Things that need to happen before I can use R and eventually Julia instead of SAS
1. Ability to process large amounts of data on disk via chunking methods
Once someone writes an R binding for Apache, it's the beginning of the end of my relationship SAS.
Currently SAS is often used as data manipulation even though I believe SAS originally stood for Statistics Analysis System. I have enormouse respect for the founders of SAS (who are billionaires) but I believe the world is moved and SAS's continuing relevance is due more to inertia than excellence.
Many of SAS products were painful to use including Enterprise Guide, the SAS language is painful to use. I had tried to stay away from SAS until a recent consulting project where the main tool is SAS. But I had to use SAS for some parts of the projects, but only the part that deals with large amounts of data.
I believe Microsoft has a proprietry xdf format that can work as in on-disk format but it's proprietry. Sigh.