With the explosive growth of the tablet form factor and greater availability of pen-based direct input, online writer identification is increasingly becoming critical for person identification, digital forensics as well as downstream applications such as intelligent and adaptive user environments, search, indexing and retrieval of handwritten documents. Extant research has approached writer identification by using writing styles as a discriminative function between writers. In contrast, the authors model writing styles as a shared component of an individual's handwriting. They develop a theoretical framework for this conceptualisation and model it by using a three-level hierarchical Bayesian model (Latent Dirichlet Allocation). In this text-independent, unsupervised model each writer's handwriting is modelled as a distribution over finite writing styles that are shared among writers. They test their model on a new online handwriting dataset IBM_UB_1 and also offer benchmark comparisons by using the IAM-OnDB database. Their experiments show comparable results to the current benchmarks and demonstrate the efficacy of explicitly modelling the shared writing styles.