I've been trying to get Dr. Malewicz up to speed on my research, and he has responded with a position that is common ammong many computer scientists and mathematicians, namely the mistaken assumption that information density is the primary feature of language, rather than semantic articulation. I'll lay in some problems with this view here:

• Redundancy - The simplest problem is error recovery, high density messages have, definitionaly, low redundancy, which makes error recovery difficult, and can make error detection equally difficult.
• Articulation - Articulated languages, those with internal multi-level structure, provide two extremely valuable features for comunicators:
  • Partial Evaluation - Articulated messages are often amenable to partial evalutaion, wherein semantic extraction can be performed on parts of the message at a time, and the totality of the meaning can be built up with processing costs proportional to part costs. Non articulated systems may have arbitrary mappings, producing situations in which all messages have the same, extremely expensive, evalution costs.
  • Incremental Construction - Articulated messages are amenable to incremental construction, wherein the message is constructed, and re-written, part by part, with the construction costs limited by the complexity of a given part, not by the complexity of the entire message.