Evergreen ILS Website

IRC log for #evergreen, 2013-10-04

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat

All times shown according to the server's local time.

Time Nick Message
00:10 b_bonner joined #evergreen
00:11 mtcarlson_away joined #evergreen
00:37 b_bonner_ joined #evergreen
00:38 mtcarlson_away joined #evergreen
03:50 b_bonner joined #evergreen
03:51 mtcarlson_away joined #evergreen
04:24 paxed what i really like about Koha is their help system. a link on top right of each page which opens a page on local "wiki" for that page.
06:52 phasefx joined #evergreen
07:27 jboyer-isl joined #evergreen
07:35 rjackson-isl joined #evergreen
07:53 Callender_ joined #evergreen
07:53 collum joined #evergreen
07:55 Callender__ joined #evergreen
07:57 Callender joined #evergreen
08:19 akilsdonk joined #evergreen
08:34 kbeswick joined #evergreen
08:41 mdriscoll left #evergreen
08:42 mdriscoll joined #evergreen
08:48 kmlussier joined #evergreen
08:51 timlaptop joined #evergreen
08:51 mmorgan joined #evergreen
08:52 mrpeters joined #evergreen
08:59 ericar joined #evergreen
09:38 dbwells_ joined #evergreen
09:38 Guest72764 joined #evergreen
10:26 dbwells_ joined #evergreen
10:57 RBecker joined #evergreen
10:57 RBecker joined #evergreen
11:04 BigRig_ joined #evergreen
11:05 BigRig_ joined #evergreen
11:09 misilot joined #evergreen
11:19 rfrasur joined #evergreen
11:22 rfrasur Has anyone deployed a discovery layer over EG...If so, do you have a link?
11:26 jcamins rfrasur: I'll be interested to hear what you learn about that? An EG driver for Biblionarrator is on my to-do list.
11:26 jcamins s/\?/./
11:27 kmlussier I thought Pennsylavani was using VuFind, but I just looked at their web site and only see tpac.
11:28 rfrasur jcamins: whatever I find out, I'll share (if it's not in here...ya know...redundancy)
11:28 rfrasur kmlussier: I've heard urban legends of VuFind....
11:29 jcamins rfrasur: why are you looking at discovery layers, out of curiosity?
11:29 kmlussier rfrasur: http://list.georgialibraries.org/pipermai​l/open-ils-general/2013-April/008394.html
11:30 jcamins Huh. I didn't realize IISG was Evergreen.
11:31 rfrasur jcamins: it's a conversation amongst our consortium members about pulling all our resources together.  I know EG has the technical capability to handle anything...BUT we have some policy things that are in the way...at least right now.
11:31 jcamins rfrasur: I've moved away from my belief in ILSes. Not that my belief in the current crop of discovery layers is any stronger.
11:32 phasefx2 joined #evergreen
11:35 rfrasur jcamins: I dunno what the answer is.  I just know that it's something that we (librarians, at least.  I dunno other people's perspectives beyond anecdotes) need to be actively thinking/doing something about our disparate resources.
11:36 rfrasur kmlussier: ty.  That's exactly what I want to see.  I wish they were closer.
11:36 * rfrasur would hijack them at a conference or something.
11:37 kmlussier rfrasur: I've also heard that KCLS is planning to move to BiblioCommons, but it doesn't look like the move has happened yet.
11:37 jcamins rfrasur: I think that a discovery layer that put aside the card catalog mentality is part of the answer. I'm not sure what the other part is.
11:37 kmlussier http://www.librarytechnology.o​rg/ltg-displaytext.pl?RC=17777
11:39 rfrasur explain the card catalog mentality briefly.
11:41 jcamins rfrasur: ILSes, OPACs, and the current crop of discovery layers take the approach that online catalogs can be used like a card catalog.
11:41 jcamins Enter term, flip through cards, select books, win.
11:41 jcamins But when you work with a card catalog, how many cards can you flip through in, say, thirty seconds?
11:41 jcamins Dozens.
11:42 jcamins Maybe hundreds.
11:42 * rfrasur listens
11:42 jcamins With an online catalog?
11:42 jcamins 20.
11:42 jcamins Maybe.
11:42 jcamins And meanwhile we have more books.
11:42 jcamins So that's a problem.
11:42 rfrasur what might the alternative look like?
