Frequently Asked Questions¶
Big Questions¶
What's the latest version of DEMETER?
The current version number is given on the left under the “Helpful links” heading
Where do I get the latest version of DEMETER?
The latest version is always available at http://bruceravel.github.io/demeter/.
How do I join the IFEFFIT mailing list?
Check it out: http://cars9.uchicago.edu/mailman/listinfo/ifeffit/. Just fill in the form and click “Subscribe”!
Why did Bruce direct me to the mailing list rather than answering the question in the email I sent him?
One answer is that the mailing list is a very useful resource. Questions asked on the mailing list might garner answers from multiple experts with different perspectives. Also questions asked on the mailing list might get an answer even when I am on travel or otherwise unavailable to answer questions. Folks on the list tend to be friendly and very helpful.
Also the mailing list is archived. A useful answer might be discovered by a confused person in the future. That's a very good thing.
Another answer is that there are a lot of people using DEMETER these days. I simply do not have the time to provide individual support to the hundreds of people using my software. The mailing list helps me manage my time while still helping people get over their hurdles.
The most blunt answer is that supporting this software is not actually my job. I mean that in the sense that it is not specifically mentioned in my job description – y'know, the thing I actually get paid for. Supporting my software is something I do because it is of value and because it is often rewarding. But, look ... I give you functional software for free. I wrote documentation that is often adequate and give that away for free. I give away lecture notes and training materials for free. I answer questions on the mailing list. That's a lot of stuff. The only thing I ask for in return is that questions be directed to the mailing list rather than to me personally.1 That's not a big ask.
How do I cite this software?
DEMETER
B. Ravel and M. Newville, ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT, Journal of Synchrotron Radiation 12, 537–541 (2005) doi:10.1107/S0909049505012719
IFEFFIT
M. Newville, IFEFFIT : interactive XAFS analysis and FEFF fitting J. Synchrotron Rad. (2001). 8, 322-324 doi:10.1107/S0909049500016964
LARCH
M. Newville, Larch: An Analysis Package for XAFS and Related Spectroscopies, Journal of Physics: Conference Series, Volume 430, 012007 (link)
How do I suggest a topic for this FAQ?
The easiest way to make a FAQ suggestion is to go to Demeter's issues page at GitHub and open a new issue with your suggestion. If you'd like to suggest the answer as well – superb!
You can also edit the source for this FAQ page by editing the file
documentation/SinglePage/faq.rst
in the source code repository. Sphinx is used to make the web pages and the source uses reStructured Text. Go ahead and edit the file, then make a pull request at the Demeter's GitHub page.Each entry looks like this:
#. .. admonition:: What about my question? :class: faq Here is a great answer!
The hash (
#
) should be flush against the left margin to get the numbering correct. The.. admonition:
and:class: faq
markup is used to typeset the question and answer in an attractive manner. The indentation is important – everything should be lined up underneath the “a” in “admonition”.
Questions about Athena¶
I imported a lot of data into ATHENA and now it is misbehaving. What's up with that?
This is discussed at length on the mailing list. Here is one good example. The bottom line is that IFEFFIT, the library that is used by Demeter for math and XAFS related functionality, is written in Fortran. It used static memory allocation, which means that it will eventually run out of memory if you import a lot of data. When that happens, other weird things will happen.
The solution is not to overload ATHENA. Split your work up into groups of 30 or 40 spectra. Quit ATHENA and restart to process the next group of 30 or 40 spectra.
This is one of many areas where LARCH is an improvement over IFEFFIT.
ATHENA performed an “autosave” and now it is unresponsive. What's up with that?
This is discussed on the mailing list here.
The bottom line is that there is a bug in some versions. Two work-arounds are
- Figure out where the autosave file is on disk and delete it.
It's called
athena.autosave
. - Find the
demeter.ini
file (it should be in$HOME/.horae
or%APPDATA%\demeter\
) and edit it with a text editor. Find the[athena]
section, change “autosave” to “false”.
- Figure out where the autosave file is on disk and delete it.
It's called
How do I import my data from SSRL? DUBBLE? Photon Factory? etc...
Some beamlines send their users home with strange data files. ATHENA has a plugin mechanism for managing some of those strange data files. Be sure to enable the plugin for your beamline. Here is an explanation for how that is done.
If your problematic data is not from a beamline that ATHENA already knows about, then you must include an example of the data file when you post your question on the mailing list.
The real problem here is that your beamline saves its data in an ill-conceived format. You should convince your beamline scientist to implement the XAS Data Interchange format
Why can't ATHENA import data from my 36-element detector?
It can, but it might need help. In fairness, it's not necessarily a reasonable thing to ask of a general purpose program like ATHENA. Here's an interesting discussion of this topic.
The bottom line is that you might consider processing your data prior to importing it into ATHENA by summing the individual columns and removing any bad channels.
Questions about Artemis¶
How do I measure coordination number? Why can't N be a variable?
Whew! This is a big topic. Try starting here.
Searching Google is a good idea.
In short, coordination number is inherently difficult to measure because of its correlations with σ2, with the photoelectron scattering amplitude, and with other parameters of the data processing and analysis.
N cannot be a variable because that's how ARTEMIS works. S02 can be a math expression made of variables.
I ran FEFF. Why are paths missing and coordination numbers surprising?
ARTEMIS offers a feature called fuzzy degeneracy where paths of similar length are grouped together. It is explained in detail in the document.
How do I use FEFF9 in ARTEMIS?
The answer is: you don't. Using FEFF seamlessly and in a way that is transparent for the user is difficult. ARTEMIS is designed to use the version of FEFF6 that comes with the package. Work is being done to incorporate feff85exafs into ARTEMIS.
A better question is: Why do you think you need a different version of FEFF? Read this or this for Bruce's take on the scant effect of self-consistent potentials on EXAFS analysis.
Questions about Hephaestus¶
Questions about Windows¶
ATHENA and ARTEMIS were working yesterday, but today they crash at start up (or whenever I import data, or something). How do I fix this?
There are two common culprits.
- A file in
%AppData%\demeter
has been corrupted.%APPDATA%
is typicallyC:\Users\<username>\AppData\Roaming
where<username>
is your login name. Try deleting the files in that folder and restarting the program. Make a copy of the files in the%AppData%\Demeter
folder to help diagnose what caused the file corruption. - The antivirus definitions on your computer have updated since you last used DEMETER. You may need to configure your antivirus software to allow you to run DEMETER programs and load all of its libraries.
- A file in
Questions about Macintosh¶
I can't get DEMETER to install on my Mac. What's up with that?
Questions about Linux¶
Why does it take so long for DEMETER to install from source?
The first time you install DEMETER on your computer, you have to install all of the dependencies. There are quite a lot. It simply takes time. When you upgrade on the same computer, the dependencies will already be in place, so the upgrade will be speedy.
Footnotes
[1] | Well, ok ... that's not quite true. I also ask that you cite the reference for ATHENA and ARTEMIS when you publish. |
DEMETER is copyright © 2009-2016 Bruce Ravel – This document is copyright © 2016 Bruce Ravel
This document is licensed under The Creative Commons Attribution-ShareAlike License.
If DEMETER and this document are useful to you, please consider supporting The Creative Commons.