ICE compounds get ported by converting them to scripts / nodes in the other applications. Those who have built them should have the chops to get it done since at its core, it is just math and logic.
I'm 100% behind Raf here. Unless you're working at a studio where you need to continue to dev Softimage plugins in order to get current or up coming jobs out the door it doesn't make sense to continue to build plugins for Softimage only. Even in its prime, the 3rd party tools market for Softimage is minimal. If you can dev for multiple platforms at the same time that's going to be helpful as well. Mootz and his early test with Fabric is, in my opinion, the correct direction. Support your plugins for Softimage but diversify to others are the same time. For those who are replying and don't know how to script, I don't believe Raf's comments were for you. Its for those who are developers and TD's. It's a valid point that for these people to make a living, they need to transition faster than most. We already saw another thread where someone lost a dev job because they announced the end of Softimage. This is just the beginning. Fabric is one of the only options I see as a viable platform to continue with. Splice works in Softimage and Maya currently with Houdini and I believe Max coming within the year. This hybrid approach allows you to keep your work relevant while having a comfortable transition using Softimage during that time until you're fully comfortable in say Maya. I know it's not artistic friendly yet, but the only way to ensure that is to get more developers on board to push it there asap. Raf, myself, and many others have been hard core fans of Softimage for years. It's not like we're are giving up on a solution that is in its prime. It's on its last leg with a drop dead date of April 2016.