11:43 jcamins Better relevancy ranking, not just based on search terms, but also based on what the patron is probably looking for.
11:43 jcamins How many times do you go past... say, three pages on Google?
11:43 rfrasur more of an Amazon search?
11:44 rfrasur Never.  Well, I have, but it was a fool's errand.
11:44 kmlussier jcamins: I'm with you on the better relevancy ranking, but how do we do that? Popularity metrics or are you thinking we need more information about the user in question?
11:44 jcamins Amazon does okay for buying, but we're libraries.
11:44 jcamins Why do people come to libraries?
11:44 phasefx we were resistant to giving Evergreen "browse" early on
11:44 * jcamins thinks they come to libraries for the following reasons:
11:44 jcamins 1) to find something out,
11:45 jcamins 2) to find something to do,
11:45 jcamins 3) because they need a public bathroom.
11:45 jcamins Disregard three, they won't need an OPAC.
11:45 rfrasur lol...maybe
11:45 jcamins The reason Google wins over the OPAC is reason 1.
11:46 moodaepo joined #evergreen
11:46 jcamins The OPAC is a bunch of digitized catalog cards. If you want information about something other than the book, you're SOL. On the other hand, search Google with the same term (say, a name), and the first result tells you a little something about the person.
11:47 jcamins Birth date, death date, maybe a one sentence summary ("Military officer and provocateur, John Smith was known for getting lots of people angry at him, then running away across the ocean.")
11:47 rfrasur ahhh...okay.  So the beef is with the information that we use.  The MARC/card.
11:48 jcamins My vision for a discovery layer is one that combines relevancy ranking that takes into account closely related materials (you searched for "john smith pocahontas," so if the first ten results are all about Jamestown Colony, let's show you more about Jamestown Colony, eh?), and also integrates authority data.
11:48 jcamins I use Wikipedia to look up everything.
11:49 jcamins Then if I want more, I go to the library catalog and copy and paste things out of Wikipedia.
11:49 jcamins It's faster.
11:49 rfrasur yep
11:49 * rfrasur does too
11:49 * jcamins thinks it's absurd that we don't have any of that information in our catalogs.
11:50 jcamins kmlussier: relevance is hard. And I'm definitely not an expert.
11:51 jcamins But I do know that if I can say "this search results in lots of books that have these N things in common," I can probably also say "... so I might as well show the user more books with these N things, because I can't collect the level of information I'd need to answer definitively what the user wants" (i.e. replicate Google's data mining).
11:52 jcamins I'd also like to see the catalog provide more ways to take in information. Computer screens are not optimized for reading long scrolling lists.
11:52 jcamins So I'm trying maps, etc.
11:53 jboyer-isl I don't want to put down your idea because there are lots of good parts, but I think people underestimate the volumes of data that Google has on you, specifically, when they wish that opac searches could be more like Google. The only way to be as effective as the largest advertising company in the world, is to be the largest advertising company in the world.
11:53 rfrasur Honestly...and this shows how much I know...I thought each search into the catalog was a type of data mining
11:54 jcamins jboyer-isl: right, that's why I think we *can't* replicate the relevance.
11:54 * rfrasur takes it all in.
11:54 jcamins The best we can do is my incredibly weak heuristic.
11:55 jcamins That said, I think it's a better heuristic than none at all, which is the traditional approach.
11:55 jboyer-isl Would that mean adding a ton of records for things not already in the catalog (or using authority records for people/places not necessarily referenced in records directly?)
11:56 phasefx brain storming; maybe we can define/expose contexts that a user can choose from, like, "I'm a layperson", "I'm a student studying ___", etc. to help inform the search
11:56 jcamins jboyer-isl: yeah, I don't see how we can make the OPAC useful without adding Overdrive, etc.
11:56 jcamins As for authorities, it doesn't make sense to add authorities that aren't used, IMO.
11:56 smyers_ joined #evergreen
11:57 rfrasur phasefx: build a small reference interview into the experience?
11:57 jcamins Unless your goal is to just provide a ready reference tool, without any bibliographic/holdings component, in which case you may as well just use Wikipedia, IMO.
11:57 phasefx rfrasur: as an option
11:58 phasefx Wikipedia is CC-BY-SA too, so we could slurp that data in and make use of it somehow
11:58 jcamins But I think Wikipedia is probably more reliable than Encyclopedia Britannica, on average, so take my opinion with a grain of salt.
