Good question. First, we need to recognize that PaaS offerings vary greatly in levels of lockin. Generally, lockin is higher when there is significant deviation from what is generally accepted by those in the industry (e.g. lock-in is very high when a new language is introduced whose execution scope is isolated to the PaaS that owns the language). Given that, there isn’t a correlation between using a PaaS offering and some dire lock-in scenario.

Second, given what I’ve described, there is always some degree of lockin using PaaS. The question becomes whether the lock-in risk is significantly outweighed by the large value delivered by PaaS concepts. PaaS based solutions will generally experience huge savings in time to market, project costs, maintenance costs etc.

A good analogy is the operating system. Could you write software that is started at boot with a PC so that you don’t experience OS lockin? Sure, but then you spend enormous amount of time writing drivers and hardware communication, and lose out on the abstraction the OS created. Clearly, this is a purposefully trivial example but it makes the point that a layer like a PaaS offering can be extremely beneficial to those writing apps.

I would say it depends which kind of platform you are considering.

PaaS could mean technical integration solving technical redundancies such as single login screen for several solutions, … But it could also mean complete integration in terms of look & feel, naviguation, …

My opinion is that users will ultimatly benefit from a platform if that one is answering different business needs in a coherent and uniform environment.