In this talk I shall give a brief introduction to the Hierarchically Tiled 
Array (HTA), 
a data type on which researchers from UIUC, IBM and UDC (Spain), have been 
working for some years.  HTAs enable the specification of locality and parallelism
in object-oriented languages in a straightforward way by means of operations on tiles.
We have found that codes written with HTAs are very readable, and their performance
is similar to that of other approaches to express parallelism. Implementations in 
MATLAB and C++ have been developed, with the focus to date being on 
distributed memory systems.  In the second part of the talk I shall discuss the 
design options, challenges and opportunities we face when considering a shared
memory implementation of HTAs .