IQSS logo

IRC log for #dvn, 2013-05-03

We've moved! Please join #dataverse instead. The new logs are at http://irclog.iq.harvard.edu/dataverse/today

| Channels | #dvn index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

All times shown according to UTC.

Time S Nick Message
08:07 ruebot joined #dvn
15:25 sbmarks joined #dvn
17:47 pdurbin nice. just did a `git pull` on develop and there's a fancy favicon now :)
17:48 pdurbin sbmarks: you're back! :)
17:49 sbmarks I LIVE
17:49 sbmarks just been super busy and out of the loop the past few weeks
17:49 sbmarks but the subnetworks thing put me in mind to get on here
17:50 pdurbin hoboy
17:50 sbmarks that stuff looks good. probably a very good feature for us
17:50 pdurbin good good
17:50 pdurbin it was kinda fun to click the google doc and see people checking it out right in the doc itself
17:51 pdurbin various animals appearing at the top of the page
17:56 pdurbin curious koala or someting. I forget
19:52 sbmarks pdurbin: oh! I had another question
19:52 pdurbin uh oh ;)
19:52 sbmarks do you have any involvement with the OJS/DVN integration stuff?
19:52 sbmarks haha yeah, I'm a killer
19:53 pdurbin nope. here are my notes about OJS: http://irclog.iq.harvard.edu/dvn/2012-12-17
19:55 sbmarks haha  ok. i ask just because we've got a call about getting involved in that aspect
19:56 pdurbin ok
19:56 pdurbin you probably know more than I do
19:56 sbmarks maybe not
19:56 pdurbin sbmarks: now you owe me some faceted search testing ;)
19:56 sbmarks sure! what do you need?
19:57 pdurbin do you like reading? :)
19:57 pdurbin http://dvn.github.io/devguide/features/search/faceted/roadmap
19:57 pdurbin I just wrote that up this morning
19:57 pdurbin and now I'm working on the "converting multi-pass to single-pass" bit
19:57 pdurbin the "within these results" bit
19:57 pdurbin as a test case
19:58 pdurbin here, let me deploy this war to a public server...
19:58 pdurbin (it's buggy but why not) ;)
19:59 sbmarks cool
19:59 * pdurbin deploys to http://dvn-3.hmdc.harvard.edu/dvn/
19:59 pdurbin firstly, does that roadmap make any sense at all?
20:00 pdurbin and secondly, do you want help finding the bugs on dvn-3? ;)
20:03 sbmarks 1) Yes, I think so. Search was fine, but the way it works is incompatible with faceted search. So in order to enable faceted, you need to rework the whole search, but also not make it suck.
20:04 sbmarks 2) of course, happy to help!
20:04 pdurbin heh
20:05 sbmarks but more or less right, right?
20:06 pdurbin yep eyp
20:06 pdurbin yep
20:06 pdurbin does search suck?
20:06 sbmarks no, not that I'm aware of!
20:06 pdurbin oh good
20:06 pdurbin I mean, I don't know, really
20:06 sbmarks i just meant, since you have to rewrite it, it still has to be at least as good as it was before
20:06 pdurbin sure
20:07 pdurbin I don't want to break how it works
20:07 pdurbin the whole collections concept complicates things
20:07 pdurbin (sorry Gary)
20:07 pdurbin who I tweeted at yesterday: https://twitter.com/philipdurbin/status/330010633309409281
20:08 pdurbin hmm, whom I guess
20:08 sbmarks the collections thing is useful though
20:08 pdurbin anyway, I'm trying to think about adjusting the lucene query to be aware of a list of studyIds
20:08 pdurbin oh sure
20:09 pdurbin I was trying to come up with a real world analogy for collections
20:10 pdurbin collections are like having a couple fish tanks on your desk and a couple laptops on your desk with webcams to other people's fish tanks. and you say, "look at my collection of fish!"
20:10 pdurbin and then you say
20:10 pdurbin "let's search for the yellow fish!"
20:11 pdurbin "in my collection!"
20:11 pdurbin and you don't mean just the real fish in the water in front of you
20:11 pdurbin you mean ALL the fish in your collection
20:13 sbmarks_ joined #dvn
20:13 sbmarks_ hrm
20:14 sbmarks firefox, go home you're drunk
20:15 pdurbin I'm trying to make this testing make sense... slovenia... 2003
20:16 * pdurbin tail's server.log
20:17 pdurbin yeah, it's horribly broken
20:17 pdurbin but I'm not sure it's even easy to tell :)
20:18 pdurbin If you click "democracy" at http://dvn-3.hmdc.harvard.edu/dvn/ you should get 1142 studies
20:18 pdurbin Then, if you check "within these results" and type "replication" and search you should get 229 studies
20:18 pdurbin If you click "Slovenia (34)" you actually get 34 studies
20:18 pdurbin And if you click "2003 (3)" you actually get 3 studies
20:19 sbmarks yup
20:20 pdurbin But then if you click "Country/Nation = Slovenia (click to remove)" you get 48 studies, which probably isn't right. or at least the actual query doesn't reflect what you probably think it is
20:22 pdurbin at that point the query is for "replication" and productionDate=2003 ... with nothing about democracy at all
20:23 * pdurbin wishes the lucene devs had some input on my ramblings this morning: http://colabti.org/irclogger/irclogger_log/lucene-dev?date=2013-05-03#l36
20:24 pdurbin sbmarks: how much are you going to cry if you don't see facets everwhere? what if they simply aren't there in certain places?
20:24 pdurbin :)
20:27 pdurbin it's *some* progress that "within these results" shows facets at all... this is the block of code I'm replacing with the new code: https://github.com/IQSS/dvn/blob/205ab6466b0f243111ebc1e81c797e8b45cfc97c/src/DVN-EJB/src/java/edu/harvard/iq/dvn/core/index/Indexer.java#L1546
20:31 pdurbin I went ahead and commited and pushed it
20:31 pdurbin sbmarks: I linked to this conversation: https://github.com/IQSS/dvn/commit/29b8464673938832230ae614a0a392c17ecda2b8#L2R1350 :)
20:32 sbmarks pdurbin: sorry for dropping out, I am on a call
20:32 pdurbin heh. no worries. it helps to talk :)
20:32 pdurbin "talk to the bear" http://cm.bell-labs.com/cm/cs/tpop/debugging.html :)
20:47 pdurbin I'm getting punchy.. I just sent an email with the subject "search: from multi-pass to single-pass for faceted delight" :)
23:24 ruebot joined #dvn
23:24 ruebot joined #dvn

| Channels | #dvn index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

We've moved! Please join #dataverse instead. The new logs are at http://irclog.iq.harvard.edu/dataverse/today