Paul Erdõs (1913-1996) was an eccentric, hungarian mathematician. With more than 1500 published papers, he's one of the most prolific scientists ever. According to Wikipedia he "devoted his waking hours to mathematics, even into his later years—indeed, his death came only hours after solving a geometry problem in a conference in Warsaw.".
He preferred to work together with other mathematicians, more than 500 coauthors. He basically lived in his suitcase, traveling between scientific conferences, universities and the homes of colleagues all over the world, always working. Hence "Another roof, another proof".
His extensive collaboration is the background for the famous Erdõs number. Any of those 500+ who published a paper with him has an Erdõs number of 1 . Anyone who has published a paper with a number 1 author (themselves not one of them), is assigned number 2. Number 3 is given to everyone having published with a number 2, and so on.
I guess that most people having published something with a collaborator in the natural sciences have a finite Erdõs number.
There are similar numbers out there. We have for instance the Kevin Bacon number, describing similar professional relations between actors.
pibbur who claims to be pibbur number 0.
PS. He created a rather peculiar vocabulary for himself. For instance, he used to refer to children (he had none himself) as epsilons, after the symbol ε typically used in calculus to denote arbitrarily small positive quantities, for instance in the definition of limits. DS.
PPS. Now you also know why mathematicians may find "let ε<0" funny. DS.
PPPS. Come to think of it, they might laugh even more if I proposed "let |ε|<0". Except if I wrote it during an exam. DS.