The elasitcube description is fairly correct.
The QlikView data model is relational, i.e. tabular. And, yes, associative databases are completely different beasts. The associative part of QlikView is in the user interface, how you work with the data. It is not in how QlikView internally stores data. If our marketing material talks about an "associative database", then it is a mistake that should be corrected.
The QlikView techology is unique: No other tool that I know of uses color coding with group theory to display data so that the user can work in an associative way: Using any field as a starting point for a chain of selections; no drill-down paths that you have to follow (you can have them when it helps the user though); all calculations on the fly so that you aren't limited by what is pre-defined. A proficient user can add his own calculation in the user interface.
Further, the type of join used (inner, outer, left, right) is determined by the user selection, so that it is contextual, to further support the user-driven associative experience.
The QlikView associative experience is about letting the user be in charge, so that the user's train of though determines how the analysis is made.
In many times I have read in QlikView community and
QlikView whitepapers that QlikView "has associative database technique "
In fact I just finished writing a research about the
advanced techniques which used by QlikView ( which I am personally interested
in them) and in my research I depended that QlikView "has associative
database technique "
But now I must to research again about the technique which used by
QlikView to associate the data element values in the RAM in detail
In fact I don’t know why QlikView used this sentence in their marketing material or
community or what they attend by doing that