Integrating Lucene with HBase by Boris Lublinsky and Mike Segel.
You have to get to the conclusion for the punch line:
The simple implementation, described in this paper fully supports all of the Lucene functionality as validated by many unit tests from both Lucene core and contrib modules. It can be used as a foundation of building a very scalable search implementation leveraging inherent scalability of HBase and its fully symmetric design, allowing for adding any number of processes serving HBase data. It also avoids the necessity to close an open Lucene Index reader to incorporate newly indexed data, which will be automatically available to user with possible delay controlled by the cache time to live parameter. In the next article we will show how to extend this implementation to incorporate geospatial search support.
Put why your article is important in the introduction as well.
The second article does better:
Implementing Lucene Spatial Support
In our previous article [1], we discussed how to integrate Lucene with HBase for improved scalability and availability. In this article I will show how to extend this Implementation with the spatial support.
Lucene spatial contribution package [2, 3, 4, 5] provides powerful support for spatial search, but is limited to finding the closest point. In reality spatial search often has significantly more requirements, for example, which points belong to a given shape (circle, bounding box, polygon), which shapes intersect with a given shape and so on. Solution, presented in this article allows solving all of the above problems.