How to do Tiny Data?
By Rajko Thon
Tiny Data - What is that supposed to be?
Imagine the following situations.
A few months ago, you had a great idea and wrote it down in an electronic file. Suddenly you need your notes, but how do you find them? Because somehow they are not in the place you remember them to be. And your operating system doesn’t know about a thing called “note”. But you could still search your entire drive…
Or how about that e-mail you received a while ago which included those important documents. Of course you just don’t remember who sent it to you or when it was sent or any useful word of the context. Must it really be so difficult to find things quickly when you need them?
Not to forget this book you always wanted to buy. And after you bought it you soon realize that you already own it. If only you could have known that…
Where these examples come from there are countless more. Of course there are already applications that manage e-mails, notes, documents, inventories of books and other objects - but is there a single one which handles all of it? What’s more, is there such an application, that also keeps my data private? And could it still allow me to browse my data as comfortably as if it were on the web?
Interesting, but we have not arrived yet!
So the idea unfolds
When big data is associated with the collection of huge, rather unstructured datasets, which are subsequently evaluated in an arbitrary way using automatic mechanisms, can a desktop application with a small database even provide comparable benefit? What if we trade the potential benefits of large unstructured amounts of data against the tangible benefits of manageable structured ones? What if our few, locally stored data were deprived of the access of large centralized services? Will we still be able to effectively search them, like we could with a large web-based search engine? What is it worth if we can decide ourselves with whom we share which of our private data? Is it even conceivable to decentralize the Internet more strongly without losing functionality we already love and use on a daily basis?
We think this may be possible. And if you look around, you may find others to think in similar ways. Look at this for example: Web Decentralisation Project.
So what to do instead of Big Data? We introduce as an alternate approach (drum roll): Tiny Data!
So the question arises: How to do Tiny Data?
How to organize personal data in a manner to keep them private and to extract as much benefit as possible? The idea of Tiny Data has evolved into a concept for a real application. This application will make use of an architecture based on a relational database, an accompanying process management system and a graphical user interface. Our languages of choice are PSQL, Python 3 and Object-Pascal (so far). In subsequent posts we will explain the system and its components in more detail. Furthermore, we will demonstrate the progress on the way based on concrete things we want to manage.