With a vote of 5-0, the Alicebot and AIML Architecture Committee approved the removal of <if> from the AIML specification. The text of the 28 September 2001 motion by Noel Bush was:

This proposal reintroduces a topic that was perhaps a casualty of the parliamentary juggernaut.

Several points:

1. No one has articulated any function that *can* be performed by <if> that cannot be performed with <condition>.
2. Jon, Rich and others have both expressed support for the idea of eliminating <if>.
3. In this early stage of AIML's life, it doesn't make sense to be trapped by our own process, as Kim has pointed out.
4. With all due respect and deep gratitude for Tom and Pedro's ongoing effort to get Program D into good shape, it does not make sense to preserve <if> as extra fat on the spec "since it's already implemented anyway" any more than it would have made sense to preserve <typeof_xxx/> because it was already implemented. I do get the sense that the pressure to retain <if> is originating in that way, out of "guilt" that Tom & Pedro faithfully proceeded on it based on the spec rather than on a rational view of what's best for the user community and "posterity", so to speak. I think everyone can continue to show appreciation to Tom and Pedro and still make an honest correction to the spec.
5. So, take this as a formal proposal to remove <if> entirely. There's been significant discussion about this but it was basically ignored based on the fact that the motion to correct the form of <if> was already on the floor. I'd like to see us reverse that mistake before it's too late.