Building Uncle Logic's Daily Crossword: Harder Than It Looks
From Uncle's Workshop · Development Journal

A little while ago, we decided Uncle Logic needed a daily crossword.
Not a particularly unusual idea. Crosswords have been around for more than a century, millions of people enjoy them, and they seem like a natural fit for a website dedicated to puzzles.
So we got to work.
We needed a grid, some words, a bunch of clues, and a way for people to fill everything in.
How complicated could that be?
Well.
As it turns out, a little more complicated than we expected.
Making a crossword and making a good crossword
Here's something we've learned along the way: getting words to fit into a crossword grid is not the same thing as making a crossword people will actually enjoy solving.
You can fill a grid with perfectly valid dictionary words and still end up with a pretty terrible puzzle.
Maybe the answers are technically correct, but nobody has used half of them in an ordinary conversation since 1847.
Maybe the same three-letter words keep appearing.
Or perhaps you've created a puzzle where two obscure answers cross at a letter that nobody could reasonably figure out.
Congratulations. You've made a crossword.
We're just not sure anyone will thank you for it.
What we're trying to build is something better: puzzles with familiar vocabulary, interesting clues, sensible crossings, and enough variety that solving one feels different from solving the next.
And that's considerably harder than simply filling empty squares.
The vocabulary rabbit hole
One of the first challenges was building a useful collection of crossword answers.
At first, that sounds easy. English has an enormous number of words.
Surely we can just collect a dictionary and start fitting things together?
Unfortunately, dictionaries contain all sorts of things that aren't especially helpful in a general-audience crossword.
Obscure abbreviations. Unfamiliar technical terms. Archaic words. Proper names that might be recognizable to a handful of people but meaningless to everyone else.
And then there are perfectly ordinary words that become awkward when you try to write a fair crossword clue for them.
So we've spent a substantial amount of time researching, reviewing, and refining our vocabulary.
The goal isn't to have the largest possible word list.
It's to have a collection of answers that can help produce enjoyable puzzles.
That distinction has turned out to matter quite a bit.
Difficulty is more than choosing harder words
Like many Uncle Logic games, we're building our crossword around a weekly difficulty progression.
Monday should be approachable. The challenges should generally become more demanding as the week progresses, with Sunday offering its own accessible featured experience.
But making a crossword harder doesn't mean filling it with increasingly obscure vocabulary.
There's a difference between a difficult clue and an unfair one.
A clever clue might make you think about a familiar word in an unfamiliar way. You struggle for a moment, find a crossing letter, and suddenly everything clicks.
That's satisfying.
A clue that requires you to recognize an obscure historical abbreviation you've never encountered?
Not quite the same feeling.
We're trying to develop puzzles that reward reasoning, vocabulary, and those little moments of recognition.
And that means the clues need just as much attention as the grid itself.
Sometimes more.
The problem with the word "daily"
Here's where our original plan became especially ambitious.
We aren't building just one crossword.
We're building a daily crossword.
That means we need a steady supply of puzzles that meet the same standards.
A good puzzle today doesn't help very much if tomorrow's puzzle isn't ready. And having a dozen puzzles prepared doesn't necessarily mean you've developed a reliable process for producing hundreds more.
Each grid needs to be constructed, filled, clued, checked, and tested.
We also need to consider whether its difficulty actually fits the day it's intended for.
A puzzle that feels like a pleasant Monday warm-up to one person might feel like a Thursday challenge to someone else.
That's why human testing matters.
A computer can tell us whether a grid is valid. It can verify that answers fit, letters agree, and technical rules are satisfied.
What it can't reliably tell us is whether a person will enjoy solving it.
And that's a fairly important part of making a puzzle game.
So, where are we now?
We've made real progress.
The crossword-solving interface has been developed, and we've built and tested an initial collection of puzzles.
But we're still working on the part that will determine whether this can become a sustainable daily feature: producing enough consistently good puzzles, with the right difficulty and editorial quality, without treating each new crossword like an entirely separate development project.
That process has taken longer than we originally anticipated.
We could rush out what we have and figure out the rest afterward.
But we'd rather do the work now than introduce a new daily game and immediately struggle to keep it supplied with puzzles worth playing.
A daily crossword should become something people look forward to, not something we scramble to finish every morning.
We've already got enough puzzles around here that involve racing against a clock.
We don't need to make publishing one of them.
What's next for the crossword?
There's still work to do.
We're continuing to improve the construction process, refine our vocabulary, review clues, and calibrate difficulty.
We also want to make sure the experience is comfortable on phones, tablets, and computers. A crossword should be challenging because of its clues, not because you're fighting with the keyboard.
We don't have a public release date to announce just yet.
And rather than give you an optimistic date that we'll spend the next several weeks trying to defend, we'd prefer to share meaningful progress as we make it.
That's part of why we're starting From Uncle's Workshop.
We want to show you a little more of what goes into building Uncle Logic: the decisions, the experiments, the unexpected difficulties, and occasionally the things that sounded much easier when we first thought of them.
The daily crossword is a pretty good place to start.
Worth getting right
It's easy to underestimate something familiar.
Most of us have encountered crosswords for so long that we don't really think about the work involved in making one.
You open the puzzle, read a clue, fill in a few letters, and hopefully enjoy yourself.
That's exactly how it should feel.
The work behind the scenes shouldn't be the player's problem.
But now that we've spent some time on the other side of the grid, we have a much greater appreciation for the people who create these puzzles every day.
And we're determined to make something that deserves a place alongside them.
So the crossword is still coming.
It's just taking a little longer to get all the pieces to fit.
Which, now that we think about it, is a pretty appropriate problem to have when you're building a crossword.
— From Uncle's Workshop, at Uncle Logic
