -----------------------------------------------------------
 
--> 
+
 
 Not every feature of FireWorks is covered in this tutorial. Please visit the
 the documentation website https://materialsproject.github.io/fireworks/ for