The following is a project that makes heavy use of QueryData! -- We are happy to announce the availability of a free, open-source Perl package that implements...
Hi Jason, 1. Is the bug below fixed as yet? I don't recall seeing an announcement to the effect... ... 2. I need to be able to access the WordNet 1.7 data...
Hi Jason, Ignore question 1. I just saw changelog for version 1.18 that fixes the bug below. Thanks, and sorry for the question. WRT question 2, I understand...
... Fixed, as you figured out... ... Unix and Windows use different sequences of characters to separate lines ("\n" vs. "\r\n"), but I think perl will be smart...
I probably should have made these changes a while ago, but school has kept me busy, so I'm just getting around to it now. I'm going to make changes #1b, #2 &...
As promised, a new version (1.21) is available: http://www.ai.mit.edu/~jrennie/WordNet/ Any collocation that QueryData returns will use the underscore '_' to...
... I went through a couple versions, thinking each would be the one I'd really release. So, yeah, the web page had a few different version numbers over the...
QueryData.pm version 1.22 is released. Get it here: http://www.ai.mit.edu/~jrennie/WordNet/ I've changed Makefile.PL so that it will die if 1) WNHOME isn't...
... Thanks again! That sure does need to be fixed. I also noticed a similar bug in getWordPointers (passing $tgt to getWord without converting). Hopefully...
... Hi Jason, I had this loop running, looking for the hypernyms of a whole lot of synets. The code was breaking for two for them. Here are the queries for...
... Great! I've added a test for the hypernym of sunset#n#1 so that the same bug won't accidentally crop up in the future :) Version 1.23 has been released. ...
... Thanks for finding this. I wasn't parsing the source/target pointers correctly (more hex issues...) Anyway, I've released 1.24 and added a test to make...
... Ah :) So, querySense will only tell you about things that have entries in the data.* files. It will not try morphological modifications to find the right...
... My impression has been that the polysemy count is sometimes tied to the number of senses observed in the SEMCOR data (a manually sense tagged portion of...
... That's easy enough to do :) ... Yeah. I think it's kind of weird too. Certainly a good first approximation, but probably not much more. ... Oh, you might...
Thanks Sid, I think your clarification answers this point pretty well. I am not sure what I was thinking of in regards to links between WordNet and SEMCOR -...
Hi Jason, ... I tracked the warning Mona was getting -- "Attempt to free unreferenced scalar during global destruction" -- to QueryData. She gets this warning ...
... I tried running Mona's query (distance.pl -type res --trace --allsenses car bicycle) and didn't have any trouble. I used QueryData 1.24 and distance.pl...
... Hi Jason, She gets this distance 0.11 and with any program that runs QueryData. In fact, I was also unable to recreate the warnings on my machine. I found...
... I'd appreciate being sent copies of any scripts that make such warnings, especially small ones that might make it easier for me to isolate the problem. ......
... Perl 5 is the minimum requirement. It sounds like you are using a version of perl 5 that may have significant bugs (5.004). 5.004 was part of the 5.00x...
... equate ... they're ... I was doing a little reading recently that may explain somewhat why WordNet equates polsemy with familiarity. Apparently George Zipf...
Proper nouns in WordNet are capitalized (e.g., World War II), but all words in QueryData.pm seem to be converted to lower case. For my application, it would be...
... So, right now, QueryData lowercases everything. This isn't necessary. It is necessary to do the lowercase for the index.* data file lookup, but there's no...
Hi Jason and Peter (and anyone else listening in...) ... This sounds like a good idea. It's fairly easy for a QueryData user to lowercase things if we wish,...
... This solution would work, but I would prefer a new function, such as ProperNoun($word) (returns true/false) or Capitalization($word) (return capitalized...
... So, directly returning the capitalized versions of words would be cleaner because that's how they're stored in the WordNet database. It would also better...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
