NoSQL Usergroup May 2012

Just one week before the NoSQL matters conference (http://www.nosql-matters.org) will happen we want to meet again to discuss NoSQL databases and related technologies.

We've our first talk confirmed! Harald Vajkonny will talk about a MongoDB project he's currently working on:

When talking about noSQL people mostly focus on speed or the ability of handling big data. Another important feature, amorphness of record structures seems to be a minor feature which comes along with noSQL as a matter of course, but is not really a core property.

In my paper I would like to show an application area in which especially amorphness becomes crucial: linguistic data (being representative for all kinds of scientific data).

When starting out a new research, data usually comes in unstructured and chaotic. Of course one usually has a hypothesis about which data is important and what is connected with what, but there is often not enough knowledge to completely design a consistent database structure. Databases are usually built, when we know our subject matter well enough and need only to gather quantities of proof. We do not use them for primary discovery procedures. Designing SQL tables and rows too early can even become dangerous if the database design hides important correlations or imposes a specific theory on the data.

The data being presented in my paper comes from Early Celtic Philology and Judaistics. I will present data such as commentaries and dictionaries, which both appear quite complex and heterogenous. A nightmare for SQL designers. But something to be relaxed with noSQL.

During the discussion I would like to reason about similar areas in the web world, which can be compared to scientific data, and about the general question "What means database design in noSQL where after all we do not need necessarily any design?"