I view it as a shame that I haven't encountered more instances of prose in this (scientific/technical) context. I think it adds to the 'quality' of the piece. From a selfish point of view I would enjoy if I encountered more of this style of writing while being introduced to topics. Although, I'm also not contributing any writing of any quality to any corpus. So there is dissonance there.

As a complete aside, I thought your writing here http://canonical.org/~kragen/memory-models/ was helpful. And I now know about BNF!

I agree about the style. I often find higher-quality prose in older papers, prior to 01950 and especially prior to 01900.

Thank you very much!