Until recently, I was pen solver (with plenty of cross-outs). When my printer conked out a while back, I switched to digital and haven't gone back to paper.

I'm not sure there's any innovation I could come up with that some clever constructor hasn't already done. With all of the self-published crossword sites online now, a constructor doesn't need to worry what the public is ready for.

I don't get hate mail. If the WSJ mailroom gets any, they're nice enough not to share it with me. I do get occasional lukewarm-or-worse reviews on some of the puzzle blogs.

As an editor, I don't like an answer word in the grid to appear in a clue for a different answer. My test-solvers are very good at pointing those out.

Sort of a do-it-yourself crossword dictionary. Nice.

I wish I had a better answer, but no, no particular word I've been dying to include. And I don't keep track of debut words, so I guess I don't have one to be proud of.

Filling grids becomes easier--you get to recognize which areas need to be tackled first and what words will fit. Coming up with new themes gets harder, since you feel like you've seen everything.

I've noticed that too, but it is just coincidence. I don't collaborate with other editors to plant common words. (Though maybe there's a super-meta contest idea in that...)

Obviously there's often something hidden in the Friday contest puzzles. Other than those, I've probably hidden messages in puzzles in the past, though nothing comes to mind.

Yes, lots of those words that appear so often they're hard to clue in fresh ways. But they do glue the rest of the puzzle together.

We have occasionally Friday metas from Patrick Berry and Peter Gordon. But since I need to test-solve every meta to gauge its solvability, I'm need to be picky about those I'll invite.

The only words I can't include are the really offensive ones, and anything longer than the grid will fit.

Will and I are friends. We worked together for years at Games Magazine, and I'm a judge every year at his crossword tournament.

I'm not too worried. I've been doing this for decades now, so if it goes away I'll be ready to retire. Though I hope it lasts another decade or two.

Yeah, no favorites. Too hard to come up with fresh clues for those folks.

I didn't have those ideas until I got to college, where I made a crossword for the daily student paper. Then I thought it would be great to make a living making puzzles, but I knew that wasn't realistic. Years later, I'm still surprised.

Hi, Maddie here from the WSJ interrupting for a minute. Here's a link to Rainbow Connection: https://www.wsj.com/puzzles/crossword/20180629/41573/index.html?mod=wsjreddit

One of my early favorites was "Play with matches" for TENNIS.

At some point, I just thought "I wonder if I could make a puzzle." So I did. (The early results were pretty bad.) Editing is a lot like constructing, asking "How could this puzzle or clue be better?"

Get a copy of Patrick Berry's Crossword Constructor’s Handbook from his website. And look for the Facebook groups for new constructors.

If I consider it crosswordese, I'd be happy to never see it again. My current top nonfavorite word is STYE.

I solve a few crosswords. I do Matt Gaffney's Weekly Crossword Contest puzzles, and Peter Gordon's Fireball crosswords. I solve the Times puzzle sometimes, especially the unthemed Saturdays.

Many of them are. The key to making a good computer-generated crossword is to spend a lot of time tweaking the word list. And even after that, the computer can't write good, clever clues.

I'm not sure I have an absolute favorite. One of my favorite Friday contest puzzles was "Rainbow Connection." (I won't spoil it in case you can find it on the WSJ site.)

The first step is coming up with a theme, and finding theme answers of good lengths that balance each other. Then I place them in a grid, decide where black squares look promising, and then fill in the restof the words. Clues come last.