One of the foundations of open science is to have open approaches to scientific research. Open and transparent contents of computational software play an important role in ensuring research replicability. The project for upgrading the MELA focuses on developing software content in collaboration with the scientific community.
MELA is software for the analysis and planning of forestry operations, specifically developed for Finnish conditions. Its earliest versions have been in use since the 1970s. During its four-year duration, the MENU project will completely modernise and upgrade the software. The MENU project does not use or maintain the old MELA source code; this is done in other LUKE projects. Instead, the aim is to develop a more versatile, transparent, scalable and user-friendly software that is easier to maintain. The development work makes use of open source code and co-development.
Five elements in upgrading software content
The software development consists of five important elements (see figure): modeller, MELA team, model function library, co-development and meta-descriptions.
The basis for all the development work is the modeller, which is the content provider in the system. For example, a researcher wanting to provide their computational models or model functions resulting from their own research for the use of other researchers or decision makers.
In the new MELA software the system searches for model functions suitable for each calculation or application based on the meta-descriptions. In practice, the model function library stores single functions, which are linked with additional information or meta-descriptions, for example on the model’s variables or its suitability for different datasets or uses. Different uses can be for instance reports estimating the feasibility of investments based on wood, non-wood forest products or other ecosystem services.
In order to achieve that the different model functions of several researchers can function properly together, the MELA team harmonises or certifies the functions of open source code that the researchers wish to share with other users through the MELA model function library. This certification process for model packages between the modeller and the MELA team is done iteratively. As a result, there will be a certified model function library, which will be published under open source code licence. This model function library can be used either as a part of the MELA software or as a separate component in other applications.
The new MELA provides researchers with the possibility to use in different simulations not only their own model functions, but also those developed by other modellers. The developer of the original model gains visibility to their research, and the accessibility of the system is increased when content is being developed by several modellers. In addition, this results in improved transparency and easier maintenance of the software.
At present, the implementation of the model function library is progressing according to plan as a part of the software development project.
Researcher workshops are some of the most important practices in software co-development. In order to introduce the co-development of MELA software data content, a researcher workshop was organised in the autumn of 2019. In addition to the presentations at the workshop, we collected feedback in order to define practices for open source code and software co-development. The plan is to organise at least one researcher workshop during 2020.
Figure. Plan of the model function library processes in the project to upgrade the MELA software.
Authors: Daniel Melander and Tuula Packalen