11:58 jcamins phasefx: exactly!
11:58 jcamins Think of it! Use viaf to disambiguate between your NACO headings and Wikipedia, then people get biographies and can check "is that really the author I want?" in the OPAC.
11:59 rfrasur how? (hah!  just kidding...I don't expect an answer...right now)
11:59 jcamins http://libdemo.biblionarrator.c​om/search?q=henry+ward+beecher#
11:59 jcamins That's the University of Michigan's records. I actually loaded it prior to adjusting relevance, so relevance is backwards.
12:00 smyers_ joined #evergreen
12:00 jcamins And I haven't written my Wikipedia harvester yet. But if I'm researching Henry Ward Beecher, having his biography right there would be kind of useful.
12:01 phasefx another idea from the past, I think it would be neat for librarians to be able to create "canned searches" with wiki-like augmentations
12:01 jdouma joined #evergreen
12:01 phasefx the whole wordpress/opac thing was neat
12:01 phasefx scriblio
12:01 jcamins And for a more complicated search, the visualization can lead you to things you wouldn't expect.
12:02 kmlussier dbwells++ #Release Notes are looking much better. :)
12:02 jcamins "Oh, I see, these three books with random titles are actually in the center of my conceptual space."
12:02 phasefx one advantage that a lot of libraries has is that they know their audience/community; google has to cater to everyone
12:02 jcamins ... except replace "conceptual space" with something that real people say.
12:03 jcamins phasefx: that's true, especially in special libraries.
12:04 * phasefx used to think that having essentially finite records was an advantage; not so sure these days
12:05 phasefx jcamins: I've heard the notion of making the OPAC into a research tool/workspace before, where folks could gather/create/edit information
12:05 phasefx share
12:06 jcamins phasefx: yes. This. That is exactly what I feel it needs to be.
12:06 phasefx a step in that direction would be spatial bookbags, where you can drag your resources wherever like on a post-it board
12:07 jcamins Ooh, I like that idea.
12:07 phasefx all good OPACs have lists/bookbags, and can "share", but none can really "consume"
12:07 * jcamins borrows it.
12:08 * phasefx wants to take jcamins' reading list and incorporate it into his workspace
12:08 jcamins My list concept only involved one dimensional organization, but two dimensions is much better.
12:08 phasefx humans have excellent spatial memory
12:08 jcamins Three dimensions and a thingy-flow would be even cooler, of course.
12:09 phasefx I can remember where on a page better than a page number
12:09 jcamins phasefx: you can probably even remember where in a book better than a page number.
12:09 * phasefx nods
12:10 phasefx jcamins: let me know if you do the spatial bookbag thing; would love to see it
12:10 jcamins phasefx: I think I definitely will. It's brilliant.
12:10 jcamins And well beyond my current capabilities, d3js-wise! What's not to love?
12:11 phasefx :D
12:11 * phasefx can't imagine how to get it into the javascript-lite TPAC :)
12:11 jcamins phasefx: use a discovery layer, probably.
12:12 jcamins Poor rfrasur. I hope I have not scared her away from discovery layers.
12:12 phasefx jcamins: make it like android, where if you drag a resource onto another resource, they get grouped
12:13 jcamins phasefx: oh, that reminds me.
12:13 rfrasur I'm sorry...I was actually listening in the conference and thinking at the same time.  It takes much to scare me.
12:13 jcamins The other thing about discovery layers.
12:13 jcamins And catalogs generally.
12:13 jcamins Desktops are a lousy interface for them.
12:13 rfrasur You made me think about how we're dealing with the information...and what the priority...at my end...needs to be.
12:14 jcamins At first they were all we had, but now we have tablets and kiosks... and we're not using those devices' capabilities.
12:14 rfrasur The reference interview is the BEST interface.
12:14 phasefx man, a mobile app with gps and awareness of a library's physical layout would be interesting
12:14 rfrasur If you could plug the librarian person into the catalog...that'd be the best discovery layer (which I think we do).
12:15 rfrasur librarians = original discovery layer
12:15 phasefx imagine the geocaching type games you could get kids to play
12:15 phasefx with books
12:16 jcamins rfrasur: a lot of libraries don't have the luxury of enough librarians, and a good catalog is the best they can do.
