I’m still at Sun Engineering and I ran a BoF on my Applications’ Citizenship theory. For those (all?) of you who’ve not heard them, I attempt to classify an applications sociability by observing its behaviour. Applications can thus, be good citizens and co-exist with all apps that do not themselves exhibit some form of non-social behaviour. The three classes of non-social behaviours are bad citizens, which jeopardise the operating system, anti-social citizens which cause a limited, known and bound list of applications to fail and a special class of the anti-social where an application cannot exist with additional instances of itself (Pschitzo). Having created a classification model, we can do a couple of things. The first is to design behaviours for the infrastructure to cope with the good and non-social citizens (by building Clubs, Gaols and Asylums) and and the second is to understand the nature of the non-social behaviour. This is usually a poor and/or exclusive use of scarce resources. If we could list these scarce resources, then applications developers could aim at writing applications with “Good” citizenship and/or infrastructure designers understand if they need a rationing or containerisation response to the applications. Sadly it has not been well attended.
Originally posted on my Sub/Oracle blog, republished in Feb ’16.