Unfortunately, readability is at odds with other, more important goals of the authors, like shoe-horning in references to all important papers in the field, especially if they're likely to be written by reviewers.

Also, a lot of work must be shown to demonstrate completeness and rigorousness, and that doesn't help comprehension either.

The trick is to read only the segments of the paper that are actually worth reading, which is often only the abstract, conclusion and maybe implementation details.