摘要: Heap data is potentially unbounded and seemingly arbitrary. Hence, unlike stack static data, heap cannot be abstracted in terms of a fixed set program variables. This makes it an interesting topic study there abundance literature employing abstractions. Although most studies have addressed similar concerns, insights gained one description abstraction may not directly carry over to some other description. In our search unified theme, we view as consisting two steps: (a) modelling, which the process representing memory (i.e., concrete locations) model abstract locations), (b) summarization, bounding by merging multiple locations into summary locations. We classify models storeless, store based, hybrid. describe various summarization techniques based on k-limiting, allocation sites, patterns, variables, generic instrumentation predicates, higher-order logics. approach allows us compare large number dissimilar abstractions also paves way for creating new mix match techniques.