[10:10:55] <gehel>	 lunch
[10:27:26] <dcausse>	 lunch 2
[12:53:11] <pfischer>	 dcausse: I ran the hive query regarding deduplication and refined it to compare file_size, too. As a result, there are 2014 files in en wiki and commons that match in title and file_size.
[13:00:16] <dcausse>	 pfischer: out of curiosity, why would you be interested in file_size being the same?
[13:12:40] <pfischer>	 Well, I wanted to narrow the criteria of duplicate. An identical file size + title would be a better indicator for one and the same file in two different places. By just looking at the file name, I might be two completely different things that accidentally share the same name.
[13:15:36] <dcausse>	 pfischer: the current index time behavior that we'd like to remove does only take the page title into consideration
[14:12:56] <pfischer>	 I know and I’m trying to challenge this approach. Wouldn’t we risk hiding valid search results by just looking at the title?
[14:19:29] <dcausse>	 possibly, but I'm not sure how can force MW to go to commons if the same title is locally available on the wiki
[15:00:50] <gehel>	 dcausse: Wednesday meeting is starting: https://meet.google.com/eki-rafx-cxi
[15:02:19] <dcausse>	 oops
[16:12:51] <dcausse>	 dr0ptp4kt: you froze
[16:27:22] <dcausse>	 dinner