Paper compression

Our paper is just about ready for Wednesday's conference deadline. This might be the earliest I've ever ‘finished’ a paper – is it ever really finished? – nearly 36 hours before the deadline. :)

Of course, 5 hours ago it was a page and a half too long, so I started applying my time-tested ‘paper compression’ techniques:

  • reduce the vertical space taken up by figures containing code and examples, even if that means abandoning customary line-breaking and indentation,
  • look for paragraphs with just one or two words on the last line, and rewrite them to make them a line shorter,
  • hunt down bibliographic entries that go on too long, and shorten Transactions to Trans., November to Nov., etc.
It's amazing how much mileage you can get out of these silly tricks. Now it's at the point where if certain key paragraphs need one more word inserted, the whole thing spills onto page 11.

This is a new field for me. I've read a fair bit, but still it's strange and scary to be submitting somewhere that I don't recognize any of the names on the program committee. (No offense folks, if any referees look me up and read this!) I had an idea that was vaguely related to compiler implementation, but actually could be generalized to a different field, and that seemed the best place for it. Plus this work was far more accessible to my M.S. students – one of them is my co-author – than any of my work on type theory. There's something to be said for that. I don't think I'll post any version of the paper here until I hear whether it is accepted, but for the curious I put a poster online some time ago – with very preliminary data – that describes much of the result.

©20022015 Christopher League