Finder Philosophy

The XWord Info Finder page helps constructors create better crosswords. It’s the most popular page on our site, but also the most daunting. Fortunately, 90% of what constructors need can be accomplished with the standard Search button using only letters, question marks, and asterisks.

Learn a few more concepts like $c, $v, square brackets, and letter ranges (still using the standard Search button), and you cover 95% of cases. Obviously, I just made those numbers up, but they feel close.

So, what’s the point of the OneLook button, or RegEx, or all the stuff hidden under “More…”, or Python? None are ever required, but they can be handy when attempting certain theme types. Plus (and this is a personal note that may not apply to you), learning this is fun!

What’s the big diff?

The three main buttons, Search, OneLook, and RegEx, accomplish similar goals. They answer questions about how a grid might be filled, and (unlike comparable sites) they provide those answers within the context of what we know about crosswords. We know a lot about crosswords. We know how frequently and how recently words have been used in the NYT, and how Jeff Chen has scored them. We provide links to every clue ever used, and links to the puzzles that used them.

Let’s do a high-level comparison of those three buttons. (See Advanced help for details.)

Search

Search is the simplest to use. It scans through every NYT crossword ever published, Jeff Chen’s word list, and a few dictionaries.

There are limitations. Search can’t look beyond our internal (large but finite) database. Our lists don’t know about word breaks, so it’s not great for finding phrases

OneLook

OneLook uses only external data from the whole internet. It does tricks Search can’t, like requiring or excluding letters you specify. It knows about phrases, so searches with spaces work. Its Artificial Intelligence engine makes good guesses about meanings, so it can search within categories.

It’s not particularly focused on crosswords, so it can be random, and each query is limited to 1,000 results.

Regular Expressions (RegEx)

RegEx is the most complicated option, but the most powerful. It uses the same internal database as Search. It can find palindromes, or letter patterns needed for special themes.

The main downside is that it’s tricky to learn.

Why is Regex now restricted and what’s the workaround?

When I first introduced RegEx to XWord Info, I didn’t know if anyone would use it. It has become popular (you’re a bunch of nerds!) and the queries people try have become increasingly complex. So complex, that some were bogging down our website, frustrating everyone who uses it. That also hurts our Google results because search engines think our site is too slow to be interesting. So now, super-complex queries time out, and everyone is happy except for the poor constructor working on some amazing creation.

But I have good news. If you’re sophisticated enough to build up these arcane RegEx queries, you’re clever enough to run them on your own computer so our web servers aren’t impacted. You’d need a simple Python script where you could enter your super-amazing RegEx query and see the results.

We have such a script! See example 4 here.

This will run whatever query you like against one of our word lists or even against one of your own.

PS. Don’t worry. Only a teeny tiny fraction of RegEx queries are affected. Almost all that are, can be tweaked to make the syntax simpler, or be broken down into two or more different queries.

Otherwise, crank up a Python environment. The free one here works great. And have fun!