The Incremental Maintenance of Transition Tour

Abstract

While evolutionary development methodologies have become increasingly prevalent, incremental testing methods are lagging behind. Most traditional test generation algorithms -including the Transition Tour method -rebuild test sequences from scratch even if minimal changes to the system have been made. In the current paper we propose two incremental algorithms to update a Transition Tour test sequence after changes in a deterministic finite state machine model. Our solution uses existing information -the Eulerian graph of a previous version of the system and an Euler tour in it -to update the test cases of the system in response to modification. The first algorithm keeps an Eulerian graph up to date, while the second algorithm maintains an Euler tour in the augmented graph. Analytical and practical analyses show that our algorithms are very efficient in the case of changing specifications. We also demonstrate our methods through an example.Finite State Machines (FSMs) have been widely used for decades to model systems in various areas, such as sequential circuits [11], communication protocols [14], some types of programs [1] (in lexical analysis, pattern matching, etc.) and object-oriented software testing [2]. Several specification languages, such as SDL [15] and ESTELLE [31], are extensions of the FSM formalism.An FSM M is a quadruple M = (I, O, S, T ) where I, O, and S are the finite and nonempty sets of input symbols, output symbols and states, respectively. T ⊆ S × I × O × S is the finite set of transitions between states. Each transition t ∈ T is a quadruple t = (s j , i, o, s k ), where s j ∈ S is the start state, i ∈ I is an input symbol, o ∈ O is an output symbol and s k ∈ S is the next state. FSM M is deterministic if for all (s j , i) state-input pairs there exists at most one transition in T .An FSM can be represented by a state transition graph, a directed edge-labeled graph whose nodes correspond to the states of the machine and whose edges to the state transitions. Each edge is labeled with a pair of input and output symbols, which is associated with the corresponding transition.In the description of our algorithms, we use the graph terminology because it describes a more general approach. However, it is important to emphasize that the purpose of our algorithms is to maintain a test sequence of a system, whose description is given in the form of an FSM.In the rest of the paper we consider strongly connected, deterministic FSMs.

MBT in agile/lightweight processes: a process‐centred review

“…In the context of MBT, incremental FSM-based testing, which is extensively researched in the past few years [37][38][39][40][41][42], can provide promising ideas for addressing the requirements that have not been addressed in existing MBT-for-agile techniques. This approach aims at modularising test case generation and execution based on evolving test models.…”

Test generation algorithm for the All-Transition-State criteria of Finite State Machines

“…to identify the effects of changes in the test suite derived for a previous system version and to only update those parts that are necessary. As our algorithm reused some fundamental parts of the TT-method, many parts of the incremental TT [20] method can be utilized to fulfill this purpose. We also plan to extend our method for Extended Finite State Machine models, where the guarding conditions over variable values can also be considered when generating the test suite.…”

Untitled

“…The TT-method has the least assumptions about specification and implementation FSMs, but it only guarantees to find output faults. For continuously evolving FSMs, the incremental TT-method is proposed in (Németh & Pap, 2014) that updates the TT test sequence after changes applied to the model.…”