I’ve just returned from the Digital Humanities conference in Utrecht, and very excited to share details of some of the workshops with those who were unfortunate enough to miss it. One workshop I went to was called “I am the one building the tool“. It was quite interesting.
Around the table sat software engineers from Germany, the UK, the US and Israel who develop digital humanities related software discussing the problems we confront. It was fun meeting people that deal with the same problems we do.
The most pressing issue, we agreed, was lack of academic recognition for research software development. People who develop research software (sometimes referred to as RSEs – Research Software Engineers) do not get enough academic recognition. It is hard for them to publish, which makes it harder for them to proceed in the academic ranks.
Some suggestions were made – such as developing a way to peer-review software, so source code can be published academically, or maybe try and publish the data model as an academic publication.
Figuring out how to model the data – which database to use, how to divide the data between tables or collections or how to format the documents (in cases of a document database) is probably the most important part of a DH software development project. If you get your data model wrong, the entire development software suffers. Finding the right data model is hard. It needs expertise. It needs deep knowledge of the problem’s domain. It is something others can learn from.
We have approached some of our DH clients and suggested this to them. Some of them have already agreed – in the next couple of years they are going to publish papers explaining the way the data is organized in their systems. I hope others follow suit.