The text on this page was automatically translated and hence may differ from the original. No rights can be derived from this translation.
The KB, the national library of the Netherlands, provides access to historical text collections such as digitised newspapers and books through its online services. The KB found a gap between user-friendly search services like Delpher and DBNL on one hand, and the service Dataservices for advanced digital research on the other. Dataservices is not accessible to many users because it requires them to store requested data and program algorithms themselves. At the same time, the KB sees increasing possibilities for digital research on its text collections. Therefore, the KB commissioned Dialogic to explore whether there is a need among its users for an analysis platform where multiple collections (from the KB and beyond) can be integrated and analysed in a so-called "text suite".
To understand how a text suite can support users in their research on (historical) text collections, we have created a schema of different research phases and needs that may arise per phase, based on literature research. This schema is shown in the figure below. Through interviews and a survey of 873 users of KB services, various potential functional needs were tested to determine how a text suite can add value.
[Image of a research process text suite schema]
We conclude that there is no clear need for advanced capabilities in the Analyse phase. Although this was the starting point of the exploration, interview partners and survey respondents indicate less need for this and less usage if offered. Three main arguments are highlighted. Firstly, due to the large heterogeneity of source material from the KB and beyond, researchers prefer to bring everything together on their own computer for analysis. The alternative of allowing a text suite to import sources raises questions about the sustainable preservation of compiled collections. Secondly, with rapid developments, especially in quantitative analysis tools, interview partners see it as a risk that the KB offering tools may quickly become outdated if not used enough to justify continuous development efforts. Lastly, we observe that where analysis functionality is already offered in existing platforms (e.g. the n-gram viewer in DBNL or frequency analysis in Nederlab), it does not seem to lead to significant recognition and broad application for new research questions. The latent need for such functionalities thus appears limited.
On the other hand, we find a clear need for more advanced capabilities in the Discover and Select phases. Our recommendation to the KB is therefore to position a text suite as a user-friendly tool for users and researchers to make their own selection of data that they can export for analysis using their own tools.
The full report can be downloaded at [https://doi.org/10.5281/zenodo.6591571](https://doi.org/10.5281/zenodo.6591571). Based on our findings, the KB has already decided to develop a service to support advanced capabilities for discovery and selection.
