IRC log for #koha, 2006-10-04

All times shown according to UTC.

Time S Nick Message
11:33 hdl ok.
11:43 kados hdl: still there?
11:43 dewey there is a minor diff in <div>s, that I missed
11:43 kados hdl: got a quick question regarding itemization of serials
11:44 kados hdl: it seems that barcodes and itemcallnumbers aren't recalled when editing back issues of a serial
11:44 hdl yes
11:44 kados hdl: have you seen this behaviour?
11:44 hdl this is a problem.
11:44 hdl This is because there is no "link" between serial and items.
11:45 hdl I am striving to solve this on rel_3_0
11:45 kados hdl: your client won't be adding barcodes to their serials?
11:45 kados hdl: or itemcallnumbers?
11:45 hdl Another problem would be for multiple items reception.
11:46 hdl kados : tey are.
11:46 hdl kados : they are.
11:46 kados hdl: but if they receive a serial, with a barcode and itemcallnumber, then next month, repeat, but the barcode and itemcallnumber aren't included for the back issues, it will break their items
11:47 hdl No.
11:47 hdl Becaus serial doesnot have an itemnum field a the moment.
11:48 kados hdl: it will add a new blank item then
11:48 kados hdl: blank meaning it will not have barcode or itemcallnumber, but will have serial info
11:48 hdl I should hide the display of items for old elements.
11:48 hdl would that solve your problem ?
11:49 kados hdl: only partly
11:49 kados there doesn't seem to be a way to update the Vol No info in an item record
11:49 kados using serials control
11:49 kados you can only update the serials tables
11:50 hdl for item records, you should use biblios.
11:50 hdl BUT.
11:50 hdl I thought on this problem.
11:50 hdl There are many problems to solve.
11:50 kados why not just add itemnumber to serials?
11:50 kados it would solve a lot of problems
11:51 hdl only partly.
11:51 hdl If we receive 4 times an issue number... you are doomed.
11:51 kados why?
11:52 hdl Because you want to say : OK I got this number in my library.
11:52 hdl And not see it 4 times in your statecollection.
11:53 hdl What I can propose is to write on the wiki what I thought of to manage this stuff.
11:53 kados ok
11:53 kados but I also think that we shouldn't advertise itemization of serials as a new feature
11:53 hdl And how this could be implemented and designed.
11:53 kados because it is clearly not ready for production
11:54 hdl I have done this on the rush. And couldnot think about linking and updating.
11:54 kados I have some cases of duplicated barcodes
11:54 kados with the one client I have using this
11:55 kados and items that can't be deleted because they have no itemnumber
11:55 kados but no way for the librarian to fix
11:55 kados (and even for me ... very difficult)
11:56 hdl I could work on this but maybe after authorities management.
11:56 kados right
11:57 hdl kados : Documentation about bib1.att and record.abs and gils on Indexdata site isOK for use on zebra1.3 ?
11:57 hdl Is there another Doc source ?
11:58 kados hdl: some of the doc is for 2.0, some is for 1.3 ;-)
11:59 hdl kados : How did you manage to build bib1.att and record and usmarc.mar and .chr files ?
12:00 hdl kados :Error /retry ?
12:00 kados hdl: it took many many months :-)
12:00 kados hdl: now you will be happy when I say we need to split the roles of koha developers ;-)
12:01 kados hdl: because I would vote myself as the Search manager :-)
12:01 hdl I would be happy only if you tell us that you cope with UNIMARC too :D
12:01 hdl which I doubt ;)
12:01 kados hdl: I will
12:02 kados hdl: but only after NPL goes live ;-)
12:02 hdl So I shall strive to understand what you did.
12:02 kados hdl: and only if we find someone else to do a good job on MARC21 Biblio management
15:27 owen kados: have you tested frameworks in dev_week?
15:28 kados owen: what part of frameworks? management of them?
15:28 owen Yeah. We had problems in previous versions with the process for editing frameworks
15:29 kados I don't trust the frameworks editor
15:30 kados I rely on carefully constructed sql files ;-)
15:30 kados why, is there something you'd like to change?
15:30 owen We need to set NPL up with a specialized framework for MORE items
15:30 kados ahh, right
15:30 kados ok, tell me a bit about it
15:32 owen Currently, MORE items are added with minimal information: title, itemtype.
15:32 kados just those fields?
15:32 kados plus an item I'm guessing
15:33 kados with barcode
15:33 owen Yeah, of course.
15:33 kados any other fields?
15:34 owen No. The records don't need to be searchable other than by title, since we put the MORE tracking number into the title field
15:34 kados why not put it in callnumber field?
