This is a transcript of today’s XWord Info Press Conference. Unfortunately, due to technical issues, some of the recording has been garbled and the text is unrecoverable. We regret the error.
JH: Thank you for that kind introduction. Please, sit down, everyone. Yes, I know, I know, please sit. Thank you. Really, thank you. It's great to have a chance to address you again as XWord Info reaches yet another milestone. I'll go over all the usual details and answer all the questions you've submitted in advance and, yes, there will be a chance for additional questions at the end with the usual caveats cited in your press package. There have been many surprises this year and once again I'll give you my perspective on what that means about the current state of crosswords and the industry and where I believe we're going given new trends in newspapers, in software, in cloud services, and in mobile devices. The most fascinating and utterly surprising change has been how quickly, basically over the past few months, constructors, editors and solvers have...
At this point, a portion of the 26-minute initial remarks, as well as the first few questions, have been lost. We pick it up as the fourth question is being answered.
JH: ...some over 40,000 times with one IP address responsible for nearly 70,000 word look-ups. For reasons I mentioned earlier, I never expected that. The ability to search crosswords, Variety puzzles and a dictionary at the same time turns out to be valuable.
Q: How can you insist that XWord Info is not about crossword statistics when it so clearly is?
Let me try to explain this another way. XWord Info is based on a database so, yes, there are lots of statistics that are easily available and I expose them through the UI. I happen to think many are interesting but you're free to ignore them if you disagree. The stated goal of the site, though, "to celebrate NYT crosswords and the people who create them," has nothing to do with stats. I try to have the most accurate digital representation of the crosswords, even the non-standard ones, in the belief that the historical record is worth preserving. That's why I go to the trouble of making sure that puzzles with special gimmicks — writing outside the boxes, answers that turn corners or go backwards, images embedded in the grids, etc. — are properly displayed. All rebus entries are carefully untangled. All non-linear puzzles have correct answers. I think of myself a curator of NYT crossword art.
Q: Then why so many pages about the most Qs and fewest blocks and so on?
I happen to find these interesting. Bloggers might celebrate a new record, might mention it in passing, or might pointedly ignore it. Their job is to express opinions. My site is different. It presents the data dispassionately. The few times I editorialize are clearly marked as JNotes. If you're curious about the records, XWord Info makes it easy to find them. That kind of searching is tougher in blogs.
Q: Do you really get so much hate mail about XWord Info? That seems incredible.
That's not quite what I said. I expressed surprise at the number of people who seem to go out of their way to tell me they never use XWord Info, or who want to inform me that they've heard second hand that someone else hates it. Complaints seem to fall into three buckets:
- The presumed focus on stats encourages pangrams and stacks and other supposed construction sins.
- Making the data available to the public puts too much stress on absolute originality. Who cares if there was a similar theme in 1994? Without databases, nobody would notice.
- Keeping even some of XWord Info outside a pay wall steals search traffic from the blogs so instead of getting thoughtful analysis and a chance to participate in a community of like-minded enthusiasts, searchers see the raw answers and go on their way. They’d rather that none of the pages were free. I do sympathize with this one, which is part of the reason I provide links to blogs. I hope to make the blogs more successful.
Q: How many of the top editors, constructors, and bloggers have XWord Info memberships?
I don’t give out any information on customers, sorry.
Q: What’s the point of XWord Info given that Cruciverb.com and several other databases already exist?
This question comes up every year. There are two main differences between Cruciverb and XWord Info. First, Cruciverb caters mostly to constructors and I mostly target solvers. There is a lot of overlap, of course, but Cruciverb has great value in many areas I don’t try to replicate. Kevin provides tips for constructors, help with grid design, identification of thematic answers, forum and email support for the community of professional and hobbyist puzzle creators, and on and on. Second, Cruciverb contains much more data since every puzzle from every major publisher is included. XWord Info focuses exclusively on NYT crosswords, including many Variety puzzles which I don’t think are found in any other database. I do have some features just for constructors like the Finder Page and Crossword Analysis but most pages are for fans.
Q: So your value is that you make the puzzles, even odd ones, look pretty?
I suppose that’s part of it but there’s more. I show full clues and answers. I don’t think anyone else tracks lists of rebus puzzles or Schrödinger puzzles or pretty grids or asymmetric ones or grids you draw on and several other categories. Don’t forget the Variety puzzles. There is a lot of great creativity there, including many of the best clues.
Q: What unique feature are you most proud of?
I think I have the only database that carefully tracks info by constructor. This is tricky since names can change over time or have alternate forms. For each Shortz-era constructor, I have lists of puzzles with thumbnails, lists of words debuted by that constructor, lifetime Scrabble average, etc. My absolute favorite page is Across Lite links by constructor. It’s an easy way to print out a set of older puzzles you may have missed by whomever your favorite constructor might be.
Q: What’s your role in the Pre-Shortzian Puzzle Project?
I sit on the Board of Advisors. Will Shortz and Stan Newman are Editorial Advisors and I’m the Technical Advisor. As new, I mean old, puzzles get litzed and proofed, I host them on XWord Info. This exactly fits my goal of preserving and showcasing the art of NYT crosswords.
Q: You famously have a plugin-free Acrostic solver that works far better than the one on the NYT website. Why don’t you make it available to them?
I created it as a proof of concept for the Times. Turns out, they weren’t interested. Someone else? Yes, you at the back.
Q: What’s the single most often asked question you get asked about XWord Info?
Why don’t I make my Acrostic solver available to the New York Times.
Q: You also have a regular crossword solver written in HTML 5, right? Why can’t I use that to solve NYT puzzles?
That solver is available for constructors. When they Analyze a puzzle they wrote, they can create a link for test solvers or for the public that uses my solver. I can’t make it available for published puzzles because I don’t have permission to do so. That makes sense because NYT charges for electronic puzzle access. They finally have an HTML 5 solver of their own which is not too bad. It’s missing some important features and they haven’t figured out the perf issues yet but they’re learning to code modern websites and they’ll get there over time. They’re also helping to wean the world from the evils of Java which is a wonderful thing.
Q: What’s the point of all your gimmicks like the thumbnail tooltips?
I suppose I think they’re useful or fun. Not all browsers can show the tooltips but modern browsers do. I use tooltips to show thumbnails of the grids if you hover over dates, and constructor photos when I have them. I also use them in the grids so you can hover over squares to see the associated Across and Down clues. Some of this was just to investigate some modern programming techniques I wanted to explore for my day job. I provide ways to turn them off if they bug you. None of them work on an iPad which doesn’t handle hover states at all. Microsoft Surface tablets work fine.
Q: How do you get constructors to send you photos?
I ask for them. The nice ones oblige. I’m grateful to all who do.
Q: You spoke in your opening remarks about the future of XWord Info and some of it sounds intriguing. Can you provide some hints on the timeframe involved?
Sorry, no. We have time for one last question. Yes, you. Go ahead.
Q: You mentioned that you do analytics on your web traffic. Where do most of your hits come from?
Depends on how you look at the data but I find the breakdown by top cities interesting. Ones near the top include New York City, Chicago, Ottawa, Ashburn Virginia, Brooklyn, Minneapolis, Miami, Los Angeles, San Francisco, San Diego, Saint Paul, Toronto, Vancouver, Mountain View, Montreal, Beijing (!), Calgary and Victoria. I’m not sure what it means that so many Canadian cities are on the list. Ok, we’re out of time. Thanks everyone. Feel free to use the comments below for additional follow up.