(arjen) While I cannot make it to the meeting, I do want to raise a topic I found rather intriguing: the use of custom data structures that Brian uses in the implementation of TIP #629 , the new [lseq] command. What can we do to make such data structures more convenient?