Some Thoughts on Databases

The state of semi-lockdown in Japan, and with the universities and their libraries all closed, has allowed me some time to explore some new online material. One of the many things I have discovered is the CALD (Corpus of Arabic Legal Documents) database (http://cald.irht.cnrs.fr/php/ilm.php). This is one output of the ‘Islamic Law Materialized’ (ILM) project, sponsored by the ERC and based at the IRHT/CRNS in Paris. 

The CALD database is accessed via the homepage of the project (https://www.ilm-project.net/en), which itself relates information about the aims and scope of the project. Primarily, the project seeks to enable researchers to work with fully searchable modern editions of Islamic legal documents from the medieval period. The aim of this is primarily twofold: to allow researchers easy access to documents that otherwise are difficult to read due to the cursive nature of the script; and to draw together documents from different collections—including previously edited volumes—to enable comparisons across time and place to be made. It is not immediately obvious when the project ran, although in the ‘News’ section the last added items are from 2013, so I presume it finished then.

The database, which has an interface in both English and Arabic, is accessed from the ILM homepage. The database’s front page contains information about the database itself, including that the timeframe covers the 2^nd-9^th/8^th-15^th centuries, and that the documents are presented as Arabic editions ‘in modern spelling’, along with what is referred to as ‘full bibliographical data’. It then details that the user can either do a full-text search using any Arabic term or browse the documents by the first letter of the city in which the document is now kept (for example, ‘V’ would bring up ‘Vienna’).

While the ‘search’ option is extremely useful, this ‘browse’ function is rather odd. I am not sure why a researcher would want to browse through only one library collection given that most research is carried out with a focus on a certain time period, type of document, or subject, for example. This is partly mitigated, however, by the fact that everything I have mentioned so far can be accessed without identifying yourself. There is, however, an option to register and login, thereby gaining access to more functionality. This extra functionality is very welcome, but I do not understand why one has to register to access it. There is no benefit to the user as far as I can see from this system of registration; I did not even receive an email acknowledging my doing so.

These leads to an issue that I observe fairly often when there is a website that has been constructed by a small team in a research institution: there seems to be little to no consideration of what a website’s ultimate users want. In other words, there is often no process of User-Interface (UI) research and testing beyond what the research team themselves would expect. This may, and often does seem to, differ significantly from that which end users would require. To those working on similar online databases in future, I would urge you to have some conversations with potential end users to see what they would want and expect out of such a product in order to construct a site that will be as user friendly as possible.

Once you have registered and logged in, though, it gets much better. There are a number of useful search tools, including ability to search for a date range, in both hijri and western calendars, a dropdown menu of 29 different types of documents that can be chosen from, and five dropdown ‘descriptor’ boxes in which it is possible to search via a document’s provenance, material on which it is written, and keywords. As such, once logged in the site gives a useful and very flexible search experience: 

So, as my research covers the sixth/twelfth century eastern Mediterranean, I chose a date range of A.H. 490-590 and provenance from Egypt. The resultant search provides ten documents. This is, however, unusual; searches for both the area of modern Turkey and for Syria yield no results. The multiple results for Egypt are primarily the result of the inclusion of various Geniza documents. While the lack of Arabic documentary sources before the Mamluk period (1250-1517) is well known, for the period 690-790 there are only six items from Egypt, three from Palestine, and none from Turkey, so the limited number is due to the limited work done by the research team rather than the number of documents in existence.

This leads to a second issue, which I touched on in a previous post: any such database as this risks being seen, particularly by students or very early career researchers, as being the go-to resource for Arabic legal documents, and thus being the primary basis for most research on the subject. This, however, is a rather dangerous position because it is just one set of documents that have been chosen almost at random by researchers over a number of years. There are thousands more relevant documents, particularly from the Mamluk period, in libraries in Egypt and elsewhere that may give wholly different results to those found here. Any database should, therefore, be used extremely carefully.

Those documents that are contained within the database are, however, excellent. The editions are extremely good with sufficient metadata included, while the site also contains images of each document to allow for a comparison of the edition and the original. I do, however, have one issue with how the site deals with these documents: no translation is provided. While this is fine for those who have excellent Arabic, it means the material is out of reach for those just learning or those studying historical topics who have no Arabic. This is especially so given that they often contain technical language that requires an expert to translate properly.

This database is a useful resource, but it also feels, to some extent, like a wasted opportunity. The issues outlined above are problematic, and could have been solved by some (or more?) UI testing. Unfortunately, this database now seems fossilized and will have no further changes made, like so many others when the project of which they are part has ended. In future, hopefully project budgets will contain within them funds for making alterations to their websites well after the original project has come to a close.

Note: All images on this blog are copyright CALD and ILM project, CNRS (Paris).