Time  Nick      Message
11:21 kados     hdl: I'm here now
11:21 hdl       hi
11:21 kados     hdl: is paul here?
11:21 kados     it seems no
11:22 kados     hdl: could you briefly explain the 'Link' feature in MARC subfields constraints?
11:22 kados     hdl: fun :-)
11:22 hdl       QUERY : Title= "chaos" at /home/hdl/Koha/Code/savannah/koha-dev/koha//C4/SearchMarc.pm line 237., referer: http://localhost:8083/cgi-bin/koha/search.marc/search.pl?type=intranet
11:22 hdl       [Thu Feb 23 15:21:16 2006] [error] [client] Use of uninitialized value in subroutine entry at /usr/lib/perl5/site_perl/5.8.7/i386-linux/ZOOM.pm line 280., referer: http://localhost:8083/cgi-bin/koha/search.marc/search.pl?type=intranet
11:22 hdl       [Thu Feb 23 15:21:16 2006] [error] [client] Can't call method "option" on an undefined value at /home/hdl/Koha/Code/savannah/koha-dev/koha//C4/SearchMarc.pm line 242., referer: http://localhost:8083/cgi-bin/koha/search.marc/search.pl?type=intranet
11:22 hdl       
11:24 hdl       kados : link feature seems to allow ppl to search for data like the one selected but I donot know more.
11:24 kados     hdl: it is different than 'Search also' thohugh right?
11:24 hdl       When paul comes ack, he can explain
11:24 hdl       Yes.
11:25 hdl       Search also will search in other fields.
11:25 hdl       Link, will display a glass after author or subject for instance.
11:26 hdl       On which you can click to browse other biblio on the same author/subject
11:26 hdl       *kados : can you exmplain what I did wrong ?
11:27 hdl       my tab is in /usr/local/share/idzebra/tab
11:27 hdl       rebuild_zebra.pl doesnot seem to work :
11:28 hdl       Error 10000 : Connect failed
11:28 hdl       ZOOM error 10000 "Connect failed" (addinfo: "localhost:2100") from diag-set 'ZOOM'
11:29 hdl       updatedatabase worked fine... apart from some  : no mapping found for some characters.
11:30 hdl       and DBD::mysql::db do failed: The used table type doesn't support FULLTEXT indexes at ./updatedatabase line 1029. for bibliosubtitle or bibliothesaurus
11:31 kados     interesting
11:31 kados     that DBD error is quite strange
11:32 kados     I don't think we're doing any FULLTEXT indesing
11:32 kados     indexing even
11:32 kados     hdl: how did you start zebra?
11:32 hdl       zebrasrv @2100
11:32 kados     you need to use:
11:32 kados     zebrasrv localhost:2100/kohatest
11:32 kados     where kohatest is the db name in koha.conf
11:33 hdl       I followed your notes ;)
11:33 kados     yea, they are a bit outdated :-)
11:33 kados     I will write a new set very soon
11:39 hdl       one need to launch a zebra server before rebuild_zebra.pl
11:39 hdl       Then run zebra index engine
11:39 kados     correct
11:39 hdl       ?
11:39 kados     start zebra
11:39 kados     using above command
11:39 kados     then run rebuild_zebra.pl
11:39 kados     thd: are you using latest cvs?
11:40 hdl       I(hdl) am now
11:40 kados     oops .. I mean hdl :-)
11:40 hdl       :)
11:40 hdl       :D
11:40 kados     heh
11:40 thd       kados: not since last night.
11:40 kados     thd: morning :-)
11:40 kados     thd: http://www.loc.gov/marc/bibliographic/nlr/nlr00x.html
11:40 hdl       morning thd
11:40 kados     thd: been thiking about that list and Koha's current frameworks
11:40 thd       good morning kados
11:41 kados     thd: it would require many many frameworks to handle all those requirements
11:41 hdl       kados : why rm -f biblios ?
11:41 kados     hdl: not anymore
11:41 kados     hdl: those instructions are too old to use
11:42 kados     hdl: are you converting a 2.2.5 system to use zebra?
11:42 hdl       2.2.4 in fact
11:43 kados     same thing
11:43 kados     so here's what to do:
11:43 kados     check out latest head cvs
11:43 hdl       done
11:43 kados     edit koha.conf to include:
11:44 kados     zebradb=localhost:2100/kohatest
11:44 kados     at the bottom
11:44 kados     go to cvs koha directory
11:44 kados     type:
11:44 kados     ln -s misc/zebra/unimarc zebra
11:44 thd       kados: you would need about 25 frameworks.
11:45 hdl       now done
11:45 kados     cd zebra
11:45 hdl       ok
11:45 kados     zebraidx update biblios
11:45 kados     you may need to 'mkdir lock' first ... it will warn you
11:46 kados     that command tells zebra: OK ... this is a valid zebra database ... set up the basics for me
11:46 kados     when update happens successfully (hopefully you have no records in biblios, I forgot to mention)
11:46 kados     then type:
11:46 kados     zebrasrv localhost:2100/kohatest
11:46 kados     it will start zebra
11:47 kados     NOW ... you are ready to attempt to rebuild_zebra.pl
11:47 kados     let me know when you get that far
11:47 hdl       zebraidx update biblios : what does it do ?