15:34 owen The issue previously was that the title field seemed to be the best place to put that number in order for it to be searchable
15:35 kados that's no longer the case ;-)
15:35 owen I'm not sure it makes any more sense to make it a call number than part of the title, since it's a number like '29382103921'
15:37 kados right
15:37 owen We talked about setting up a MORE item type which wouldn't be searchable through the OPAC
15:37 kados yep, that can still be done
15:37 kados not sure it will happen before golive though
15:38 owen Lauren approved that, so it can go on the to-do list
15:39 kados approved the new itemtype?
15:40 owen There had be some question about statistics with the itemtype scheme, but she said it's fine
15:40 owen had been, I mean
15:41 kados right
15:42 owen You mentioned at the last Liblime meeting that there would be some new requirements for the catalogers to follow in order to get all the right data into records for the purpose of the new searching
15:42 kados yes
15:43 kados they already know most if them
15:43 kados we've been discussing it
15:43 kados what we're going to need to do
15:43 kados is put together a cataloging manual ;-)
15:43 owen Donna today talked as if she's still concerned about not having all the information she needs
15:43 kados ahh, good to know
15:43 kados I'll touch base with her
15:43 kados tomorrow
15:43 dewey tomorrow is our national day
15:44 kados hehe
15:44 owen dewey: forget tomorrow
15:44 dewey owen: I forgot tomorrow
15:44 owen We talked today about who all would have edit privileges on the catalog
15:45 owen That's why I'm thinking about this stuff now
15:45 kados it's tricky
15:45 kados because iirc some folks create MORE items
15:46 kados but we wouldn't want them to have access to the regular items, right?
15:49 owen In an ideal world we'd have several levels of cataloging access: Full, "Simple" (for things like MORE items), and deletion-only
15:49 kados right
15:49 owen Even the deletions can be tricky, although less so now if we've got Koha stopping deletions of items still on issue
15:49 kados in an ideal world I'd have a team of 12 programmers to work on koha :-)
15:50 kados owen: right
15:50 owen :) My point is that we're fine having only one level of catalog access because we understand we've got to work with what we've got
15:51 kados gotcha
15:55 tumer[A] owen:you can still delete a bibliographic record (together wit all the items) even if any is on issue , but don't tell them
15:55 kados tumer: not in dev_week
15:56 tumer have you updated i did not see that
15:56 kados tumer: yea, check out the DelBiblio routine
15:56 tumer k
15:57 tumer this new thing that came to head instead of dev_week is it?
15:57 kados don't think so
15:57 kados should be just in dev_week
15:59 tumer its in dev week
15:59 tumer we are all doing the same things differenly allover again
16:00 kados tumer: we are?
16:01 tumer yes i have written that just after dev_week to but somebody lost it
16:01 kados :(
16:01 tumer so i have it in head only
16:02 tumer i think barcode checking of items was lost as well
16:02 kados I recently discovered that in rel_2_2 (and probably elsewhere), rebuildnonmarc completely destroys items data
16:02 tumer you can edit a barcode and have duplicate barcodes
16:03 kados as in it takes the not-up-to-date marc data for items and puts it in items table, rather than using the authoritative data from the items table itself
16:03 owen Another thing to be fixed in 2.2.6b :)
16:03 kados tumer: I haven't tested that in dev_week, but in rel_2_2 it's definitely not permitted to have duplicate barcodes
16:03 kados owen: really?
16:04 tumer not when creating a new one
16:04 owen I mean it /should/ be.
16:04 tumer but when editing it does not
16:04 tumer i have that fixed in head
16:04 kados wow, that's something I didn't know
16:04 kados will have to make a note of it
16:04 owen I don't mean it /will/ be
16:04 kados is there a bug filed for it?
16:05 tumer i dont remember
16:06 chris hmm in the olden days, barcode was a unique column in mysql .. so you couldnt have duplicates mysql wouldnt let you .. i think that got ripped out at some point, and no one replaced it with checks in the code
16:06 tumer hi chris
16:06 kados hey chris
16:06 chris hi guys
16:06 chris seen this
16:06 dewey I haven't seen 'this', chris
16:07 chris ?
16:07 kados what I've seen a lot of recently, is items saved to the marc_subfield_table, but not to items table
16:07 chris yikes
16:07 kados so you end up with an item you can't use ... and can't edit or delete either
16:07 kados yea, it's a disaster
16:07 chris classy
16:08 kados chris: looking nice
16:08 chris was their launch yesterday
16:08 kados w00t
16:08 kados gonna do a press release?
