Developing the taxonomy and application
If you have read the first two articles in this series, you will know that they centred on design considerations for a desktop content graph. Now it's time to move on to more practical matters; developing the taxonomy and building the desktop content graph application. I've split my description of this stage of the work into two parts.
Part 2 covers the design of a taxonomy and its development out of a collection of Finder tags.
In part 3 I've begun the process of developing a desktop application. I've described the application design, the user experience considerations and the way that I approached the code development. If you've read other articles by me on software, you'll know that I'm the first to say that I am not a professional software developer. I am a hobbyist programmer, and generally write software only to solve a specific information need. In this case the specific need was to find a way to link pieces of desktop content (in files) to other content, such as other files or taxonomy concepts (I talked in part 0 about the lack of existing tools for doing this). So the initial iteration of the system looks like this:
Recommended by LinkedIn
It's fairly basic. The user can drag files into the workspace from the Finder, and link related files together using a symmetric relationship derived from the content model. The system captures Finder tags from the added files and converts them in the background into SKOS concepts.
You will see that I'm not finished yet. The initial application proves the concept, of being able to build an RDF graph based on file system objects and taxonomy concepts. Beyond that, I need to show how to move this graph data into a graph database (in this case, GraphDB) or alternatively (if the user has no access to a dedicated graph database tool) to serialise and de-serialise the graph data to disk. I need to dramatically improve the user experience, particularly for linking files in the workspace. I'm working on these right now and will be writing about them soon. I will also make the code (written in Xojo) available in github, once it reaches a stage that I feel confident to share!
For now, I hope you will read and enjoy the articles.