Because we intend for the dictionary to be available online and fully searchable, and our primary goal is to produce this version, as opposed to the production of a print version which will then be put on-line, the first step in creating the dictionary is to choose the correct technical framework. It is important that we take the time necessary to create the correct infrastructure appropriate for our specific needs. This requires the initial establishment of two things: Means of data storage, and method of data retrieval, and this is where our on-going discussions with members of the Heidelberg Research Architecture as well as with the creators and maintainers of similar database/dictionary projects, such as the Middle English Dictionary, the Digitale Familiennamenwörterbuch Deutschlands, and the Dictionary of Medieval Latin from British Sources, come in. The two primary questions that we hope to resolve in these discussions are:
- Do we want to use TEI for our encoding (as the DFD does) or create a custom schema (as the DMLBS did)?
- What is more appropriate, an XML database (such as eXist-db) or a relational database (such as MySQL)? Both the DFD and the DMBLS use XML databases with XQuery; however, the initial dataset from which the DMNES will be built is currently in MySQL. Two posts which discuss the relative merits for each type of database can be found here and here.
We want to be sure that we pick an infrastructure that is optimal in many different aspects: in terms of search speed and efficiency, in terms of long-term maintainability, in terms of ease of use for non-specialists, in terms of cost and access (as a principle, the DMNES intends to work with open-source tools unless there is absolutely no alternative). Doing so is no easy task and we do not intend to rush things merely to be able to make headway on the fun part: Collating and interpreting the data!