16:09 chris there were 3 reporters there yesterday
16:09 chris plus the Minister for libraries
16:10 kados we need to be better about updating the news on the site while I think about it
16:10 kados chris: sweet!
16:10 chris so im just seeing what was written about it
16:10 tumer nice work chris
16:10 chris kados: yes we do
16:10 kados I'd submit some stuff but I can't remember how :-)
16:11 chris hmmm fire it at russ :)
16:11 chris thats how i do it
16:11 kados heh
16:11 kados hey russ
16:11 kados russ: can you remind me of how to access the editing interface for
16:12 russ how do
16:12 russ sure
16:12 russ two secs
16:12 russ to browse the edit site
16:13 russ to access the editor
16:13 russ the news editor is at
16:13 kados thx
16:14 russ
16:14 russ two secs i'll change it
16:36 russ kados you there?
16:38 kados russ: yep
16:39 russ when you add a new item for the news, check the homepage box otherwise it will just show up in the news section
16:39 kados ahh ... oops :-=)
16:39 kados sorry bout that
03:25 paul_away cd koha./nick paul_indispo
07:59 hdl kados ?
07:59 dewey hmmm... kados is becoming a true Perl Monger...
08:21 kados hi hdl
08:27 hdl hi how are you ?
08:27 kados a bit tired ... but otherwise, well
08:27 hdl Still trying to get an authority file into zebra.
08:27 hdl But *.att and *.abs files are quite a mystery to me.
08:28 kados :-)
08:28 hdl and dont know whether there are norms, since bib1 and gils seem to be generalized.
08:29 hdl or if we can build up our own.
08:29 kados I wrote a quite comprehensive bib1.att ... do you have that one?
08:30 kados at the top of the file I have:
08:30 kados # $Id: bib1.att,v 2006/09/18 20:02:47 kados Exp $
08:30 kados it adheres to bib1 as closely as possible, but I did expand on the standard in the 8XXX and 9XXX attribute sets
08:32 kados and my record.abs is likewise, very close to the standard
08:33 hdl yes but bib1 seems to be lacking some ways to question headings.
08:33 kados headings? you mean subject headings?
08:33 hdl If you want to query only $a heading or whole heading  or rejected forms ...
08:34 hdl You donot have enough distinctions between them.
08:34 kados the file I committed has some explaination of bib1, based on online bib1 syntax documents
08:34 hdl Maybe I want to have too much precisions.
08:34 hdl OK. I shall invest.
08:35 hdl I did get my stuf indexed.
08:35 kados cool!
08:35 hdl But yaz-client closes connexion before making any search.
08:35 hdl Target closed connection
08:35 hdl Do you know why this can occur ?
08:35 kados hdl: Mike R has already submited a UNIMARC patch :-)
08:35 kados hdl: check your email
08:36 kados hmm ... not sure why it closes connection
08:38 kados hdl: I have discovered something very important (a bug)
08:39 hdl kados: got this. I will try.
08:39 kados hdl: does not properly re-construct items data
08:39 kados hdl: so do not run it on any client data or items will be destroyed!
08:40 hdl what do you mean by  "does not properly re-construct items data"?
08:40 kados hdl: it seems to assume that the authoritative items data is in the MARC
08:40 kados hdl: rather than in the items data
08:40 kados hdl: items table I mean
08:40 hdl It does and it is the way rebuild non marc work.
08:41 kados morning owen
08:41 hdl It states the fact taht MARC values are more valuable than non-MARC ones.
08:41 owen Hi
08:41 dewey salut, owen
08:41 hdl I see no problem in that.
08:41 hdl Hi owen.
08:42 kados hdl: but in fact, it's incorrect for all versions of Koha except 2.2.6
08:42 kados hdl: because before one month ago, marc_subfield_table was never updated with new items information if items were edited
08:43 kados hdl: and moreover, marc_subfield_table has no understanding of how to properly order items data ... so if you have items ordered as Vol 1, Vol 2, etc ... it will re-arrange them randomly when you rebuildnonmarc
08:44 hdl Does it ?
08:44 kados yep
08:44 hdl What about subfieldorder ?
08:44 hdl or tagorder ?
08:45 hdl (tagorder is here for that purpose)
08:45 kados all I know is, I run rebuildnonmarc, and it moves items data around ... and items data is missing
08:45 hdl imho
08:45 kados I haven't investigated carefully why
08:46 hdl I ran rebuild over and over and didnot have such a problem.
08:47 kados in 3.0 I think we should make the items table the authoritative source of items data
08:48 kados since our MARC management is unreliable
08:48 hdl zebra closes connexion before yaz-client can make any search. Target closed connection
08:48 hdl kados :Do you know why this can occur ?
