On 2/22/07, chicao74 <[EMAIL PROTECTED]> wrote: > Hello, > > First of all, thanks for Quackle. > > I've noticed how big the gaddag files are. So, I've looked and hacked > the makeminidawg.cpp into a makeminigaddag.cpp and ran it on the > gadddag. The result a file of about 4M instead of 16M. I had to change > the code in gaddag.h because of negative indexing, but it all seems to > work fine. Is this on any interest?
Francisco, Absolutely. Please send us your changes and we can run some tests to see if the output is identical. This is something we have meant to do for a while, but not gotten around to. > Second, when anagram is called on the first move, it calls spit() > which generates all possible words without considering the player > rack. Any chance that can be improved? I've had code I wrote on a tool > that used DAWG to build works on a list of characters, including > blank, so I know it is possible. The spit() in generator.cpp traverses the DAWG recursively, so it might appear that it is outputting the entire dictionary, but its search is restricted by the quantity of each tile held by the player. If I remember, this is done with an array called count[]. John O'Laughlin
