Why not use LDAP?

Protocols are most successful when focused on initial design goals. LDAP was designed for global, authoratative user directory, not ACAP/IMSP design goals.
Protocols get more complicated and less implementable as features are tacked on. We don’t want to contribute to creeping featurism in LDAP as it is already useful in email world to solve the global authoratative user directory problem.
LDAP doesn’t have experimental background for addressing the requirements of ACAP/IMSP. The LDAP model is sufficiently different from IMSP that it could create unforseen problems.
Deployed LDAPv2 infrastructure is missing fundamental ACAP requirements. Upgrading a deployed and much needed infrastructure is far more expensive then adding a new protocol.
LDAP’s binary X.500 based model makes the protocol difficult to debug. The general application configuration problem needs a simpler protocol if it’s going to end up in many applications.