FAQ

Frequently asked questions

Q: How do I acknowledge the use of Especia?

To acknowledge the use of Especia, please cite:

Quast, Ralf (2016): Evolutionary spectrum inversion and analysis (Especia). Zenodo. DOI 10.5281/zenodo.785071.

Quast, Ralf; Baade, Robert; Reimers, Dieter (2005): Evolution strategies applied to the problem of line profile decomposition in QSO spectra. Astronomy and Astrophysics 431 (3) 1167. DOI 10.1051/0004-6361:20041601.

Q: How do I uninstall Especia?

By default make install installs into $HOME/bin. Read the install_manifest.txt generated by make install to find out what was installed and where. To remove the installed files cd into your build directory and type make uninstall.

Q: How does Especia model the instrument function?

The generic instrument line spread function (LSF) is modelled by a Gaussian, which is truncated at four half widths at half maximum (4 HWHM). The instrument convolution is piecewise integrated by parts.

The use of custom LSFs necessary for instruments like Hubble's Cosmic Origins Spectrograph might become feasible through future developments.

Q: How does Especia perform on unresolved narrow lines?

Tests with simulated data show that Especia models unresolved narrow lines of half the width of the instrumental line spread function reasonably well, if the spectrum data are per se fourfold oversampled. If the spectrum data are not oversampled, Especia executes a super-sampled computation of the instrumental convolution to establish accuracy.

Irrespective of the oversampling factor, Especia will not obtain much information from lines narrower than half the width of the instrumental function. Line broadening and column density parameters of optically thick unresolved narrow lines will exhibit much larger uncertainty than usual.

Q: Where are the executable files installed?

By default make install puts the executable files into $HOME/bin. Read the install_manifest.txt generated by make install to find out where the files are installed. You may move them, however, to any other location.

Q: Why do Especia result files include the model definition?

Besides the actual result, Especia result files include all information necessary to reconstruct the source information that lead to the result: the command line used to invoke Especia, the model definition file, the spectroscopic data, and an Especia version identifier.

This redundancy is inspired by the idea of reversible computation. Any computer algorithm A that maps m bits of input plus n spare bits (all zero) to m + n bits of output (the first m bits containing a copy of the input and the last n bits containing the actual result) is reversible. The reverse algorithm R merely zeroes the n result bits.

Especia is a reversible computer algorithm to proactively enable traceability of code, spectroscopic data, model defintions, and results.

Q: Why is Especia written in C++?

When I started developing Especia, C++ was my favourite programming language. Today I would probably use modern Fortran because of its vector and matrix facilities. However, C++ has been a very good choice.

Q: Why is this software named Especia?

It is a form of acronym for Evolutionary spectrum inversion and analysis. It is the Spanish word for spice, too, which is known to be the most important and valuable substance in the universe of Dune, an epic science fiction novel by American author Frank Herbert.