Listener No. 4385, PD: A Setters’ Blog by Dragon
Posted by Listen With Others on 6 March 2016
Perhaps established Listener setters make notes on the construction process as they go along, planning for their Fifteen-squared blogs. Dragon is a new team and didn’t. So memories of the earliest events and feelings may have become distorted or magnified. The quotations below may not be entirely accurate and some should be taken with a large splash of soda.
Dragon consists of Robert Whale (veteran Listener solver) and David Harry (rookie Listener setter), both enthusiastic Azed competitors (R and D hereafter). When D’s All Change by Dilwitch, was published in August 2013, it was widely criticised for its lack of a PDM – “Clever construction – But Where’s My PDM!!” [Abandoned in the struggle to actually get the grid constructed.] However, the comments gave him the idea of a producing a puzzle in which every word dropped d(s) or p(s) or both before entry into the grid — “I’ll give them Penny Dropping!!!” Printer’s Devilry is D’s favourite type of Azed Special and he’d been looking for an idea which lent itself to using it in a puzzle. The coincidence of the standard abbreviation for Printer’s Devilry being PD was too good to miss: Across clues could be PD, distracting from the real gimmick, and the Down answers would be pd-deficient. For an endgame, highlighting the hidden word NOM (i.e. no “M”) seemed appropriate.
This left D with a problem. He already had another project going, with its newly constructed grid awaiting clues. He is a very, very slow clue-writer. Creating for PD the necessary Sympathy dictionary (wordlist) of p&d-less words was going to take a lot of time as would writing a second set of clues. It seemed sensible to have the basic idea checked out and also see if he could unload some of the work. So he turned to R who had nobly test-solved for him on a number of occasions. Although R spends far less time on his Azed competition clues than D, he still gets much better results, so the puzzle would definitely benefit. D sent R the draft preamble asking if he liked the puzzle’s gimmick and if he would like to come on board and write the clues. The replies came: “Yes, though I don’t much like NOM” and “Yes, if you’re not in a hurry”. We were off.
The problem of finding an endgame was put to one side while the dictionaries were prepared. This entailed writing a little VBA macro to find and strip suitable words from Ross Beresford’s UKACD dictionary (and our thanks to him for GNU licensing it). The Down lights would be filled in from that dictionary. At some point it was realised that since PENNY-DROPPING itself contained both ps and ds, it would make an ideal word to be discovered. It could be put, unclued, in the central column of a 13×13 grid with unchecked cells above and below. Later came the thought of making each column represent a year with the central one as 1971, the year of decimalisation. In the columns before 1971, dropped pennies would be ds but not ps, the opposite afterwards and both letters would drop in the year in which both had represented pennies. [P]ENNY-[D]RO[PP]ING could sit symmetrically within the numbers of the year. Some tweaks to the macro produced two further dictionaries, exploiting Sympathy’s ability to allocate specific dictionaries to specific lights.
Some grid-fills on these lines were produced and sent to R. The central column contained [P]ENNY-[D]RO[PP]ING sitting between 19 and 71 with the numbers barred-off from the rest of the grid. This made the final deductive jump rather a big one, R felt. He suggested making the 1’s part of two Across answers to turn the large deductive step into two smaller ones and integrate the numbers into the rest of the puzzle. This worked well, though even with this help some solvers faltered at the final step.
Doing this spotlighted a previously unnoticed problem. The most attractive nine-letter word starting with ONE is ONEROUSLY. Could you do a reasonable PD clue to it? If you can there’s a free pint in it for you (or a glass of something else). The same consideration applied to all the Across answers, potentially. It was no use D presenting R with a filled grid, R writing lots of clues, then hitting an intractable Across and having to start afresh with a new grid-fill only to have the same thing happen again. D needed to prove each Across word was PD-able as he went, rejigging fresh grid-fills around the gradually enlarging clued sections. (Thanks here to Mark Owen for Qat – D would still be writing those clues without it.) D had also had the thought that, since each column represented a different year, ideally each column should show at least one P (left side) or D (right side) to indicate that they weren’t pennies in that particular year. This meant two more macro tweaks, two more dictionaries and many more refusals from the Autofill because of the further reduction in available vocabulary. But it proved possible.
Eventually a PD’d fill was fixed and it was over to R to edit D’s clues and produce the Downs. By July 2014 the puzzle was ready, test-solved and sent off. In due course some clue changes were requested, others were provided and D learned a new word – concision. Azed competition clues don’t have the same length and solvability constraints as clues for real crosswords and it showed in those submitted. Our thanks to Vetters Shane and Roger the size of whose input is not appreciated by solvers, and to all those who made appreciative comments by letter and on the Web.