What are Cliques?
 

Paul Leyland's recent email to the NFSNET mailing list described the second stage of post-sieving as clique removal. There was no mention of removing doubletons, so I'm guessing that clique removal is a generalization of the doubleton process. Can this be explained in a few paragraphs? If so, should it be linked into the NFSNET FAQ?

The paper "Strategies in filtering in the number field sieve" at
[url]http://db.cwi.nl/rapporten/index.php?persnr=533[/url] helps explain this concept.