Dear Jonathan,

A long time ago you wrote me about a new planarity algorithm that
I'd presented at SODA in 1999.  At the time I could not take you 
up on the offer to present the algorithm to the DDJ audience, though
I very much wanted to.

The key issues included finishing the dissertation and not 
preempting a refereed journal publication.  Although I graduated
in 2001, the journal paper has taken quite a bit longer, but the
cause of theoretical perspicuity has final prevailed (this is my
term for applying the tenets of software engineering to the 
valuation of algorithms).

I recently saw a quote from Dykstra that said "Computer science
is as much about computers as astronomy is about telescopes."
I respectfully disagree; the analogy doesn't work because 
astronomy is not called telescopics!  Computer science can get
as mathematical and algorithmically intense as one likes, but
the mission statement of computer science should be to develop
methods for the automated solution of problems.  As the editor of
DDJ, I'm betting you couldn't agree more, and I'm very excited
by the prospect of a DDJ "homecoming" for my planarity algorithm.
Hopefully, I'll be just in time for the April Algorithms issue!

To be honest, there was a third reason I couldn't send you a
paper on planarity after the SODA conference.  Basically, my 
supervisor and I had developed an initial algorithm that was simpler 
than the almost impossible prior methods.  However, that method was 
still far more complex than my developer sensibilities could allow, 
so I spent six months rearchitecting it.  But my supervisor was
unwilling to go with this version for the SODA submission even
though I had implemented it because the deadline was less a 
month away when I broke the news to her.  It has turned out to
be a mistake, but I cannot blame her for wanting to play it 
safe, and I should have showed her what I was up to sooner.

Anyway, the fact remains that a DDJ paper on the algorithm
presented to SODA would have been impossible because it was
too hard and also didn't reflect my implementation.  Since then
my own implementation has matured substantially, and its
exposition is simpler than anything out there.  Moreover, the
readership will benefit from finding out about a number of
independent implementations in other languages.  None of
this would have been possible previously either.

So I think it has turned out for the best, and I think you
will be able to make no small amount of marketing brouhaha
over the presentation of the world's fastest and simplest
linear time planarity algorithms.

Please let me know if you accept this submission, and what
steps you'd like to take to proceed.

Thank you,
John Boyer
