Photo Credit: Michael Rose

This is the official site for the Tonto library.

Tonto is available to non-commercial entities under the GPLv2 open source license.

## Contacts

If you are experiencing technical problems you can contact Dylan Jayatilaka.

Commercial entities or those requiring consultation for using diffraction-related applications should contact Simon Grabowsky.

## What it does

Tonto is best known for it’s ability to:

• Refine X-ray crystal structures for molecular crystals using ab initio wavefunctions — using “tailor-made” structure factors. This allows the refinement of ** Hydrogen atom positions and atomic displacement parameters (ADPs) ** Anharmonic vibration parameters for heavy atoms

• Obtain “experimental wavefunctions” by constraining a single determinant Hartree-Fock (HF) or density functional theory (DFT) wavefunctions to reproduce X-ray diffraction data.

• Obtain a large array of wavefunction analyses including one- and two-electron properties, and bond-index analyses.

## History

Tonto was written at the turn of the century from scratch to be object-oriented and Fortran-based. At that time it was not clear if Fortran would continue as the dominant language for numerical analysis and high performance computing. Therefore, bets were hedged and the code was written in a custom language Foo which translated into Fortran. The intention was to be able to transform Foo code into C++ or another language should it become necessary.

Fifteen years on, the situation is as unclear as it was before.

Certainly Fortran 2008 and Foo have converged, but the latter still has extra and useful mixin and procedure annotation features. And now there are other contenders such as the Julia language. Developments in functional programming languages mean that the perl-based Foo translator can be replaced by efficient and easy-to-use monadic parsers generators, say using the beautiful Haskell language. As a matter of fact, this development now make it a realistic possibility for an individual or small group to develop a fully compiled language — say through the the LLVM compiler back end — exactly as the Julia folk have done.

It seems the decision to use a domain-specific language has been vindicated and the future looks bright.

## Contributors

The contributors are

• Lukas Bucinsky
• Hans-Beat Buergi
• Patrick Cassam-Chenai
• Birger Dittrich
• Simon Grabowsky
• Daniel Grimwood
• Dylan Jayatilaka
• Peter Spackman
• Mark Spackman
• Mike J. Turner
• Magda Woinska
• Stephen K. Wolff