12:16 phasefx rfrasur: the way it seems to work most of the time these days is "ask your friends, and they ask the internet"
12:16 rfrasur jcamins: phasefx: you're both right
12:16 jcamins Well, a good catalog is the best they _could_ do.
12:17 jcamins Often they don't do, but in a theoretical sense, there could be a good catalog.
12:23 jcamins jboyer-isl: I suppose if you knew that your constituency was particularly interested in a particular thing you might load authority records relating to it, even if you didn't have any books on it.
12:23 jcamins For example, a library with a large Russian immigrant population might choose to load authority records for major Russian authors, even if the acquisitions budget is too small to acquire anything by some of the less well-known ones.
12:24 jboyer-isl jcamins: I think the only big point to unused auth records could be following the see or see also refs for records that match, even if there aren't direct uses of those records. I was spitballing a bit at the time.
12:24 jcamins jboyer-isl: oh, I may have missed making one point about my vision clear... only valid links should show up.
12:25 jcamins At present I don't have to worry about that in Biblionarrator, because I am loading only those records that are linked.
12:25 jboyer-isl i see.
12:25 jcamins That's definitely a crucial point, that there not be any blind links.
12:25 * bshum catches up on the scrollback
12:25 jcamins That's why I ended up with a graph database.
12:26 bshum kmlussier: Houston Public Library in SITKA was an example of a bibliocommons using Evergreen lib:  http://houstonbc.bibliocommons.com/
12:26 bshum We looked at them once upon a time
12:26 * rfrasur checks out Houston.
12:26 kmlussier dbwells: I'm almost ready to push a bunch of new release note entries for features that were missing notes. I just want to make sure I should still be pushing them to the subdirectories of the RELEASE_NOTES_NEXT folder? Will you be running the script again at release time?
12:26 bshum Their library website seems to have links to both the SITKA catalogue as well as Bibliocommons.
12:26 jcamins I really liked BiblioCommons the first time I looked at it, but recently it seems to have gotten less functional.
12:26 bshum So maybe they're still deciding
12:27 bshum Or TPAC is swaying them back :)
12:27 kmlussier Nice! I really like BiblioCommons. But I haven't looked at it recently.
12:28 bshum As for VuFind, the SPARK libs in PA probably switched back to TPAC I think when ESI took over as their hosts.  Prior to that they definitely were using VuFind.
12:28 bshum I can't find a nice link to the Tasmania libraries
12:28 bshum They've been quiet, but in the past had indicated using Evergreen + VuFind
12:28 bshum PALS used to use VuFind too, though I'm unclear on whether they've actively done anything with it, or if they're also switched to TPAC.
12:30 jcamins bshum: who is PALS? I think there are several library systems with that acronym.
12:30 bshum jcamins: Sorry, erm... MNPALS (Minnesota).  It's the only PALS I know of in the Evergreen sphere :)
12:30 bshum But looking around them, I don't see much about VuFind and all the libraries on Evergreen link to TPACs
12:31 bshum So they must have went a different way
12:31 jcamins Thanks. I know the catalog librarian at a PALS in NJ, and she never once mentioned migrating to Evergreen, so I figured it wasn't them.
12:31 dbwells kmlussier: yes, I plan to redo them again.  The way I did it for the beta was pretty ad hoc, so hopefully the second time through I can be a little more systematic about it.
12:32 kmlussier dbwells: Great! That's what I was hoping you would say. :)
12:35 yboston joined #evergreen
12:38 bshum dbwells: So we're testing that fix for receiving
12:38 bshum dbwells: And it looks like it received properly and we created the item
12:38 bshum But apparently the range in the summary holdings doesn't show the updated issue
12:40 dbwells bshum: what holdings display are you using?
12:40 bshum Mary was looking at http://acorn.biblio.org/eg/opac/record/2236299?e​xpand=issues;sid=1644;stype=basic;sepath=2013;se​path=09;seoffset=0;seoffset=0;seoffset=0#issues for example.
12:40 bshum Where the last issue that was just added was the Sept 15 one
12:40 bshum But the range isn't updated
12:41 ericar joined #evergreen
12:43 dbwells bshum: looking at it, but by the way, if you put those caption parts in parens (), they won't display so oddly, e.g. "$i (year)" instead of "$i year".
