This is where it's all happening at the moment. The new engine is playing; some early releases have been made and it's already as good as the legacy version. Some general plans for the engine:

• Fast, bitboard-based infrastructure. The legacy Deep Brew is pretty fast, but the new one is shaping up to be significantly quicker.
• More sophisticated checkers knowledge. I've acquired a bunch of domain knowledge since I wrote the legacy version; the new eval will be significantly more subtle, and (hopefully) won't slow things down too much. End result: much stronger play.
• Clean hashtable implementation. The legacy version has a rather baroque hashtable implementation which seems slightly broken. The new one will work :-)
• Opening library - this won't be in the first release, but I don't know that a draughts program can be world-class without one.
• Endgame tablebases - if there are GPL tools to build them, we'll use them; otherwise we'll write our own. Obviously this is a little way over the horizon, but this is another area essential to top-level competition.
• Portable - implementation on Windows, Linux, Palm etc