Help talk:Chains of relations and attributes

[edit] Page needs cleaning and updating

Sorry to say, but this page is brutal to understand. Please simplify and use better examples so that we can use the software!

Thanks.

See Help:Semantic search for simpler queries.--Patrick 12:01, 11 June 2007 (CEST)

Overall, I agree with the first writer. This is a very important help page, on an advanced but very useful topic, and an effort should be made to make it as "intellectually" accessible as all the other help pages. I.e.: The use of generic sets such as A, B, C and V and generic relations such as R1, R2 might be logically elegant but will put "normal" people off. For example take the first sentence in the first section:

• For a given A, a set of A's, or unrestricted A, we can find C for which there is a B such that C R1 B R2 A.

If I get its meaning correctly, it can be rephrased into something like this:

• "Suppose we have a number of pages about rivers. We know that some of these rivers are related to cities, as they pass through them, and a proper semantic relation has been created for this purpose: [[flows through::<city>]]. All cities are then related to nations, through the relation [[is located in::<nation>]]. Now, how do we find to which nations is each river related to?

Less elegant and definitely much more verbose, but far more intuitive.

Additionally, there's the issue of the "R1inv", which I interpret as "the inverse relation of R1". The question that popped in my mind while going through the page is this: is there a syntax to get the inverse of a relation or do I have to explicitly create appropriately reciprocal relations such as (city) [["is located in::<nation>"]] and (nation) [[is location of::<city>]]?

Manu3d 01:35, 5 July 2007 (CEST)

"The first method requires annotations for R1inv" implies that corresponding annotations for R1 are not sufficient. It would be useful if they would be.--Patrick 10:00, 12 July 2007 (CEST)

The first writer is absolutely correct. To be sure, a more complex query requires a more complex explanation, but what's needed is a clear explanation.