08:49 kados sorry, no
09:04 hdl have you read what I proposed on
09:04 kados not yet
09:05 hdl how can I analyse what is in db with zebra ?
09:05 kados hdl: there are some explainations already on the wiki
09:05 hdl To know if indexation didnot failed ?
09:05 kados[…]ramming_resources
09:05 kados Zebra Programmer Guide
09:05 kados has some useful tips
09:09 hdl As soon as I type f in a yaz-client connexion with db is closed.
09:13 hdl It does so only for authorities and not for biblios.
09:13 hdl Strange.
10:03 kados hdl: do you understand the MARCmarc2koha sub?
10:03 hdl yes.
10:04 kados it seems somewhat recursive in a way that I'm not completely comfortable with
10:04 kados because I'm not 100% sure of the behavior
10:04 kados for instance:
10:04 kados while (($field)=$sth2->fetchrow) {
10:04 kados        $result=&MARCmarc2kohaOneField($sth,"bibli​o",$field,$record,$result,$frameworkcode);
10:04 kados    }
10:04 kados the first time this is called, $result is null
10:05 kados right?
10:05 kados but $field is not
10:07 kados in rebuildnonmarc, MARCmarc2koha is used in a few places
10:08 kados sub localNEWmoditem has:
10:08 kados    my $olditem = MARCmarc2koha( $dbh, $record,$frameworkcode );
10:08 kados    C4::Biblio::_koha_modify_item( $dbh, $olditem );
10:08 kados $record contains only items fields
10:09 kados but if I understand correctly, MARCmarc2koha still stuff to biblio and biblioitems fields
10:09 kados it looks like someone was gonna check to see witch one to do
10:10 kados first line is:
10:10 kados my $sth=$dbh->prepare("select tagfield,tagsubfield from marc_subfield_structure where frameworkcode=? and kohafield=?");
10:10 kados but that $sth is never eecuted ...
10:10 kados and there is no flag passed to say which of biblio, biblioitems,items is to be pulled out
10:11 kados and it looks also like $result is overwritten every time something is added
10:13 hdl It seems to me that $result could be a ref to a hash.
10:13 kados hdl: or is there something magical happening because $result is a ref to hash?
10:13 kados hdl: you beat me :-)
10:13 hdl That will be completed over and over.
10:14 hdl In order to caintain biblio, biblioitem and items information.
10:14 hdl caintain stands for contain.
10:14 kados ok
10:14 kados I think I understand it now
10:14 kados but I am still confused about whether items table or MARC is authority on items data
10:15 kados here it seems that it's MARC that's authoritative ... but if something in items is unmapped, it is lost?
10:16 hdl If you merely overwrite informations from items table, then yes.
10:17 hdl But you can be careful,
10:17 kados for instance ... if datelastborrowed is not in framework, it is lost when rebuildnonmarc is run, right?
10:20 hdl seems so.
10:21 hdl because all fields are in OLDmoditem
10:22 hdl kados : have you tested work with more than One base on zebra ? Does this work ?
10:22 kados hdl: I haven't tested it
10:22 kados hdl: well ...
10:23 kados hdl: I have tested it, but not put data into more than one database on zebra
10:27 hdl the second database seems not receiving queries.
10:27 hdl tumer around ?
10:33 kados hdl: we almost need a 'rebuildmarc' script to add the stuff in items to MARC
10:33 kados hdl: because unless I'm mistaken, the MARC items data shouldn't be authoritative ...
10:34 kados hdl: for one, there aren't enough subfields in a tag to contain all the items mappings
10:34 kados hdl: also, at least in rel_2_2, we don't update the MARC data on circ like we do the items table
10:35 kados hdl: do you know if paul has any opinions on this issue?
10:35 hdl I think that what could be done in not UPDATE ALL the fields in items table but ONLY those who are found in MARC.
10:35 hdl paul_indispo
10:36 kados perhaps that is what is already done, I'm not patient enough to step through every line of the code ;-)
10:36 kados paul_indispo: are you present?
10:36 hdl (he is reading)
10:37 kados ok
10:37 kados in dev_week, I decided that items table would be the authority on items data
10:37 kados because updating items in MARC is too slow during circulation
10:40 kados so there is a periodic cron job that updates the MARC items data (in biblioitems.marc and in Zebra)
10:40 kados but i wanted to clarify what is done in rel_2_2 and rel_3_0
10:43 kados hdl: is paul_indispo with you?
10:45 hdl No he is managing a trainig and each time you beep him, client is aware ;)
10:45 kados :-)

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