My first pass at thoughts about the future:
1. Keep up with FreeBSD (base the next release on 6, etc.)
2. I agree a core/module approach would work well and improve some
current issues and allow for future growth.
3. PHP and Java aren't well suited to the m0n0wall tenets (small, fast,
clean), in my opinion. Java is slow and overly complex, PHP is not
complex enough. I will refrain from starting a programming holy war but
humbly offer Perl/python as suitable core bases.
4. IDS plugin's or a modular IDS solution should be part of the larger
5. Fixes for general issues that plague me personally, like slow
recovery of IPSEC tunnels, etc.
6. Revamp the tunnel/partition terminology and implementation to make it
simpler to decide what you want and how to implement it (do I want a
partition or a b/w cap, etc?) I hate to say "Wizard" but it couldn't
7. Keep up the excellent level of quality and durability that has made
me such a big fan of m0n0wall in the first place.