Bob Jackman's 'Word Builder' book ( http://members.ozemail.com.au/ ~rjackman/wordbuilder/title.html ) differentiates between 'familiar' as less well-known words, and the familiar set could form a useful lexicon (as, indeed, could the unfamiliar set).
Of course, exactly what is familiar varies, but it appears to me to be a well-chosen list and likely to have fewer idiosyncrasies than a list generated using an algorithm from a larger lexicon (like the TANKBUSTER example), because the extra work of checking the individual words has already been done. A couple of caveats are that it would still be some work to generate a text file if there isn't one already, and there is the possible issue of copyright if the list is for other than personal use. Also, the selection does not go up to the 15s (but it does include all 2-8 and quite a few 9s), although presumably most parlor games do not feature large numbers of words this long. Nick Ball On Nov 30, 2008, at 2:52 AM, tinytim1234uk wrote: > Quackle uses many unusual words in its plays thanks to its > comprehensive lexica. I found myself wanting to use a restricted > lexicon that made me feel like I was playing a parlour game rather > than a tournament game. > > As such, I set out to create my own, more restricted, lexicon. I ran > into two challenges: > > - Technical: building a .dawg lexicon file. > > - Editorial: Which words to choose? > > I have made some progress on the technical challenge. I took the > source code for makeminidawg and produced a single makeminidawg.cpp > module that will build as a standalone command-line application > without the need for the rest of Quackle or Qt. I have built it on > Linux, Windows and Mac OS X. I hope that this will help people create > their own customized lexica. > > The editorial challenge is trickier and arguably unsolvable. I > produced a list of 63,012 words by taking the intersection of > > - Sowpods; and > > - A list I found of all the words in the British National Corpus that > appear six or more times. > > The result is not entirely satisfactory; for example, TANKBUSTERS is > in the list but not TANKBUSTER. However, I find that I personally > prefer playing with this list so that when Quackle thrashes me I know > most of the words it has used. > > I have put a tarball on my website at http://www.nihilist.org.uk > containing: > > - The restricted lexicon; > > - The list of words that I ran through makeminidawg to create the > restricted lexicon; > > - A Windows executable version of makeminidawg; > > - The source code for said makeminidawg (as per the GNU GPLv2); > > - A readme file explaining the contents; > > - A sample batch file for building the Windows executable. > > Enjoy! > > -- Nihilist >
