Most of the Mathematica users beyond beginner level will be aware of the phenomenon of Packed Arrays. The user-accessible functions related to packed arrays live in the Developercontext and have been there for quite some time. I wonder what this means for their future use.

Is Developer more stable than Experimental? Can we expect this functionality to move into the kernel and what calling strategies could we adopt to minimize the effects such a move would have?

-

I don't think that Developer is going to go away; there is too much (also internal) stuff depending on it. What I do at the beginning of a package:

pack = DeveloperToPackedArray;
packedQ = DeveloperPackedArrayQ;


then only use those; if anything changes then it's only one place I need to change it. On the other hand a re-factoring: of DeveloperToPackagedArray is not the end of the world.

Concerning moving this to the kernel: unlikely. It has not been done in the past (for probably good reasons), why should it happen now. I don't think there is generally a way to classify stability of packages; if it is in the kernel, it is more stable than if it is in a package that you have to explicitly load.

-
What would be nice if the Developer package would get loaded by default. Then it would not matter if people type DeveloperToPackedArray or ToPackedArray. Just like JLink these days (which did not load by default on startup until Mathematica 6). –  Rolf Mertig Jan 24 '12 at 18:36
@RolfMertig, I'll mention that, but I have a feeling that this is not going to happen. –  user21 Jan 24 '12 at 20:06