Open source geospatial systems and tools have really taken off in recent years, as has the open source community at large. While an open source geospatial solution may be enticing to some GIS managers, there remain a number of hurdles for them to be effectively employed.
Chief among these hurdles are the lack of documentation, training, and the added difficulties of configuring or extending open source solutions (it’s tough to get a good ArcSDE administrator, it’s tougher still to get a good PostGIS administrator). In addition to this, while open source is starting to provide a rather comprehensive suite of software that includes desktop and web based GIS tools and services, as well as database storage and support, they still tend to lack a good data management framework. I suspect that the reason for this is because developers of open source spatial software want their work to be able to do something relatively fast. So they tend to make them work on existing GIS data (such as shapefiles) rather then put a heavy emphasis on creating and managing spatial data from scratch.
Most GIS may be reluctant to move too quickly towards an open source solution when they already have a functionally proprietary system in place. Even though moving towards open source may be compelling due to the potential cost savings in licensing fees, those savings are offset by concerns over being able to manage the system and maintain current capabilities.
The most likely scenario in the near future is to start seeing the appearance of hybrid solutions.
As an example, a GIS shop may decide that their use of ArcEditor (or ArcInfo), ArcSDE and ArcIMS (or ArcServer) may be overkill for their specific requirements. While this software provides an enormous amount of capability, this may not actually be required. It may be possible for them to effectively manage their spatial data using an ArcMap license to manage the spatial data, and then have that data pushed into a PostGIS database for corporate or enterprise attribute editing and visualization. The key is to leverage the power of both commercial and open source solutions towards cheaper and oftentimes more powerful solution.