From xiao_dong2001 at hotmail.com Wed Mar 14 10:41:35 2007 From: xiao_dong2001 at hotmail.com (Xiao Dong) Date: Wed Mar 14 10:10:00 2007 Subject: [Mead] Error in running MEAD on GA3 Message-ID: An HTML attachment was scrubbed... URL: http://lists.si.umich.edu/pipermail/mead/attachments/20070314/689cfc71/attachment.htm From kaustubhp_in at yahoo.com Tue Mar 20 05:39:15 2007 From: kaustubhp_in at yahoo.com (Kaustubh Patil) Date: Tue Mar 20 05:06:00 2007 Subject: [Mead] LexRank threshold Message-ID: <269663.84672.qm@web61125.mail.yahoo.com> Hello, I am using MEAD-3.10. I want to perform some simple experiments with LexRank by changing thresholkd value. Is the -jump option equivalent to threshold (the one in the JAIR article)? Thank you and regards, Kaustubh Patil --------------------------------- Get your own web address. Have a HUGE year through Yahoo! Small Business. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.si.umich.edu/pipermail/mead/attachments/20070320/8ecffb70/attachment.htm From Sebastian.Delachica at Colorado.EDU Sun Mar 25 22:53:46 2007 From: Sebastian.Delachica at Colorado.EDU (Sebastian Dela Chica) Date: Sun Mar 25 22:20:11 2007 Subject: [Mead] Follow-Up on CIDR and nidf files issue Message-ID: <460743CA.3090301@colorado.edu> Files: cidr.pl, nidf.pag, nidf.dir I realize it may be seen as poor form to follow up my own post, but thought I'd share my perspective on this issue that seems to keep rearing its head at different points in time and in slightly different shapes within the MEAD community... OK and I really need to get CIDR working very soon for my research to move forward at a reasonable pace :-) Based on perl documentation on dbmopen/DBM and the like, it would seem the whole idea of distributing .pag and .dir files can lead to problems across platforms as much depends on how the perl binary was built on the final target machine. So while these .pag and .dir files may work well on several machines for cidr.pl, it is just a matter of time until one runs into a perl binary that does not quite jibe with the distributed files (such as mine and others reported in this mailing list). Honestly I lack the time to dig much deeper from the perl binary build process end, but it seems to me from the perl docs that there are no guarantees of cross-platform compatibility for .pag and .dir files, regardless of whether one uses dbmopen or tie. Given this (and assuming I am right, perhaps someone would be willing to comment here?), would it be possible to obtain the .txt version of this nidf database and provide a utility to create the .pag and .dir files on the target machine? The advantage is that once the .pag and .dir files are created on a target machine by its perl executable, they should work. I am willing to do the work and provide the utility back to the community: it should be fairly straightforward assuming the nidf.txt file looks something like the enidf.txt file distributed in the $MEAD_HOME/etc directory. At least would it be possible to get a better idea for the provenance of the perl binary used in the development of the cidr.pl program? Of course if I am totally off track, or if there is a simpler solution please do let me know. Thanks for your time, Seb -- Sebastian de la Chica, Research Assistant Institute of Cognitive Science Department of Computer Science University of Colorado at Boulder http://ucsu.colorado.edu/~delachic