12:43 bshum dbwells: I'll mention that to her.
12:43 kmlussier bshum: Any chance it is related to bug 1000397?
12:43 pinesol_green Launchpad bug 1000397 in Evergreen "Receiving unitless issues in the batch receive interface does not update tpac holdings summary" (affected: 2, heat: 10) [Undecided,Triaged] https://launchpad.net/bugs/1000397
12:43 bshum I'll ask
12:44 dbwells I see a barcode, so I don't think it is that, but it could be somehow related.
12:46 akilsdonk joined #evergreen
12:46 dbwells I should probably just say "Oh, 'use fully compressed holdings'?  That's senator's problem." ;)
12:50 mllewellyn joined #evergreen
12:58 _bott_ I hate weird problems.
12:58 dbwells bshum: I can see what you are saying on the record you posted, but I can't duplicate it on my test box.  I'd need to see the exact holding codes for your Sep.08 and Sep.15 issues, I guess, then go from there.
12:59 * dbwells goes to lunch, back in a bit
12:59 _bott_ Has anyone found the status_changed_time trigger not firing consistently?   I have random cases where the status is changed at checkin, but the status_changed_time is not updated, it's left at the checkout time.  This in turn screws up reshelving delay.
13:08 bshum dbwells++ # we figured it out, it was a local holding_code problem
13:08 bshum The holding_code had been modified with a bad copy/paste and the date in the code was 08 not 15
13:08 bshum User error
13:08 bshum We'll test again with another sample, but I think this means your patch works to resolve the receiving problem.
13:10 ericar joined #evergreen
13:11 pinesol_green [evergreen|Kathy Lussier] Move various release notes into correct directory - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=30a69c4>
13:11 pinesol_green [evergreen|Kathy Lussier] Adding new OPAC features to Release Notes. - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=e66a212>
13:12 pinesol_green [evergreen|Kathy Lussier] Release note additions for new acquisitions features - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=51f6c9c>
13:12 pinesol_green [evergreen|Kathy Lussier] More release note entries - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=ebb94c3>
13:19 rfrasur joined #evergreen
13:20 mllewellyn pleading guilty to the operator causing the error.
13:31 RoganH joined #evergreen
13:42 dbwells bshum: yay :)
13:42 bshum dbwells: Yep, she tested on a different issue and it received perfectly fine.
13:42 bshum I'll sign off on your fix and get that pushed to master.
13:43 bshum dbwells++
13:45 rfrasur jcamins: this speaker is talking exactly what you were talking about earlier...
13:46 pinesol_green [evergreen|Dan Wells] Relax MFHD subfield 'a' requirement for caption/patterns - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=a420b0e>
13:46 jcamins rfrasur: oh? What speaker?
13:46 rfrasur Ted Fons from OCLC'
13:47 jcamins What conference are you at?
13:49 rfrasur Midwest Cooperative for Library Services Annual Meeting (it's about open stuff this year...open access...data...source)
13:49 * rfrasur thinks we should pool our money and buy some really expensive stuff from Google
13:51 rfrasur He's talking about knowledge graphs which sounds a lot like the maps you were talking about.
13:52 jcamins Does Google have really expensive stuff to sell us?
13:53 rfrasur I dunno.
13:53 rfrasur But they must, right?
13:53 jcamins Expensive stuff, yeah, but I don't think they sell it.
13:53 rfrasur https://docs.google.com/document/d/1ZU9sQrAkI88_​NSGSDUyMiYRnnNDQcUO2SJOjJANtj0o/edit?usp=sharing My notes
13:55 jcamins Interesting.
13:55 * jcamins watches you type. :P
13:55 dbwells I am seeing an issue with title browsing on 2.5-beta1, and am hoping for someone more familiar with the feature to troubleshoot or verify what is happening.  On a clean install, I imported three records, but see 6 titles in browse.  For four of the six, the only difference is trailing punctuation.
13:56 dbwells For the other pair, one has a subtitle, one doesn't.
13:56 rfrasur jcamins: you can see how often I delete
14:00 jcamins rfrasur: was his list of how OCLC was doing this as hand wavy as it looks?
14:01 kayals joined #evergreen
14:01 rfrasur It wasn't to begin with...it was to end with.
14:03 jcamins "Network of links and entity identifiers" really bugs me.
