By Redpencil on Feb 29, 2020
We write free-form content day-in day-out. This is free-form text right here. Sometimes the text is prose, sometimes it contains factual information. Oftentimes it results in new facts. Sometimes informational, sometimes actionable. There is one experience that joins them: the embedded information is rarely ever reused in software applications.
So what if we could express such facts whilst we are writing text? What if we could use those facts in other applications and reuse them in other documents? What would happen if your editor would complete your input?
Meet Say, the editor that knows what you are talking about.
Say is a project created by the Flemish Government, Agency for Local and Provincial Government. In this context, Say is configured to understand the legislative process. It understands the domain of mandates and the general model for decision making. As legislation needs to stay available in the long run, the editor produces HTML with embedded RDFa which is expected to stand the test of time. The benefits are clear. Databases which used to be hand-filled are now automatically constructed with much higher accuracy and the long-term benefits for searching through annotated archives can be stunning. The Flemish Government shares all off this goodness in the liberal MIT license.
Say is not an AI gimmick, though you could turn it into one. Say is configured and programmed to understand your domain. And that is exactly what we'll do in this hands-on developer session. We will start with a fairly blank say-editor configuration and extend it to make references to an external source like dbpedia.
[Image retrieved from https://de.wikipedia.org/wiki/Datei:Schreibkugel_von_Hans_Rasmus_Johann_Malling_Hansen_01.jpg , photografer: Pavel Eremeev, (c) upashi.eu, cropped]