2 September 2013 (Current release)

Source code and Mac-OS binaries

For this release large parts of the code which create the isochrones have been re-written, which has resulted in a simpler structure, within which it is easy to add new models. As a result there is a far richer range of options available, yet the code is lightly shorter than the previous release. Old hands, used to previous versions should note the following points.

1) The absolute values of tau^2 have changed. Of course this does not make any practical difference as only relative values used, but there was a conceptual "bug" in which when normalising the isochrone image I did not divide by the pixel size.

2) Monte is structured so that the binary fraction can be a function of primary mass. Now the binary fraction for stars greater than 14.4Mo is 1.5 times the normal binary fraction, and the mass ratio for such stars is distributed as described in Naylor (2009; IAU paper).

3) The extinction vector for the Tycho colours has been improved. The difference between the new values and those in Mayne & Naylor (2008) is small (delta Av = 0.05*E(B-V)) so does not affect the results in Mayne & Naylor or Naylor (2009), but would change things if you had large extinctions.

4) Some more FITS keywords have been added to the output of monte so that both DS9 and the latest version of GAIA give you the co-ordinates in colour-magnitude space.

5) The scripts have been updated and uncer changed slightly to make the code compile on 64-bit machines.

21 August 2009

cmdfit.tar.gz_21-08-09

cmdfit_mac.tar.gz_21-08-09

Bug found by Phill Cargile fixed, which caused a crash on certain Bessell atmospheres.

27 June 2009

cmdfit.tar.gz_27-06-09

cmdfit_mac.tar.gz_27-06-09

(1) This release is of the version of the software used in Naylor (2009).
(2) The major change is that the age axis is now given in log10(yrs) rather than Myr. This makes things much easier when dealing with data which have a large uncertainty in age, as this is a far more natural axis to use. It also means the prior probability is flat in logarithmic not linear space, which makes a difference if the uncertainities are more than half a dex.
(3) There are now "user" isochrones, which allow you to supply a file of either (i) luminosity, effective temperature and gravity or (ii) colour and magnitude for a given age. Now interpolation in age is carried out on this files, allowing you to make the choice as to how this interpolation is done is therefore yours. The formats are dealt with in the section of the manual on isochrones.

29 October 2008

cmdfit.tar.gz_29-10-08

Documentation fix for example 2.

15 October 2008

cmdfit.tar.gz_15-10-08

To be on the safe side, I switched off the extraoplation of the colour-Teff relationships.

24 September 2008

This is a major update in techniques, which I hope to write up, but for the moment should be quoted as Naylor (in prep). (1) I use a new normalisation, which retains the mass function in the grids, but then normalises the grid to sum to one between the brightest and faintest datapoints. This avoids the problem of the normalisation going infinite if the isochrone is vertical.
(2) To generate the uncertainties I now use the uncer method (replacing bootstrap), for which there is a brief description. Its faster and more accurate.
(3) The calculation of the expected value of tau^2 has been improved, and is now done in a program called tau2, which replaces tau. The old estimate became increasingly poor as the number of free parameters increased.

19 February 2008

(1) A correction to Example 2. The extra uncertainty to be added to the colour was given as 0.03, it should be 0.042. Both the web-page and the sample files have been updated to reflect this change.

21 August 2007

(1) V vs V-I for the Geneva isochrones.
(2) V vs B-V for the Siess and Baraffe isochrones.
(3) The option to use Bessell et al (1998) colour and bolometric corections.
(4) Grid now imposes the upper limit on tau-squared to be the tau-squared of the best fitting data point (i.e. that with the lowest tau-squared) plus some value. We do this so that if a different normalisation for tau-squared is used, which efectively adds a constant to all the tau-squared values, the answer remains the same. With the chi-squared-like normalisation the lowest tau-squared value was so close to zero this has little effect on the fits.

05 June 2007

(1) The Padova option was using the z=0.004 tracks from the Cioni et al (2006) papers, not as stated on the web pages Girardi et al (2002). The code has been changed so the web page is correct, and the web page now explicitly states its the z=0.019 models.

03 October 2006

(1) Corrected a problem whereby if the mass-range you requested outran that available in the isochrones for model numbers greater than 20 (i.e. those which come ready calibrated in colour and magnitude), you got some strange low-level effects in the 2D ischrones.
(2) Updated Example 1 so the numbers are correct for the new code (the differences are within the uncertainties).
(3) Added .in files to Example 1, which contain the precise inputs needed to run the programs.

30 September 2006

(1) Corrected bundle so it includes files which are soft links on my machine.
(2) Compiled all software statically, to improve portability.