14:04 rfrasur That part was pretty ambiguous...about the "network of links and entity identifiers"
14:04 jcamins Yeah, that's why it bugs me.
14:04 bshum dbwells: The punctuation difference is something that still irks me about browse search.
14:04 jcamins That's what jumped out as hand-wavy.
14:04 bshum Same with case sensitivity, all caps, vs. initcap vs. all lowercase
14:04 rfrasur the way he framed entity identifiers is basically authorities...although maybe a little more robust?
14:05 jcamins Robust like... the way everything ground to a halt when LC shut down their website the other day? Yeah...
14:05 jcamins I can translate for you, if you like: "web pages with URLs."
14:05 bshum dbwells: I think it was noted during the development but I'm not sure what the result was.  kmlussier / ESI would know best, but I think it needs more development to solve.
14:05 rfrasur well....robust as in LOTS of authorities
14:05 jcamins Oh, I was thinking robust as in "can't fail."
14:05 dbwells bshum: Well, what I am seeing seems to stem from the fact that browsing includes both 'Title Proper' fields, and 'Title Proper (Browse)' fields.
14:06 rfrasur everything can fail :D
14:06 rfrasur no.  everything WILL fail.
14:06 jcamins And usually right when you have to prepare reports for management.
14:06 rfrasur inevitably
14:07 kmlussier dbwells: That's right. We noticed the trailing punctuation issue and would like to change it. But it does require more development that wasn't covered under our project.
14:07 bshum dbwells: Huh... I don't have a title proper (browse) in my production DB....
14:07 kmlussier IIRC, when you are using authority records, it isn't as noticeable with author or subject browsing. But authorities don't help much with a title browse.
14:07 * bshum checks clean master
14:08 bshum Gah, wth
14:08 kmlussier I've noticed the same issue in autosuggest too.
14:08 dbwells kmlussier: I am actually less concerned about the trailing punctuation, and more concerned about the duplication.  Is there a reason we have both "Title Proper" and "Title Proper (Browse)" set as browseable fields, or is that just a mistake?
14:10 kmlussier dbwells: Ah, that's something I don't remember seeing.
14:10 dbwells Maybe it is a non-filing issue or something?
14:12 bshum Huh
14:12 senator dbwells: they definitely don't both need to be set as browseable fields
14:12 senator that's a mistake in the seed data i suppose
14:12 bshum The upgrade SQL didn't add title proper (browse)
14:12 bshum Which is why I don't see it in my production database
14:12 senator i'll look back and see if both were even meant to survive, or just title proper alone
14:13 bshum So if it's meant to live, that's missing from the upgrade SQL
14:15 rfrasur This next one is "The inevitability of open access" https://docs.google.com/document/d/1Ho3FMLfP_q2x​VjtdBDPKZXCuiT11gSTW8LovfM8kk9U/edit?usp=sharing (then I'm going home).
14:22 jeff "You hear that Mr. Anderson? That is the sound of inevitibility."
14:22 gmcharlt :)
14:24 rfrasur jeff++
14:26 jcamins Kind of a bad time to be talking about the inevitability of open access.
14:27 jcamins It was what, yesterday, that Science released their study of how peer review in open access journals is shameful?
14:28 jcamins Non-OA journals are probably equally shameful, but at least the argument for non-OA isn't "openness improves transparency."
14:31 gmcharlt jcamins: given that the author of that paper neglected to submit it to any proprietary journals, who can tell?
14:31 gmcharlt at the very least, Science comes off as rather self-serving
14:32 jcamins gmcharlt: yeah, as I said, non-OA journals are probably just as bad, but if >50% of the journals submitted to accepted it for publication, the issue should not be OA vs non-OA, it should be "how do we make scholarship work at all?"
14:33 rfrasur jcamins: that's been a running joke today...we're talking about this and THAT was on the news this morning.
14:33 gmcharlt jcamins: possibly a revamping of the peer review process
14:33 rfrasur (not a funny joke...an ironic joke)
14:33 gmcharlt around the notion that it's a continuous action of the community, not a once-and-done-to-get-pubished thing
14:34 rfrasur gmcharlt++
14:34 * rfrasur hasn't read the Science article, but will.
14:34 jcamins gmcharlt: yeah, and probably some serious tenure reform, which is why it won't get fixed any time soon.
