Having read my own posts anew, I doubt, anyone could make out something relevant and concrete up till now.
So, I will try to explain my vague attempt for a funny what-ever once more - but this time, I will just explain the technological core and hope that its application are more or less obvious and the previous rambling speeches fall into place.
Thinking inside the box
- For a while now, ontologies try to sort things into hierarchies like biologists sorted animals into the tree of evolution. They do that good, but it's still only things.
- Like many school pupils, I have spent a good portion of my youth learning Latin - a dead language.
- Since this strange language didn't see the necessity to sort the words of a sentence (S-P-O) but place them arbitrarily we drew graphs of the sentences to find the predicate, the subject, object and the rest of the crazy things.
- The idea now would be to do that for normal sentences, too - in a digital but very informal fashion.
- The user would try to divide the sentence up into some boxes that all contain a defined part of the sentence and put names on the boxes... as he sees fit.
- Every sentence that is graphed that way would be either an implementation of an existing scheme with a new subject etc. or a new declaration, others can use from that point on.
An Example
To demonstrate the boxing process, I have
opened up the NY Times and taken the first sentence, I've stumbled over
Apple’s chief executive, Tim Cook, took the unusual step Monday of apologizing to Chinese customers over the company’s warranty policy and said he would improve customer service in the country.
- As you see, it really isn't that easy and you have to repeat some of the information that are present for the reader implicitly.
- Things or persons are referenced to a normal ontology, so you get a one-to-one mapping.
- It is quite some task to split up the sentences that way and it might take a while for a semi-automatic pre-organization...
=> So - for a mere hobby - the effort seems to be quite high...
What you gain by that
- As chaotic as that sounds, it gives the user the possibility to handle proverbs (for example) "standalone" - without the normal usages of the verb. "say he would" is an example for it - I just didn't think it would be good, to take that box apart, too - so I just let it stay together ...
- Translation between different languages often don't work word-for-word since some expressions just have no direct correspondence. Using graphs of both translations, you could establish a equivalency and a correspondence on box level (the same boxes occur in both versions).
- Since the created boxes of the sentence have to make sense on their own, you fill them up with repetitions of already expressed things (Tim Cook / Apple) - but that way, you can use and search the box without dependencies to other statement parts...
- A posed question that would have a graphed sentence as an answer might be posed in the form of a graph with wild-card-boxes and the answer might just be the content of the mysterious box.
- When you only have labeled boxes as base entities, it isn't so bad to have different interpretations of a sentence - you can define rules to cut the boxes of one interpretation in half or glue the boxes of the other one together - whatever the better version is.
- The old version would be labeled deprecated and the rule for the re-interpretation would be referenced.
- That way, you could go wrong paths (in the way to structure the world of words) without having to throw away everything and start anew.
