Not just you, me too:-D. I can understand (and do share) that prickliness. When some holier-than-thou purist starts ranting against NULLs (often at the same time as claiming to be a follower of Codd, which is just bloody :crazy: ridiculous) and telling people that they shouldn't need a name for bridge tables (or MM tables or whatever you want to call them) because all tables are equal 🙁 I begin to see red, so I thought your comment let David Portas off far too lightly the other day.

I also find it a bit irritating that MS's SQL team don't bite the bullet and provide full support for the IEEE floating point standard; I know it isn't trivial (probably needs a NOT NaN constraint on a column which is to be used in an index, with errors caused for trying to put a NaN there just as there are for trying to put NULL in a column with a NOT NULL constraint) but it really would make that part of arithmetic a bit cleaner. Another good thing about it is that it would probably annoy the relational purists - effectively we'd be adding a domain-specific NULL, which might to the final end to any hope of getting a "pure" NULL-free relational model accepted. But MS's failure to provide that is a small thing compared to what the discontinuation of NULL would do.