Software component models: past, present and future

Abstract

In the early years of the CBSE Symposium, much research was focused on identifying the desiderata of CBSE and developing different approaches to CBSE. However, a common framework for defining and analysing CBSE approaches with respect to these desiderata was only introduced later: this was provided by the notion of component models. Every CBSE approach is underpinned by a component model, and therefore the study of component models, in particular how to define ones that can potentially meet the desiderata of CBSE, is pivotal to the success of CBSE. We have surveyed and studied existing CBSE approaches and their corresponding component models, and as a result we have: (i) shown that early approaches/models do not fully meet the CBSE desiderata; (ii) identified criteria for designing component models that can better meet the CBSE desiderata; (iii) defined a new component model according to (ii); (iv) defined a taxonomy of existing component models based on the desiderata. In addition to the classic desiderata, nowadays CBSE has to address new challenges posed by an unprecedented increase in the scale and complexity of software applications, in particular safety-critical ones. As a result, there are new CBSE desiderata for which we need to define new models. In this tutorial, we will: (i) present a taxonomy of existing component models, both old and new; (ii) discuss how well they meet the classic desiderata; (iii) discuss criteria that new models must meet in order to address future CBSE challenges. The CBSE Symposium celebrated its fifteenth anniversary in 2012. For the next 15 years, the study of component models will continue to play a pivotal role in future CBSE success. This tutorial aims to contribute to this effort.