IRC log for #koha, 2008-06-07

← Previous day | Today | Next day → | Search | Index

All times shown according to UTC.

Time Nick Message
14:15 hdl rch around ?
15:02 ryan hi hdl
15:02 hdl hi.
15:03 I posted you a message on your latest patch
15:08 ryan ah, ok, I see that I understand even less now what this was doing.
15:09 why do we store authtypecode as frameworkcode for a given subfield ?
15:09 what does it do ?
15:15 hdl: can you give an example of how it may be used ?
15:16 I see also that we are storing an authtypecode as frameworkcode, so there's a big bug here if it should stay, as well:
15:16 frameworkcode    | varchar(8)   |
15:16 authtypecode     | varchar(10)
15:16 hdl OK.
15:16 (on phone)
15:18 OK there is a bug should be varchar(10)
15:21 exemple :
15:21 if you want to make kind of hierarchy.
15:22 Europe  France Paris
15:22 you can either make on authority
15:22 OR
15:22 have authorities
15:22 Europe
15:22 France
15:22 Paris
15:23 gmcharlt hdl, ryan - in other words, this is a UNIMARCism that doesn't apply to MARC21
15:23 hdl And link tag 505 of France with Paris and Europe.
15:23 Don't you have any relations between authorities in MARC21 ??
15:24 BT and NT ?
15:24 gmcharlt yep, but not something that would be implemented by having authtypecode in the authorities frameworks
15:25 ryan hdl: so unimarc 505 is link to bt or nt ?
15:26 hdl gmcharlt: how would you do it ?
15:26 it is quite the same as linking authorities in biblios.
15:26 Should be the same process.
15:27 ryan: was just an example but yes.
15:28 gmcharlt hdl: well, for one thing, doing links between whole headings, not specifying that an individual subfield has a reference type while ignoring the subdivisions
15:28 hdl you would have 505$5g$9numauthority$aEurope 505$5h$9numauthority$aParis
15:29 gmcharlt: can you give example ?
15:29 our system with authorities linking is not accurate enough for UNIMARC.
15:30 one should/could link one subfield to an authority heading to allow composition.
15:30 But still, Nowadays system is better than nothing.
15:31 gmcharlt hdl: I think we've sufficiently determined that rch's patch should not be applied, as it supports functionaility needed for UNIMARC.  I agree that the $9 mechanism is not sufficient, but MARC21 authorities do not allow arbitrarily linking a subfiield to another authority - it would have to be a subdivsion authority records, and subdivions authority records can code for more than one subdivision at a time
15:32 in other words, at some time we'll need to rework authority and headings support to dig it out of its mountain of ad-hocness
15:33 hdl could you send authorities records sample for me to better take the git of what you say ?
15:34 Is there one in git atm ?
15:35 gmcharlt hdl; sorry, I really don't have time today to delve into this - maybe next week
15:35 hdl It is not so urgent.
15:36 koha30 is really much more important.
15:36 gmcharlt hdl: agreed
15:36 hdl But I just wanted to work this out with you.
15:36 If you want and can.
15:36 ryan hdl: i'll retract patch, and send a replacement to update database
15:36 to fix varchar(8).
15:37 hdl I have performance problems with page loading.
15:37 I think it is owed to all the css stuff and jquery...
15:38 ryan all pages?  or marc structure editor?
15:38 hdl quite all pages
15:38 But subscription and circulation and so on.
15:38 especially those.
15:38 Is there a way to fix it ?
15:39 Is there a way to improve performances ?
15:39 gmcharlt hdl: if you're running FireFox (2, not 3), try the yslow plugin and see if you can get more information (otherwise, it's a little hard to characterize the performance problem)
15:40 hdl I + tried to add some getJSON and ajax on subscriptionadd
15:41 But it is disappointing how slow this is
15:41 1s to load the result of a simple query on localhost.
16:01 atz hdl: yeah, it still will take a long time if you have to do the usual auth/context stuff on the back end
16:02 hdl Is there a way to make it faster ?
16:03 atz I use check_cookie_auth and my own "ajax_auth_cgi"
16:04 on the opac side
16:05 actually, same kind of thing on the staff side too
16:17 Man, I love the message I just read on the list....
16:17 "A National Workshop on Koha ILS was held at AIO University Islamabad last week, organized by PAKLAG..."
16:25 gmcharlt atz: yes; that was nice to see.  they also did something similar for OpenBiblio
16:35 atz wish they would use linux instead of windows though.
16:55 acmoore glad to hear I wasn't the only one surprised (but happy) about that.
17:56 nengard can someone tell me what the difference is between the two options for the SubscriptionHistory sys pref?
20:16 dkg hey folks--  i'm finally getting around to running the bulk import that gmcharlt advised me about a couple days ago -- it's pretty slow going, but i figure i'll just let it run in a backgrounded screen session.
