David,
I need to answer to your first assertion separately:
On 29/05/14 01:48, David Cuenca wrote:
Well, our goal it to gather the whole human knowledge, not to use it.
No, that is really not the case. Our goal is to gather carefully selected parts of the human knowledge. Our community defines what these parts are. Just like in Wikipedia.
Even if you wanted to gather "all human knowledge" this goal would not be a useful principle for deciding what to do first. For example, we know that every natural number is an element of the natural numbers. It is obviously not our goal to gather these infinitely many statements (if you disagree, you could try to propose a bot that starts to import this data ;-). Therefore, it is clear that gathering *all* knowledge is not even an abstract ideal of our community. Quite the contrary: we explicitly don't want it.
The natural numbers are just an extreme example. Many other cases exists (for instance, we do not import all free databases into Wikidata, although they are finite). The question then is: How do we know what data we want and what data we don't want? What principles do we base our decision on? For me, there are two main principles:
* practical utility (does it serve a purpose that we care about?) * simplicity and clarity (is it natural to express and easy to understand?)
You said that we cannot foresee *all* applications, but that does not mean that we should start to create data for which we cannot foresee *any*. There is just too much data of the latter kind, and we need to make a choice.
Don't get me wrong: I consider myself an "inclusionist". Better to have some useless data than to miss some important content. But there is no neutral ground here -- we all must draw a line somewhere (or start writing the natural number import bot ;-). My position is: if we have data that is very hard to capture and at the same time has no conceivable use, then we should not spend our energy on it while there is so much clearly defined, important data that we are still missing.
Markus