Lars,
I don't actually have a problem :) Either way will work, I just was wondering what OpenACS developers thought about this.
Generally, I would pick an easy solution. I don't want things to get too complex. What I do want to do is not have so much unused code. OpenACS is complex enough as it is.
Another issue would be if the dav packages were not distributed as part of OpenACS. In this case, they probably will be added to OpenACS. I can imagine an instance where optional features could be added from packages that are not part of OpenACS, so deciding how this might work is a good idea.
