On 01/02/2012 21:13, Gustavo Seabra wrote:
> Hi Mark,
>
> That's some interesting work. My guess is, as long as we still have
> the regular human-readable output file, that seems actually a pretty
> nice addition. However, how is it going to work for any possible
Thank you. I do not envisage this replacing the normal mdout file; more
a machine readable augmentation to it, in the form of an extra,
independent, output file (with the possible option of automatically
zipping it at the end of the calculation..)
> future code-changes or additions? Would any newcomer have to learn CML
> and implement a CML output routine along with whatever work they are
> doing?
With reference to the source code, the FoX library takes care of most of
this difficulty; one generally just calls set subroutines from the FoX
module. There is a very good tutorial on how to implement FoX (using
WXML) in existing MD code, here:
http://www1.gly.bris.ac.uk/~walker/FoX/WCML_tutorial.html
Hence, anyone looking to add CML output for a new feature, should be
able to follow the example on that page.
I think the key thing at the moment is developing an AMBER dictionary (
http://xml-cml.org/spec/dictionaries ). This is an example one from SIESTA:
https://github.com/andreww/ccviz/blob/master/dictionaries/siestaDict.xml
There are more examples of dictionaries in:
https://github.com/andreww/ccviz/blob/master/dictionaries
I'm currently working on getting a basic AMBER one, and someone adding
something new in future may need to add something to this dictionary.
Regards,
Mark
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Wed Feb 01 2012 - 15:00:02 PST
