The LAMDA project, http://home.strw.leidenuniv.nl/~moldata/, does not use version numbers. 
Instead they post notices of updates on the web site 
http://home.strw.leidenuniv.nl/~moldata/database.history 

This file describes revisions to the database.  When it is updated we must also update the citation to LAMDA
in  ~data/citations_data.txt 

====================

2015, June 30, Nick Abel updated files.  

Changes in the LAMDA web site:
Two molecules, isotopes of CS (13CS and C34S)
were removed between the previous version of LAMDA and the one now implemented.  
Additionally, five new molecules were added to LAMDA
OH+, H2S, H2CS, C2H, and NH2D.  

The LAMDA level energies are not given to spectroscopic accuracy.
We compute the wavelength by taking the energy difference between two 
levels and setting it equal to h*c/wavelength, some wavelengths were identical.  This was not
physical.  Rather, it was due to the energy separation between two levels being the same. 
This was caused by not having enough significant figures for the low lying levels of molecules 
with hyperfine structure, notably OH and HCN.  To rectify this, we went to the JPL website
and sorted the transitions by wavenumber, instead of frequency, and used this output to add 
extra digits to the energy levels and make the wavelengths unique.  

The LAMDA data format had changed between 2010 and 2015.  

For HCN, the ordering of the transitions did not match with the ordering of energy levels. Therefore, when the 
energy levels were reorganized by energy, some LAMDA transitions had the upper level being a smaller number than
the lower level, which is impossible.  In the HCN datafile, LAMDA had put the collision rates to zero, recognizing
that such a transition is cannot generate light.  We simply deleted these collisional rates and reduced the
number of collisional transitions in the datafile accordingly.

For HCl, there was an extra space added in the datafile, so that the first character on an new line was a space.  This was deleted.

Several of the datafiles have been reorganized so that the levels are not in order of increasing energy.  
Our code was not designed to deal with this.
However, the files where this was done belong to molecules that are not currently predicted by Cloudy, or
molecules whose collisional data has not changed in five years.  We therefore used the older files
(which have the energy ordering correct) for the molecules predicted by Cloudy.  

There were significant changes in intensities of H2O and CS lines due to improved data.  These changed
the temperature which resulted in changes in some high-J CO lines.

====================
2010 Dec 18

We reference LAMDA by date of data retrieval.  The file ~data/citations_data.txt contains
something like

 Much of the molecular emission data is from LAMDA as accessed on Dec. 18, 2010. This database is described in:
        Schoeier F.L., et al., A&A 432, 369 (2005) [http://adsabs.harvard.edu/abs/2005A%26A...432..369S]
      This database is in part based on the work of the following groups:
        JPL: Pickett H. M. et al., JQSRT, 60, 883 (1998). [http://adsabs.harvard.edu/abs/1998JQSRT..60..883P]
        CDMS: Mueller H. S. et al., A&A 370, L49 (2001). [http://adsabs.harvard.edu/abs/2001A%26A...370L..49M]
              Mueller H. S. et al., J. Mol. Struct. 742, 215 (2005). [http://adsabs.harvard.edu/abs/2005JMoSt.742..215M]
 
When data files are updated this citation must be too.
