Constructing => General Discussion => Topic started by: jpwguitar on March 13, 2014, 09:08:11 AM

Title: methods of crossword construction
Post by: jpwguitar on March 13, 2014, 09:08:11 AM
I feel that it is time in my life to dabble in crossword construction.  Before I begin, I'm interested to know if the use of software is considered a legitimate method for serious puzzlers.  What would Will Shortz say?  Should I stick to a blank 15 X 15 grid and a bale of pencils? 
Title: Re: methods of crossword construction
Post by: Pangram~Man on April 04, 2014, 03:24:59 PM
It's always good to start raw and then gradually work your way into the use of software. Using paper and pencil as a newbie may help you to see more of the insides of what crossword constructing is all about.

Title: Re: methods of crossword construction
Post by: duncanMKZ on May 02, 2014, 03:33:18 PM
And before buying a lawn mower, you should cut your lawn with a pair of scissors, so you can really get one-on-one with the grass and how it works.

Making up a crossword by hand is a difficult task. It's a good intellectual challenge - you have to think in two directions at once, considering not only the quality of the Across word, but also the likelihood of that word intersecting with valid Down words. It becomes sort of instinctive over time - but I'm not sure it's an instinct that functions as a useful skill if you use more automated methods. There's no virtue to it.

Back in 1976, my chemistry teacher insisted that we spend four weeks learning how to use the slide rule. Calculators were in widespread use by then, but she explained that calculators have batteries that run out. A slide rule never runs out. Over the years, my skill on the slide rule has been utterly useless to me. I'd say that doing crosswords by hand is the same.

Even when I was creating crosswords by hand decades ago, using specialized crossword dictionaries. I longed to find a way to automate the tedious aspects by getting my Apple II+ to do the hard work. I wrote programs for it, but creating the dictionaries was hard, and the computer was too slow to be much help. (I did create a damn fine Word Search puzzle creator, though.)

Today, the programs will do a wonderful job of helping you create a puzzle. You can use most software in a semi-automatic mode, where you enter words, and the software instantly offers potential fits. It allows the constructor to focus on the creative, dare I say, artistic side of things. Where it was once impressive to create a grid that worked at all, today the impressive accomplishments involve ingenious clues, or challenging words and grids. The use of software has made puzzles far more interesting, and lets constructors focus their energy on adding elements that make a puzzle more exciting for the solver.

Title: Re: methods of crossword construction
Post by: jorkel on June 04, 2014, 10:27:25 AM
Construction via software is more common than via manual means, but both are widely accepted with editors.  Considering that the grid requirements -- maximum word count, minimum theme letter count, maximum black letter count, etc. -- have not changed in over 20 years, one can safely assume that manual construction is not being discouraged.
Title: Re: methods of crossword construction
Post by: Thomps2525 on July 22, 2014, 09:59:43 PM
Hello, everyone! This is my first post...and wouldn't ya know I'd be asking a question instead of offering any helpful advice? But I'm asking because I have been unable to find the answer in any books or on any websites. As duncan points out, making a crossword puzzle by hand is very difficult. I've tried---and tried and tried and tried! I have a hard time coming up with criss-crossing five- or six-letter words and I certainly can't figure out how anyone can come up with three or four "stacked" fifteen-letter words which also make words vertically. If a puzzle maker wants to start with three fifteen-letter words or phrases, how does he determine which word to start with? And does he place the top word first or does he start with the second word? I can come up with thousands of fifteen-letter words and phrases---how about "cruciverbalists"?---but I can't understand how anyone can stack three or four so there are words both directions. Perhaps the people who make such puzzles are using a computer program...but still, how do they determine where to start?
Title: Re: methods of crossword construction
Post by: wbg on July 23, 2014, 09:02:00 AM
I have no idea how anyone could come up with three stacked 15-letter answers by hand.  I admire those who can do it, but I know I could not, except by some extraordinary piece of luck.  Brains are organized in different ways.  My late father-in-law had a list brain.  His wife would mention things they needed from the store over the course of the week.  He never wrote anything down.  On Saturday, he'd go to the store and buy everything.  He could rattle off the birthdays of everyone in the family.  Then there are chess brains, looking many moves ahead.  Card brains, people who can tell you who played what card for the whole game.  The doc on a ship I served on was a serious gin player.  He invariably knew what cards I was holding well before the end of the game.  NO ONE could beat him.  It seems to me that manually constructing crosswords is somewhat the same.  The skill can be honed, but you've either got the knack or you don't.

I've spent maybe 40 hours total messing with Crossword Compiler.  I like it.  Seems to me that the only sensible approach for those without the gift, like me, is to start with some theme answers, crosses you like, words you like, whatever, making liberal use of Autofill and cancel.  I think Nancy Salomon made the point that, if Autofill can't fill the grid, neither can you.  If some of what Autofill comes up with seems good, accept the fill, delete what you don't like, and keep going.  You do have to keep thinking about neat words that can be popped in as you go along, including words that are unlikely to be in any word list.

Those are just some meandering thoughts from a neophyte.
Title: Re: methods of crossword construction
Post by: jorkel on July 23, 2014, 01:40:35 PM
If you've given it the old college try -- constructing by hand -- and it just isn't working out, then it's time to buy software like Crossword Compiler.  Be advised that the word lists that come with software are not spectacular, so you may want to get onto the Cruciverb mail list and ask for someone to mentor you in crossword construction.  In the process of doing so, you'll probably gain access to a much better word list.
Title: Re: methods of crossword construction
Post by: Thomps2525 on July 25, 2014, 06:24:30 PM
wbg and jorkel recommend Crossword Compiler. Yeah, but then I'd be admitting that I'm not smart enough to compose a crossword all by my lil' self. Today's New York Times crossword has four 15-letter answers at the top and four more at the bottom. That means that there are 30 four-letter combinations which either begin or end an up-and-down word and I can never figure out how the puzzle creator can do that. I remember seeing a puzzle that had eleven 15-letter answers: four at the top, four at the bottom and three in the middle. I've noticed that the majority of the long answers contain many of the most common letters, such as E, S, R, T, D, A and I, and a lot of them are phrases where every other letter is a vowel.  Does anyone know if the puzzle maker starts with only one 15-letter word at the top and then fills in a few four-letter words at, say, 2-down and 5-down and 8-down and 12-down and then tries to come up with the second 15-letter across word? That might be easier than coming up with four 15-letter words to start with. Puzzles appear easy to construct but they really aren't. "Kids, don't try this at home!"
Title: Re: methods of crossword construction
Post by: wbg on July 26, 2014, 10:31:24 AM
Another factor in the mix is a person's attitude toward computers, maybe.  I like them; bought an original IBM PC in late 1982.  The full-power version.  64K of RAM, not just one floppy drive but two, each with a capacity of 160K!  With a small monochrome (green) monitor and a dot matrix printer, the setup cost around $4000.  Had at least one ever since.  So I enjoy using the software.

If you enjoy constructing by hand, great!  Do it!  If you like doing it using software, do that.  It doesn't seem like a question of moral superiority to me.  At the end, when you decide the grid is done, it is what it is and it doesn't matter how it was created.  (I remember seeing a book of sudoku puzzles that proudly proclaimed on the cover that the puzzles were hand-crafted by Japanese masters.  Sheesh!  A bunch of numbers!)  Then, of course, it has to be clued, and there, it seems to me, software is much less useful.