This is my 1000th writeup! Hooray!

Looks like your 996th to me.

Or your 1007th, if you want to count pmdevnotes, which I don't think should be counted. I'd guess that some of the code disagrees with me on this point but hasn't always (though the -7 difference could just be the node cache race condition in action -- if you haven't stumbitted user settings in quite a long time). And I'm not counting patches.

Update: I update my theory to thinking that some code agrees with me and some disagrees and that you stumbitted user settings approximately 4 pmdevnotes ago. (And that's without even looking at the code.)