14:34 gmcharlt some links for light reading
14:34 gmcharlt http://www.michaeleisen.org/blog/?p=1439
14:35 gmcharlt https://plus.google.com/u/0/109377​556796183035206/posts/CRHeCAtQqGq
14:35 gmcharlt http://svpow.com/2013/10/03/john-bohann​ons-peer-review-sting-against-science/
14:35 gmcharlt not that I have been following this or anything ;)
14:36 jcamins gmcharlt: I only saw it because the headline crossed my RSS feeds. I'm actually fairly disinterested, having concluded long ago that scholarship is pretty hopeless.
14:39 krvmga joined #evergreen
14:40 rfrasur He's talking about "post publication review."
14:41 * rfrasur loves the phrase "systematic crawling of the literature."
14:41 rfrasur yes please
14:42 mrpeters question -- is inactivity a potential cause for open-ils.auth to die off around the same time every day?
14:43 tsbere mrpeters: I assume that too much activity would be more problematic. Otherwise it would die off every night, wouldn't it?
14:43 jeff open-ils.auth shouldn't "die" due to inactivity, but what are you seeing?
14:43 mrpeters looking at a system that isn't "live" yet, but seeing some staff use...auth just dies around noon everyday.  wondering if people going to lunch or something is causing the service to stop listening
14:43 mrpeters tsbere: that's true
14:43 * mrpeters thinks the values for drones, etc. may be stock
14:44 mrpeters might need to be increased
14:44 jeff mrpeters: are there external symptoms (people can't use the system), or are you just looking at a log entry saying that a process exited/etc?
14:44 mrpeters jeff:  staff client login fails with the same type of alert that you'd get if you miskeyed your pw
14:46 tsbere mrpeters: Too many bad password attempts in a short period of time causing lockout?
14:46 mrpeters would that cause auth to die?
14:46 jeff mrpeters: And what do the logs say around that time?
14:46 tsbere Shouldn't kill auth, but that symptom would happen
14:47 mrpeters someone hammers it with the wrong password over and over
14:47 mrpeters well, the thing is, ejabberdctl connected-users says auth is gone
14:47 tsbere Ok, that would be an indicator of a problem, yea
14:47 tsbere Is the listener still running at that point?
14:47 mrpeters let me sanitize these logs
14:49 pastebot "mrpeters" at 64.57.241.14 pasted "open-ils.auth dying off" (29 lines) at http://paste.evergreen-ils.org/26
14:49 jeff mrpeters: In your troubleshooting, I'd start with... how long has this pattern gone on? two days, a week? Any changes during that time? Any other interesting things / errors near that time?
14:49 mrpeters jeff:  its been happening since implementation
14:49 mrpeters we got called in about a week ago on it
14:49 mrpeters i've got a lot of log digging to go, i mostly wondered if inactivity would cause that listener or drone to die off
14:50 jeff Nope, shouldn't. Good luck in your investigation. :-)
14:50 mrpeters thanks jeff
14:50 mrpeters will let everyone know what i figure out, for the logs
14:51 tsbere mrpeters: Is the *router* auth the only one gone? Or is the listener gone? If the latter, is the listener process still running, just not connected? The latter can happen if something kicks it off of ejabberd, such as a protocol level error...
14:55 yboston left #evergreen
14:55 yboston joined #evergreen
14:59 mrpeters tsbere: good question, let me confirm
14:59 mrpeters hmm, what do the listener procs look like in ejabberd-ctl
15:00 dbwells So, my failed attempt at searching Google for 'mods32' did manage to find some random SpiceJet Flight which apparently involves time travel (departs tomorrow, arrives TODAY).  https://www.google.com/search?q=mod32
15:00 tsbere mrpeters: opensrf@private.localhost/open-ils.auth_​listener_HOSTNAME_TIMESTAMP_PID I think
15:01 dbwells For anyone who wants to repeat today, for some reason.
15:01 mrpeters weird....the app servers show both open-ils.auth Listener and Drone running
15:03 mrpeters don
15:03 tsbere mrpeters: The router will drop the service if it can't talk to the last listener. Anything talking to ejabberd can be kicked off by a XMPP-level error...