20:16 in the meantime, i want to smooth out other bits of infrastructure.
20:16 the library i'm working with gets downloadable records from Follett Library Resources, who advertise that they support many different ILS formats for their exported data.
20:17 i'd like to get them exporting data that will align with the default koha MARC rules
20:17 where should i extract the koha-specific MARC field definitions to send to them?
20:17 gmcharlt dkg: here's a start: koha
20:18 dkg: oops - http://wiki.koha.org/doku.php?[…]for_vendors&s=952
20:19 dkg that's great! thank you gmcharlt.
20:19 does that cover everything that's koha-specific?
20:20 (until a couple months ago, i didn't realize that folks used different fields within MARC for the same purpose, so i'm learning here)
20:20 gmcharlt dkg: it covers the essentially item fields - if there are any more that you need to migrate, you can get the values by going to the Adminsitraiton menu and looking at the restuls of the Koha to MARC mapping page
20:22 dkg ok, i understand how to retrieve specific mappings -- i guess i'm wondering if the other mappings are standardized or not.
20:22 also, i'm a little concerned about 952‡y, which is marked as required for circulation.
20:22 how are vendors going to know what to put there?
20:23 gmcharlt dkg: 952$y would be the item type - vendors typically get told what values to put in for things like locaiton and item type
20:24 dkg gmcharlt: is there a default set of base item types that koha uses?
20:24 gmcharlt dkg: no, it's up to the library to define - there are sample item types available when you install, but those are just examples
20:26 dkg item type is used to distinguish (for example) a book from a CD, right?
20:26 gmcharlt that's one way of viewing
20:26 dkg I'm a little bit surprised that there are no official standards in the MARC world for how to do that.
20:26 can you give me another way to think about item type?
20:26 gmcharlt another way is regular loan vs. journal loan vs. reference loan vs. not for loan
20:27 i.e., not in terms of the item's physical format, but how it circulates
20:27 dkg ah.  i see.
20:27 and the physical format might be stored elsewhere.
20:28 thanks for answering all these newbie questions.
20:29 gmcharlt no problem (and yes, it is amazing that item record handling was never adequately standardized in MARC)
20:32 dkg should i be getting a ton of output from my misc/migration_tools/bulkmarcimport.pl ?  like one line per MARC field or subfield?
20:32 e.g.:
20:32 INDEXING :GRO at /home/dkg/src/koha/C4/Biblio.pm line 2303, <GEN201> line 463.
20:36 and to be clear: 952$y style "item types" have to match an entry in the "itemtypes" table, right?
20:37 gmcharlt dkg: you're using NoZebra, right?  yeah, those are expected and benign, though I ought write a patch to quell them
20:37 and yes, 952$y item types have to match the codes defined in the itemtypes table
20:38 dkg i'm pretty sure i'm using nozebra, though i only seem to get to work on this project in fits and starts, and it was a while ago.
20:38 how would i check?
20:39 i do actually have a zebrasrv running, fwiw.
20:40 gmcharlt dkg: check the value of the NoZebra syspref - if it's ON, you're not using zebra
20:43 dkg mysql> select variable,value from systempreferences WHERE variable = 'NoZebra';
20:43 +----------+-------+
20:43 | variable | value |
20:43 +----------+-------+
20:43 | NoZebra  | 1     |
20:43 +----------+-------+
20:43 1 row in set (1.24 sec)
20:43 so i guess i'm not.
20:43 if i was using zebra, would those messages go away?
20:44 gmcharlt yes; indexing would be more sophisticated, too
20:47 dkg Do you have a recommendation for when it's worthwhile to use zebra?  can you point me toward docs?
20:48 gmcharlt dkg: I recommend Zebra pretty to anybody who doesn't mind dealing with a couple daemon processes
20:49 dkg: but especially worthwhile if you have over 20,000 bibs or so
20:50 dkg ok.  i think we'll be counting at about 10K bibs, but i certainly don't mind a couple daemon processes.
20:51 i suppose that puts me on the cusp.  I'll try to sort it out and probably be back here asking more annoying questions.
20:51 gmcharlt ok
20:51 (and no, they're not annoying)
20:51 dkg should i expect any change in speed with zebra? (thanks for the reassurance about the questions)
20:52 gmcharlt searching and indexing should be faster
20:52 dkg OK, that's a good thing.
20:52 which daemon process do i need to read up on other than zebrasrv?
20:53 (or was "couple" a figure of speech?)
20:53 gmcharlt dkg: it would be either zebraqueue_daemon.pl or running rebuild_zebra.pl -a -b -z as a cronjob every couple minutes
20:53 the latter is probably better
20:54 dkg OK.  i'll read up and report back.  thanks for the help, as always.
20:55 gmcharlt you're welcome

← Previous day | Today | Next day → | Search | Index

koha1