Chris Ford in Creating music with Clojure and Overtone uses an example of a harmonic sound missing its first two harmonic components and yet when heard, our ears supply the missing components. Quite spooky when you first see it but there is no doubt that the components are “missing” quite literally from the result.
Which makes me wonder, do we generally supply semantics, appropriately or inappropriately, to data?
Unless it is written in an unknown script, we “know” what data must mean, based on what we would mean by such data.
Using “data” in the broadest sense to include all recorded information.
Even unknown scripts don’t stop us from assigning our “meanings” to texts. I will have to run down some of the 17th century works on Egyptian Hieroglyphics at some point.
Entertaining and according to current work on historical Egyptian, not even close to what we now understand the texts to mean.
The “I know what it means” (IKWIM) syndrome may be the biggest single barrier to all semantic technologies. Capturing the semantics of texts is always an expensive proposition and if I already IKWIM, then why bother?
If you are capturing something I already know, that can be shared with others. Another disincentive for capturing semantics.
To paraphrase a tweet I saw today by no-hacker-news
Why take 1 minute to document when others can waste a day guessing?