15:03 tsbere or if something else logs in with the same resource ID
15:04 mrpeters oops....don't see any auth_listener processes connected to the brick head's connected-users
15:04 mrpeters so, it would seem the listener dies first, then the router kills the process off
15:05 tsbere mrpeters: I would check the ejabberd logs, see if you can see what killed the listener's connection
15:06 mrpeters will do.  thanks
15:06 mrpeters best practice question -- the brick heads here only have opensrf.settings running.  Would it be better practice for auth to run there than the app servers?
15:08 tsbere the only best practice I know on that front is "you should only run settings on one box"
15:09 tsbere Though "running cstore, rstore, pcrud, and storage on a drone that has a direct link to the DB server if you have one" isn't bad either...
15:10 jeffdavis mrpeters: FWIW we run settings and auth on our brick heasds
15:11 tsbere mrpeters: If auth is dieing on *multiple* drones within a short period of time I would assume a failed attempt being repeated until no drones are left....which might indicate an attack, actually...
15:12 mrpeters i'd highly doubt an attack....possible, but highly doubtful.  logs don't seem to indicate it.
15:13 tsbere If you can take down a service with only a few calls, killing one listener with each as the router loops through them, it might not look like an attack at all
15:15 mrpeters jabber logs don't seem to indicate anything about the service dying
15:16 mrpeters lots of accepted legacy authentication and close sessions, but
15:16 tsbere mrpeters: The two most common issues I have seen were "too much data" and "malformed XML", at least on the XMPP side
15:16 mrpeters you'd see something like that in the ejabberd.log you say?
15:17 mrpeters no instances of those
15:17 * paxed ponders how hard it would be to implement "wiki"-like help, like in Koha...
15:17 * mrpeters debates just upping the drones for auth, but also debates moving auth to the brick head because thats how I've always had succcess
15:18 mrpeters its just weird to me that the listener and drone still show up in a "ps aux | grep auth"
15:18 mrpeters but dont show as connected to jabber
15:18 tsbere mrpeters: I doubt moving to the brickhead would help. *something* is kicking it off of jabber, you need to find out what.
15:19 mrpeters maybe jabber logs aren't set at high enough verbosity
15:50 pinesol_green [evergreen|Kathy Lussier] Move some release notes to core docs - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=a1d2326>
16:33 mrpeters left #evergreen
16:38 rfrasur joined #evergreen
16:39 dbwells senator: I see your browse fix branch.  Is it your opinion that no upgrade script is necessary?  I'm inclined to agree, since it only affects those who installed fresh from master/beta1 after the branch was merged, which should be very few people.
16:41 senator dbwells: yes
16:42 senator sorry i got into other things and didn't push that earlier
16:42 senator bug 1235452
16:42 pinesol_green Launchpad bug 1235452 in Evergreen "Remove unnecessary Title Proper (Browse) index from config.metabib_field" (affected: 1, heat: 6) [Undecided,New] https://launchpad.net/bugs/1235452
16:43 dbwells senator: no problem, will push right away.  While your thinking about it (or maybe you aren't), do you know if there is a reason we don't run chopPunctuation on the titleNonfiling field?
16:44 smyers_ joined #evergreen
16:45 dbwells senator: I think I'll open a separate bug about that, but it would clean up the trailing punctuation from the title browse, I think.
16:50 pinesol_green [evergreen|Lebbeous Fogle-Weekley] Remove unnecessary Title Proper (Browse) index from config.metabib_field - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=3a2112a>
16:51 senator dbwells: makes sense. can't think of a particular reason why chopPunctuation isn't applied to titleNonfiling
17:05 mdriscoll left #evergreen
17:14 mmorgan left #evergreen
17:22 bshum paxed: It isn't the same thing, but I've long wanted to work on getting some stock html help files in place.
17:22 bshum Like described in http://evergreen-ils.org/dokuwiki/​doku.php?id=scratchpad:custom_help
17:23 bshum So that the evergreen help action actually does something
17:23 bshum Being a wiki would make it easier to edit though for end-users
17:24 paxed bshum: i was thinking of having YAOUS for opac.wikihelp -> http://wiki.evergreen-ils.org/wiki/$V​ERSION/$LOCALE/$TEMPLATENAME#section or something.
17:25 paxed and clicking on "Help" in OPAC would go to that url (replacing the placeholders).
17:26 paxed i really like how in Koha it's easy to edit the help. it's like a wiki, but on local server.
17:27 paxed same should apply to the staff client helps.

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat