Time Nick Message 23:22 CrispyBran Have a good one. 23:22 CrispyBran Thanks for the info. 23:14 wizzyrea so that's where to start. 23:14 wizzyrea either way you'll know when you go to bz apply the patch on current master. 23:13 wizzyrea fixing a "insufficient blobs" is way different from say, a merge conflict 23:13 wizzyrea the approach really depends on how it's not applying 23:13 wizzyrea if not, find out why, and fix that. 23:13 wizzyrea reattach 23:12 wizzyrea if it works, yay 23:12 wizzyrea you could also check out a clean master, and cherry pick your patch over 23:12 wizzyrea that might do it 23:12 CrispyBran do I just to a git pull on the master, test my patch again and then obsolete and attach again? 23:10 wizzyrea (but it depends on the error message) 23:09 wizzyrea rebase it, probably 23:09 CrispyBran Joubu: when you create a patch and someone says it doesn't apply, what do I need to do on my end? 21:25 talljoy hiya rangi! 21:25 rangi hi talljoy 21:23 wahanui o/ '`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`' 21:23 wizzyrea more confetti! 21:23 wahanui confetti is, like, http://25.media.tumblr.com/tumblr_lpm3j6aNaN1qh8hleo1_400.gif 21:23 wizzyrea confetti! 21:16 alexbuckley Nope I did like your changes, better to see fewer screens by having the additional info about how to create things on the same page as the forms for example 21:16 Joubu have to run, see you tomorrow :) 21:15 Joubu just hope you did not signoff because you gave up! 21:15 alexbuckley No worries :) 21:14 Joubu thx for the quick signoff alexbuckley :) 21:14 Joubu ha dbic schema you meant, yes you will have 21:14 alexbuckley Yes thank you for the work on the onboarding tool Joubu 21:13 cait so i don't need to regenerate? 21:13 cait oh? 21:13 cait just something i haven't done so far :) 21:13 Joubu a simple c/p of the DB rev from master should be enough 21:13 Joubu cait: do not worry with the schema changes, now we have check to avoid failures on upgrade 21:12 cait have to figure out the schema changes too i think 21:11 cait i'd like to get the translators on it too - but will try to push it eraly 21:11 Joubu yep, apparently hea will not be backported this month, but should be next month :) 21:09 rangi to out of the SCO :-) 21:09 rangi ill also fix the english 21:08 rangi i think that is less clunky than he or she eh? 21:08 rangi If the user logged in is the SCO user and they try to go out the SCO module, log the user out removing the CGISESSID cookie 21:08 cait Joubu++ 21:07 rangi and its good to see hea2 live too 21:07 rangi Joubu: thanks for the work on the onboarding tool 21:01 rangi cool, ill update it now(ish) 21:01 Joubu apparently the problem does not appear on the interface 21:01 Joubu the patch is only about comments, so no translation needed 21:00 rangi if we get the base neutral, individual communities can decide how to deal with the translations 20:59 rangi and in english, they can be singular or plural, so easy to always use that 20:59 rangi yeah, mÄori has no gendered pronouns so it is easy 20:57 cait i didn't know until not so long ago 20:56 cait i think some of the he/she/they problem is non native speakers not being aware of the neutral forms and how to use them 20:56 cait so people are more aware? 20:56 cait if we want to change it permanently, maybe we should also have a coding guideline 20:53 rangi in fact, im gonna put my time where my mouth is and do a follow up doing that 20:52 rangi it is actually important 20:52 rangi http://timesofindia.indiatimes.com/city/thiruvananthapuram/Thiruvananthapuram-library-opens-a-new-page/articleshow/55395836.cms 20:48 rangi there is no reason to ever need to use gendered language in an example 20:48 rangi problem solved 20:47 rangi just use they 20:19 CrispyBran If we changed the language to female, I seriously doubt we'd lose any of the male programmers. Anyway, that's all the energy I can contribute to this topic. Moving on to problematic code. 20:17 CrispyBran I am not sure how reference to a particular gender as an EXAMPLE proves to be problematic. If a programmer has issue with this, there are deeper issues that should be addressed, rather than taking offense at an example. 20:09 CrispyBran :) 20:08 cait hm this vegetable needs to do laundry, brb 20:07 cait ? 20:07 cait when a notice with the code ACCEPTED is set up, a message will be sent from the kitten to the patron. 20:06 cait lol 20:06 cait ? 20:06 cait hm stop her 20:06 * cait has to admit a 'he' in a comment doesn't stop me 20:06 CrispyBran 'When a carrot manages the suggestion, it can set the status to "REJECTED" or "ACCEPTED".' 20:06 cait CrispyBran: can you demonstrate this in a sentence? :) 20:05 * cait waves 20:05 * CrispyBran thinks all references to librarians should be in animal or vegetable form. 20:02 * Joubu trolls and goes out 20:02 wahanui and ever, amen. 20:02 Joubu but "he or she" just make things heavy to read IMO 20:02 Joubu whatever 20:02 Joubu can be replaced by she or whateber 20:02 CrispyBran Joubu: really? Someone is going to waste programmer time with this? 20:00 Joubu to me it does not make sense to make these changes as it is only in code comments 20:00 huginn` 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=18432 trivial, P5 - low, ---, ephetteplace, Signed Off , Most code comments assume male gender 20:00 Joubu I'd like a English native speaker to take a look at bug 18432 19:52 huginn` 04Bug 18450: major, P5 - low, ---, koha-bugs, NEW , Renew in header bypasses hold block and renewal limits 19:52 CrispyBran https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=18450 17:56 * magnuse will try to squeeze in a signoff 17:45 oleonard I hope to do so 17:45 magnuse patches are welcome ;-) 17:44 magnuse gotcha! 17:44 oleonard (easily) 17:44 oleonard So you can't add the OPAC as a search engine in Firefox 17:44 magnuse ah 17:44 oleonard It can provide OpenSearch results, if I recall correctly, but it doesn't enable auto-discovery of search to browsers 17:44 magnuse is it something else that is used to show results from kete in koha and vise versa? 17:43 magnuse oleonard: i thought koha had opensearch? 17:00 oleonard I'm surprised I haven't wished Koha had OpenSearch for the OPAC enough yet to submit a patch. 16:59 oleonard I'm surprised Koha doesn't have OpenSearch for the OPAC 16:39 reiveune bye 15:11 barton good morning #koha! 14:50 Mauricio_BR Thank you Cait oleonard kidclamp ;) 14:40 cait from the urls 14:40 cait you could use a tool like piwik to get the searches 14:40 cait the search history 14:40 cait i think it#s in a cookie before you log in? 14:34 kidclamp could be a development though, seems possible in existing code to support saving as anonymous patron - but would want tied to syspref 14:33 oleonard Are searches logged by zebra? 14:33 Mauricio_BR ok, thanks guys for the information. 14:32 Mauricio_BR is only for suggestions and/or reading history items. 14:32 Mauricio_BR yes 14:30 eythian ah, suggestions not history. I misremembered. 14:29 Mauricio_BR http://translate.koha-community.org/manual/3.20/en/html/administration.html#AnonSuggestions 14:29 kidclamp we only save in the table if we have a logged in user 14:28 kidclamp in the code it seems non-logged in history is stored for session, but not in db 14:26 oleonard The AnonymousPatron pref says "(for anonymous suggestions and reading history)" 14:26 Mauricio_BR there is the user with id 0 but it seems to be the admin 14:25 Mauricio_BR so it is bound to a user 14:25 eythian isn't there a special user it becomes if it's anonymous that's configured by a syspref? 14:25 Mauricio_BR but in this table every record has a user field 14:24 Mauricio_BR yes 14:24 oleonard Mauricio_BR: Have you looked at the 'search_history' table? I think kidclamp is right: I only see entries for logged-in users. 14:23 Mauricio_BR XD 14:23 oleonard Better than my... Every other language which exists. 14:22 Mauricio_BR my english is not good as you can see... haha 14:22 Mauricio_BR oh, sorry 14:22 oleonard Thought you were asking about a text string 14:22 oleonard Oh I misunderstood what you were asking 14:21 kidclamp I don't think search history is sotred unless the user is logged in 14:21 Mauricio_BR because i am working with datamining... 14:21 oleonard Why are you looking for it? 14:21 Mauricio_BR i have the 16.05 ver. of Koha 14:20 Mauricio_BR i am looking for it in the tables of the database 14:19 oleonard Mauricio_BR: Where do you see it? 14:17 Mauricio_BR Hello friends. Please help me with something. I am trying to locate in the database a table wich store the words searched in OPAC as anonymous user (no login in OPAC). Do you know where can I find it? 14:07 marcelr Joubu: hi; any chance to have another look at the qa changes on the upload reports 17669/18300 ? 13:29 oleonard Not a big project, just a lot of small ones 13:28 cait revamping another website? :) 13:27 oleonard More than me :P 13:26 * cait didn't do much this release 13:25 oleonard cait++ 13:22 cait mveron++ :) 13:19 mveron :-) 13:18 oleonard mveron++ 12:22 oleonard I will take a look 12:21 huginn` 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=7550 normal, P5 - low, ---, veron, Needs Signoff , Self checkout: limit display of patron image to logged-in patron 12:21 mveron oleonard: Bug 7550 - what do you think about? 12:21 * cait waves :) 12:21 * mveron waves 12:03 * oleonard waves 12:01 marcelr hi #koha 11:57 toins hi all 11:49 oha we will be too! 11:47 ashimema atheia will be super happy :) 11:47 ashimema :) 11:45 magnuse i have started to look at adapting my existing NNCIPP code, so I might be able to do a signoff pretty soon 11:45 magnuse ashimema: ah, if you use it in production that should be an "argument" for getting it in 11:23 * cait waves 10:55 ashimema we use it in production allot 10:55 ashimema I'd love to see it in :) 10:34 huginn` magnuse: The operation succeeded. 10:34 magnuse @later tell atheia do you think it would make sense to try and get ILL as it stands now into 17.05? 10:34 magnuse Feature Slush for 17.05 is May 5 2017 10:28 magnuse ashimema: do you think it would make sense to try and get ILL as it stands now into 17.05? i should probably ask atheia... 09:18 wahanui niihau, eythian 09:18 eythian hi 08:28 wahanui kamelÃ¥sÃ¥ 08:28 cait hi magnuse 08:27 magnuse kia ora cait 08:11 magnus_breakfast \o/ 07:57 fridolin happy Easter Egg :) 07:56 fridolin hie there 07:30 dcook Good luck Eurofolk 07:30 dcook Anyway, I better head out 07:30 dcook hehe 07:27 oha speaking of oha vs koha, i usually add a "oha" comment when i change something so i can easily find it again. but with so many (k)oha strings this hasn't been working so well lately :) 07:20 wahanui privet, gaetan_B 07:20 gaetan_B hello 07:18 oha dcook: oh! not sure. i've used oha for more than 20 years now. 07:17 dcook I don't think so... I'm sure I've seen your full name somewhere :p 07:16 oha dcook: eheh, is it because (k)oha? :) 07:14 dcook oha... that name is familiar 07:01 oha o/ 06:51 wahanui salut, reiveune 06:51 reiveune hello 06:46 wahanui hello, alex_a 06:46 alex_a bonjour 06:46 dcook Yeah maybe 06:41 magnuse dcook: yeah, bit of a different use case, i guess 06:32 dcook And they don't import RDF at all :/ 06:32 dcook At least to outside entities 06:32 dcook Although maybe that's partially because they don't do any real linking 06:32 dcook They dont' used named graphs though which was interesting 06:31 dcook magnuse: Unfortunately, I don't think he'll be able to help too much, but it'll be good to get some more details 06:31 dcook 4-5 always seems the busiest time of day! 06:31 dcook magnuse: No worry. I have about a million different things happening at once anyway :) 06:26 josef_moravec morning #koha 06:16 mveron Good morning / daytime #koha 06:13 magnuse brinxmat probably has good advice 06:13 magnuse dcook: sorry, was busy in another window 05:25 dcook At the moment, I'm thinking we download the records via OAI-PMH then... run them through a filter which creates triples that Koha can understand. Although we could also use SPARQL CONSTRUCT although that would in theory involve more hardcoding... 05:20 dcook I think it relates to how I import/store the data though 05:20 dcook Sorry if I'm stepping on your toes at all 05:19 dcook Hey magnuse, chatting to @brinxmat on Twitter about LD atm 05:19 * dcook waves 05:14 * magnuse waves 02:43 dcook If you broke them up, your schema:name would go into that named graph... 02:43 dcook Would you stick them all in one named graph or break them up.. 02:43 dcook You'd. 02:42 dcook So if you did download from http://www.worldcat.org/title/good-omens/oclc/973430700 02:35 dcook And this year the National Library of Finland is opening up its national bibliography as linked data it seems.. 02:34 dcook https://news.minitex.umn.edu/news/digital-initiatives-metadata-education/highlights-alcts-webinar-linked-data-cataloging 02:30 dcook Actually, that all looks proposed... 02:27 dcook Not that that is even anywhere near my problem atm.. 02:27 dcook I do wonder a bit about using some other tools for handling RDF records... and hooking them more loosely into Koha.. 02:26 dcook Interesting... and they're using BIBFRAME... of some kind 02:25 dcook As I don't see a lot of data out there about that, yet that seems to be what UCDavis and Oslo both do.. 02:24 dcook I'm intrigued by a RDF->MARC conversion as well. 02:23 dcook That seems... 02:23 dcook But maybe just into the cataloguer 02:22 dcook Interesting... it seems that they do download the OCLC graph.. 02:21 dcook I'm too young to be squinting.. 02:21 dcook Gotta love super low res images.. 02:20 dcook Hmmm https://bibflow.library.ucdavis.edu/copy-cataloging/ 02:19 rangi ah yeah 02:19 dcook Tweets & replies rather than just Tweets... I guess because I @ed someone? 02:18 dcook Ahh, because I don't understand conventions I guess.. 02:18 dcook Don't know why it's not showing up in my UI 02:17 dcook Cheers 02:17 rangi https://twitter.com/minusdavid/status/854155774943154176 02:17 dcook Surely it show up under my tweets.. 02:16 dcook I haven't tweeted in too long... can't even find my own tweet now.. 02:16 dcook hehe 02:16 ibeardslee check twice, tweet once 02:15 dcook I was going to double-check! 02:15 dcook Ah balls 02:15 rangi heh, im not @rangi im @ranginui (just fyi) 02:13 dcook I think I can wrap my head around almost everything except copy cataloguing with RDF 02:13 dcook tweet sent 02:11 dcook Oh wait hackfest... those wouldn't be recorded 02:10 dcook What was the talk about? 02:10 rangi he was at Kohacon16 and gave a good talk at the hackfest 02:10 * dcook thumbs up 02:10 rangi drop him a tweet 02:10 dcook I mean... they're really the perfect people to talk to since they're interfacing with Koha too 02:09 rangi https://twitter.com/brinxmat 02:09 dcook I figure if I can chat to them... they might be able to clarify everything 02:09 rangi yup 02:09 dcook Rurik Greenall? 02:08 rangi rurik is the main project/manager tech lead 02:08 dcook Not enough experience to know for sure though 02:08 dcook Yeah, the LIBRIS records are intense... and I'm not sure if they're 100% correct.. 02:08 rangi lots and lots of others 02:08 dcook Do you know if that's all done by Petter or if they have others? 02:08 rangi (i meant libris not oslo) 02:08 dcook Didn't have enough time to totally go through it all 02:08 rangi ah cool 02:08 dcook And that saves to both the triplestore and Koha's marc database 02:08 dcook I think now they have an editor (Catalinker) 02:07 dcook Or maybe it used to be.. 02:07 dcook I took a look a bit at their github 02:07 dcook Mmm, I don't think so 02:07 rangi i think its still marc, that they render as rdf on the fly 02:07 dcook Need to talk to them too 02:07 dcook Not sure if they're still demoing or not 02:07 dcook I think LIBRIS might be too, but I'm not 100% sure yet 02:07 rangi only ones in the world afaik 02:06 rangi as they have a fully rdf based catalogue working 02:06 * dcook agrees 02:06 rangi you should really talk with teh Oslo people 02:06 rangi yep 02:03 dcook But that doesn't necessarily make sense.. 02:03 dcook And show that somewhere on your web page.. 02:03 dcook Using owl:sameAs or some alternative... perhaps you could define some rules to crawl that entity... 02:02 dcook Maybe you fill out some basic details for your indexer? 02:02 dcook Do you just do a owl:sameAs? 02:01 dcook You need a local record in your LMS so that your library patrons can find it 02:01 dcook You subscribe to Worldcat 02:01 dcook You've ordered "Good Omens" as you got a purchase request for it 02:00 dcook Well that's where my mind bends a bit.. 02:00 dcook Now if you were to import this record from OCLC... 01:59 dcook Although I suppose they could've constructed the entity for display purposes here too.. 01:59 dcook I reckon when they crawled the dbpedia URI... they must have run it through a mapper and only saved it with the mapped/filtered triples.. 01:59 dcook As that triple doesn't exist in dbpedia 01:58 dcook schema:name is something that OCLC must have generated 01:58 dcook rangi: The thing I find interesting with that example is the <http://dbpedia.org/resource/London> entry 01:58 dcook Probably wouldn't be that hard.. 01:56 dcook Seemingly with SPARQL's DESCRIBE, although I haven't played with that in terms of named graphs yet 01:56 dcook Although they do it in the downloads too 01:56 dcook And they've just aggregated them here on the page.. 01:56 dcook As are the "related entities" 01:56 dcook Looking at theirs... I feel like <http://www.worldcat.org/oclc/973430700> is probably in a named graph 01:55 rangi right 01:53 dcook http://www.worldcat.org/oclc/973430700 01:53 dcook I wish I could see what OCLC is doing behind the scenes.. as I like some of their examples 01:53 rangi probably 01:52 dcook I wonder if there's a nicer way to do subjects than schema:keywords.. 01:51 dcook That's pretty coool 01:50 dcook Nope 01:50 rangi http://linter.structured-data.org/?url=http:%2F%2Fdemo.mykoha.co.nz%2Fcgi-bin%2Fkoha%2Fopac-detail.pl%3Fbiblionumber%3D649%26query_desc%3Dkw%252Cwrdl%253A%2520fiish 01:50 rangi have you seen this? 01:50 dcook As you say, good starting point 01:50 dcook Yeah, I think they all use literals for the moment but that's OK 01:49 rangi yep a good starting point 01:49 dcook We could have those RDF statements in a triplestore... 01:49 dcook So yeah... we already have embedded RDF statements using HTML+Microdata... 01:47 * dcook gives a thumbs up 01:47 rangi we should use more of what is available here 01:47 rangi https://bib.schema.org/ 01:47 dcook We could use that in a triplestore as well 01:46 dcook So that's really good 01:46 rangi yep 01:46 dcook Well, in terms of microdata at least 01:46 rangi and it is extensible 01:46 dcook rangi: Yeah, that's what I mean 01:46 rangi we have schema.org support in koha already 01:42 dcook Hmm... maybe using the schema.org schema would be a good place to start 01:39 dcook hehe 01:38 kidclamp night 01:38 kidclamp go for it, I will talk with Joy and argue and respond :-) 01:38 dcook I might send out another email... a much shorter one 01:38 dcook Anyway, I won't keep you. Enjoy :) 01:38 dcook That's what I always say too 01:38 dcook hehe 01:38 dcook Oh man, I misread beertime as bedtime 01:38 kidclamp :D 01:38 kidclamp but he is the best one 01:38 kidclamp one 01:38 dcook You have kids? 01:38 dcook Yeah I wish our time zones coincided more too 01:37 kidclamp then bedtime 01:37 kidclamp it's my beertime 01:37 dcook That's my ideal bedtime :p 01:37 dcook 9:37pm, eh? Yeah I guess that's fair 01:37 kidclamp UTC -5? 01:37 kidclamp Vermont, North East USA 01:37 kidclamp I wish our time zones coincided when I was less tired :-) 01:37 dcook night in any case :) 01:37 dcook I don't even know where you live so I can't hassle you :p 01:36 kidclamp but I am off for the night 01:36 kidclamp supposedly at least :-) 01:36 dcook Now that I think about it... aren't we already using microdata... 01:35 dcook But then you might want to use your local schema that you could index easily.. 01:35 dcook Maybe too a library database like LIBRIS, Library of Congress, National Library of Australia, etc 01:35 dcook You'd maybe use owl:sameAS for linkedmdb and dbpedia.. 01:34 dcook Let's say we were cataloguing The Shining in Koha.. 01:34 dcook On linkedmdb they use dc:title 01:34 dcook Thte title is dbp:name 01:34 dcook Then on dbpedia: http://dbpedia.org/page/The_Shining_%28film%29 01:32 dcook I've seen this with LIBRIS... a mix of their own schema and established schemas 01:32 dcook http://data.linkedmdb.org/page/film/2014 01:30 dcook And surely there must be non-XML based mappings.. 01:30 dcook hehe 01:30 kidclamp so simple 01:25 dcook I guess that would look something like... kbv:Record/sdo:mainEntity/bf2:instance/bf2:title/bf2:InstanceTitle/bf2:mainTitle in xpath.. 01:25 dcook Just to get the title 01:25 dcook Then sdo:mainEntity then bf2:Instance then bf2:title then bf2:InstanceTitle then bf2:mainTitle 01:24 dcook At the top level we have kbv:Record 01:24 dcook But check this out: https://libris.kb.se/data/oaipmh/?verb=GetRecord&identifier=https://libris.kb.se/r93fv6w306l65x4&metadataPrefix=rdfxml 01:23 dcook Vocabulary mapping... and that's how we get LIBRIS into a format Koha knows.. 01:21 dcook Since it can handle any XML-based format 01:21 dcook In theory, Zebra could actually be used too. 01:20 dcook Well one way or another.. 01:20 dcook Actually, I think we'd still need a single target schema 01:19 dcook I keep thinking this is going to be so inefficient.. 01:19 dcook So that should make data validation a bit easier 01:19 dcook But I suppose subjects and predicates should all be in URI format 01:18 dcook I really dislike how you can't really make parameterized SPARQL queries 01:18 dcook I suppose that will have to be in the SPARQL.. 01:16 kidclamp agnosticism++ 01:15 dcook Not hard-coding things++ 01:15 dcook That's a good point 01:15 kidclamp I think as long as we have a way to index any specific scheam we can get away with being fleixible, it just means a ton of mapping work - chossing one schema and running with that doesn't preclude supporting others, it just means we focus the work in one place 01:14 dcook Well they've hindered us I think 01:14 dcook MARC is a great interchange format, but I think its limitations... 01:14 dcook To be honest, I think it's somethign that should've happened a long time ago 01:13 kidclamp yeah, I am up in the air about choosing a schema 01:13 dcook I'm being assimilated... 01:13 dcook Spanner came to mind before wrench 01:13 dcook Bloody hell.. 01:13 dcook That's the one 01:13 dcook https://xkcd.com/538/ 01:12 dcook That and the one about encryption 01:12 dcook wizzyrea: I think that's my favourite xkcd of all time 01:12 wizzyrea https://xkcd.com/927/ 01:12 dcook kidclamp: And I think that's vitally important in terms of indexing the data 01:12 dcook Whether Koha comes up with its own schema or uses an existing standard.. 01:12 dcook Folk like Oslo Public Library and LIBRIS use their own schemas... 01:11 dcook Without those mappings... I think it would be impossible.. 01:11 dcook And that makes sense 01:11 dcook I assume they're mapped somewhere in OCLC's back-end 01:11 dcook Instead of the schemas preferred by dbpedia 01:11 dcook While it points to http://dbpedia.org/resource/London, it uses the "schema" schema 01:10 dcook I think we can see that with OCLC 01:10 dcook "In order to understand as much Web data as possible, Linked Data applications translate terms from different vocabularies into a single target schema." 01:10 dcook That kind of comes upa t http://linkeddatabook.com/editions/1.0/#htoc84 01:09 dcook Because the question becomes... what vocabulary/vocabularies do we use in Koha? 01:09 dcook Those linkages using predicates that are pre-agreed upon 01:08 dcook Named graph for a Koha bib that is 01:08 dcook Then... create linkages automatically with a separate Koha named graph 01:08 dcook I use OAI-PMH to save the LIBRIS record to a named graph 01:08 dcook Mmm, or... instead of koha:derivedFrom or whatever 01:07 dcook Ouais, wahanui, ouais 01:07 wahanui somebody said Complicated was far too mild a term to describe Search.pm. 01:07 dcook Complicated 01:07 * dcook shrugs 01:07 dcook Not for updates 01:07 dcook Of course, I think this is a bit where old world and new world clash... shouldn't need OAI-PMH in theory 01:07 dcook We don't need a crawl to fetch them 01:06 dcook Since my OAI-PMH harvester is pushing up-to-date records into the triplestore 01:06 dcook So that we don't recrawl it... 01:06 dcook Maybe... koha:oai-pmh 01:06 dcook Something like that 01:06 dcook or koha:derivedFrom 01:06 dcook and then in Koha... just have something like owl:sameAs 01:06 dcook I should use the URI from LIBRIS 01:05 dcook But since Thursday I've been thinking... 01:05 dcook I've been thinking about saving those triples under a named graph with the name coming from Koha 01:05 dcook But it provides some problems.. 01:05 dcook Makes sense to me. Catalogue centrally is more efficient than duplicating effort across a country. 01:04 dcook LIBRIS being the national library's union catalogue 01:04 dcook So with Stockholm University Library, they're using OAI-PMH to get RDF records from LIBRIS 01:04 dcook But... I need to know how we're doing that a bit if I'm going to do that 01:04 dcook If I recall correctly, I'm supposed to just get the RDF in to the triplestore 01:03 dcook Arguably none of this is relevant for my work of course... 01:02 dcook hehe 01:02 kidclamp shhhhhh 01:02 kidclamp don't tell talljoy, sometimes i like the idea of marc records supplemented by linked data - use the work links to aggregate, but keep our march there as the base for searching etc 01:00 dcook Not sure how practical it is :/ 00:59 dcook I like the idea of linked data, but... 00:59 dcook Or maybe it's gone for good 00:59 dcook Maybe it's a case of a web app error and it will come back 00:59 dcook If you're crawling and you get a 404... what do you do? 00:59 dcook I mean.. 00:59 dcook Yeah I think about that as well 00:58 kidclamp please stop linking here 00:58 kidclamp need the RDF equivalent of 404 page - We deleted that info as it was old and moldy 00:57 dcook Even if we were manually updating the record on Dave... I don't think it would make sense to check at deletion time if anyone else is referring to that graph... as that would be a heavy processing job.. 00:56 dcook We have this cached Birmingham graph 00:56 dcook Because let's say that Dave moves away to the US from the UK. No longer lives near Birmingham. 00:56 dcook Checking every graph if it's referred to by another graph within the triplestore... 00:55 dcook I suppose you could have a cleanup job... 00:55 dcook Although it still seems like you could wind up with a lot of useless data over time 00:54 dcook Makes sense to me... I don't know how else you'd reasonably manage them.. 00:54 dcook It's time to refresh the data, so we start crawling the links... and we save the results to their own named graphs 00:53 dcook So in this example... let's say we have an authority record about "Dave Smith" 00:51 dcook Wonder if I hadn't read this before or just forgot about it since November 2015.. 00:48 dcook "Multiple Named Graphs can be represented together in the form of a graph set." that looks useful.. 00:47 wahanui wizzyrea: I forgot yeah 00:47 wizzyrea forget yeah 00:46 dcook http://linkeddatabook.com/editions/1.0/#htoc81 00:45 dcook And that is adding to my confusion.. 00:45 dcook I feel a bit like LIBRIS might not be using RDF "correctly"... 00:44 kidclamp interesting looking though 00:43 dcook Too bad the word semantic can be used in too many different ways :p 00:42 dcook But I don't think it's what I'm really looking for.. 00:42 dcook I'm looking at this at the moment: http://stanbol.apache.org/overview.html 00:42 dcook Too many browser tabs 00:42 dcook What was I thinking of.. 00:41 dcook I haven't looked at Kibana though so I'm just rambling 00:41 dcook Mind you, there's the whole "full text indexing" is all you need 00:41 dcook I do wonder sometimes though how sophisticated these searches are though 00:41 dcook :D 00:40 kidclamp you don't have to build a search enginge :-) 00:40 kidclamp es + kibana is most common, because it works out of the box with no config basically 00:40 wahanui i heard yeah, es was back working. Refactoring it to work with authorities at the moment 00:40 kidclamp yeah, es 00:40 kidclamp I like the 'percolation' feature - using a search of existing documents to classify a new document 00:40 dcook I met someone from ElasticSearch a couple years ago and meant to stay in touch but just been too busy :/ 00:40 dcook Mmm I'd seen the logging thing. Is that with Kibana? 00:39 kidclamp lots of statistics and logging stuff 00:37 dcook I mostly see ads for it in conjunction with other modules though yeah 00:37 dcook I admit I haven't kept up with ES too much 00:37 wahanui hmmm... Yeah is there a good way to fix that to get the new one running? 00:37 dcook Yeah? 00:37 kidclamp like ElasticSearch, we are using it for searching - that is the least of the things people use it for now 00:37 dcook Seems like 00:37 dcook Linked Data? 00:36 kidclamp I think the more it is implemented the more different things will be done though 00:35 dcook I mean... Zebra works but I don't think they quite knew how it worked at the start 00:34 dcook I wouldn't really want to repeat the Zebra debacle 00:34 dcook I feel like if I could just find one good production example... I'd have more confidence in the whole thing 00:33 dcook Although I guess caches are supposed to be throwaway.. 00:33 dcook Especially since for updates, you need to delete your existing cache 00:33 dcook Yeah, I'd like to see some real life examples of it though 00:33 kidclamp caching and regular crawling seems to be the only theory 00:32 dcook I think the keeping it up-to-date thing is what I struggle with most 00:32 dcook Then keeping that up-to-date... 00:32 dcook search/display... yeah 00:32 kidclamp yeah, you need to index the terms you are linking too if you want to search at all 00:31 dcook And potentially hardcode a lot of things... or make complex mappings 00:31 dcook it still seems like you need to do some work locally 00:31 dcook But other than that... 00:31 dcook And we all know you're talking about London, England once we've followed the links 00:31 dcook You can say "Oh, I'm talking about http://dbpedia.org/resource/London" 00:31 dcook Like that OCLC example.. 00:31 dcook But that you can point to a thing and say "Yeah... that's the thing I'm talking about" 00:30 dcook It doesn't seem like it's really meant to be interoperable per se... 00:30 dcook Or I'm misunderstanding it all haha 00:30 dcook Or maybe libraries are misunderstanding it all 00:29 dcook kidclamp: I'm so in agreement there 00:29 dcook So if you're using Lightroom or Darktable or whatever... you're using RDF O_O 00:29 kidclamp it feels like too much was done without enough thought on how it would actually work 00:29 dcook XMP sidecar files use RDF 00:29 dcook https://en.wikipedia.org/wiki/Extensible_Metadata_Platform 00:29 dcook Although little bit of trivia... 00:29 dcook That's the thing I hate about RDF... it's so loose 00:26 dcook Which makes it so much harder I think.. haha 00:26 dcook Yeah, that's another thing I've been thinking about 00:26 dcook Ah yeah I get you now 00:26 kidclamp that^ 00:25 dcook Or like "Work", "Instance", "Person" 00:25 kidclamp RDF type/class 00:25 dcook Are you meaning like auth, bib, hold or a different type of "type"? 00:25 dcook Hmm still not sure I follow 00:25 dcook This person: http://www.meanboyfriend.com/overdue_ideas/about/ 00:25 kidclamp basically if they had a bunch of things, each type of thing was a table 00:24 dcook It was suggested to me by.. 00:24 kidclamp any node that was a 'type' I think I mean class? 00:24 dcook kidclamp: Have I shown you this link? http://linkeddatabook.com/editions/1.0/#htoc84 00:24 dcook But I suppose if we use enough layers of abstraction, we shouldn't have to worry about that 00:24 dcook I suppose I'm OK with using a relational database for a triplestore... unless it introduces pecularities that make it impossible to move to a native triplestore 00:23 * kidclamp gets very bad at words when thinking talking RDF 00:23 dcook What do you mean by "any type was table"? 00:23 dcook any type? 00:23 kidclamp it is definitely difficult 00:23 dcook Nor should there be I suppose... but :S 00:23 kidclamp I was at code4lib and talked to the Boston Public Library guys who are doing a lot of work, they implemented everything via SQL - any type was table 00:22 dcook There's no "Delete: Cascade" 00:22 dcook But then what about all the triples that are linked from that URI.. 00:22 dcook Like... sure a URI can replace an ID in a relational database 00:22 dcook I find all the interconnections a bit mind boggling 00:22 dcook Although maybe it's not so different from relational databases.. 00:21 dcook I still find updates so... troubling 00:21 dcook It might make it difficult to create an API for the backend 00:21 dcook Although I figure until we know what we want to do with the backend... 00:21 dcook Yeah, I want it easy to switch backends as well 00:21 dcook hehe 00:20 kidclamp :-) 00:20 kidclamp 2 - we don't have to resolve our urls, it just seems way cooler to do so :-0) 00:20 kidclamp two things quick 1- my big concern is just that we make it easy to switch backends 00:19 kidclamp heh, as do we all likely 00:19 dcook I need to keep researching I think 00:19 kidclamp I need to reply to your post(s) by the way 00:19 kidclamp hah, that is a problem 00:19 dcook The post is from November 2015, so I guess it's not super recent.. 00:18 dcook And the two records I've linked to are already dead links :p 00:18 dcook kidclamp: Looking at an old blog post I made about linked data... 00:12 aleisha_ good thanks! 00:08 * dcook waves to folk 00:05 kidclamp pretty good, how's about you? 00:05 aleisha_ hows it going kidclamp 00:04 wizzyrea hi aleisha 00:04 kidclamp hi aleisha_ 00:04 aleisha_ hi all