11:47 kados     hdl: that command tells zebra: OK ... this is a valid zebra database
11:47 kados                    ... set up the basics for me
11:47 kados     you could also say:
11:47 kados     zebraidx init
11:47 kados     same thing
11:47 hdl       I did so.
11:48 kados     wait ... it's not the same
11:48 kados     I believe you must use zebraidx update biblios
11:48 kados     do you have a biblios directory?
11:48 kados     it should be a blank directory
11:48 kados     in the same location as zebra.cfg
11:48 hdl       But it answered :
11:48 hdl       14:24:36-23/02 zebraidx(22194) [log] zebra_start zebra.cfg 1.3.32
11:48 hdl       14:24:36-23/02 zebraidx(22194) [warn] open lock/norm..LCK [No such file or directory]
11:48 hdl       14:24:36-23/02 zebraidx(22194) [warn] open lock/shadow..LCK [No such file or directory]
11:48 hdl       14:24:36-23/02 zebraidx(22194) [log][app2] zebra_stop
11:48 hdl       
11:48 kados     mkdir lock
11:49 kados     (see above comments :-))
11:49 hdl       I hadnot at that time.
11:49 hdl       Now, I did.
11:49 hdl       But I launched zebrasrv
11:50 hdl       and rebuild_zebra
11:50 hdl       (Logs are impressive !!!)
11:50 kados     bbiab
11:59 kados     hdl: ok ...
11:59 kados     hdl: so now you are ready to attempt to run rebuild_zebra.pl
11:59 kados     make sure you export KOHA_CONF
12:00 kados     run it and see what happens
12:00 kados     cool!
12:00 kados     and it's showing up in the log?
12:00 hdl       Quite long...
12:00 kados     yea, it's slow
12:00 kados     in part because it creates a new connection for each update
12:00 kados     this week I plan to fix that
12:00 kados     and do speed tests
12:03 kados     thd: so ... getting back to the frameworks
12:04 thd       yes kados
12:04 kados     it seems like we need two for each materials designation
12:04 kados     one for original cataloging
12:04 kados     and one for editing existing records
12:04 kados     at least with the current scheme ... right?
12:05 thd       kados: I do not understand the reason for that
12:05 kados     thd: because take koha.liblime.com for example
12:06 kados     thd: if you try to edit an existing record, it will replace many of the fields with new values
12:06 kados     thd: and I'm not sure I have control over when a plugin works and when it doesn't
12:07 thd       kados: We need a minimal level and a national level.  The national level is merely to prove that national level will work on the Koha record editor.
12:07 kados     right ... we still need subfield repeatability and reordering though
12:08 kados     but I don't think we can have a single framework for the whole national level
12:08 kados     I could be wrong though
12:08 thd       kados: The plugins should test for empty values and only overwrite what is empty.
12:08 kados     thd: yes I agree, but I'm not sure they can do this currently
12:08 kados     thd: it works already for some fields, like 'date'
12:08 thd       kados: about 25 frameworks
12:08 kados     thd: is that going to confuse catalogers?
12:10 thd       kados: It should be one framework that can be transmuted by changing control field values but that is a long term goal.
12:11 thd       kados: Which fields have the problem to which you are referring for new and existing records.
12:11 thd       ?
12:11 kados     thd: all fields
12:11 kados     wait i might be wrong here
12:12 kados     I guess it's just 003
12:12 kados     we should use a different plugin for that anyway
12:12 thd       kados: your plugin for 008 seemed to accept the values that were already there.
12:12 kados     thd: yep, it does
12:12 kados     thd: are there no indicators shown  on the current editor?
12:13 thd       kados: 003 does not need to change in bibliographic records but continually repopulating it would cause no problem.
12:14 thd       kados: the indicator positions are there, at least the last time I looked.
12:14 kados     http://www.loc.gov/marc/bibliographic/nlr/nlr6xx.html
12:15 kados     why does the minimal level not include the 600s?
12:16 thd       kados: Indicators are troublesome because they do not yet have a link for plugin management in the framework.  Only manual entry and value lists are provided.
12:16 thd       kados: one value list for two indicators is a problem.
12:17 thd       kados: I also hope that the blank indicators for control fields using the '@' subfield hack are being discarded.
12:19 thd       kados: what was the last link meant to show?  That subject headings are optional at the minimal level?
12:19 kados     thd: yea
12:19 kados     thd: how is that possible?
12:20 kados     thd: does Koha conform to the minimal level now?
12:20 kados     thd: well, I mean koha.liblime.com of course
12:21 thd       kados: That is not a US practise but it does go on in some countries where the library science tradition is about correctly identifying the edition but not subject guided access.
12:22 thd       kados: Subject guided access needs an expensively maintained thesaurus to work well.
12:23 thd       kados: Actually the subject thesaurus is an invention of the last century.
12:24 thd       kados: The oldest items at LC have no subject headings only classification by the LC Classification System.
12:27 thd       kados: LCC dates from the beginning of the last century but LCSH are closer to mid-century.
12:29 paul_away hdl ?
12:29 thd       kados: The disparity of library science practise causes problems for the IFLA where both rich county and poor country practise must be accommodated.
12:30 thd       paul: did you wee my question from much earlier this morning?
12:31 paul      nope. I had dozens of "connection lost" because of a connection failure.
12:32 thd       paul hdl: Do you plan for framework specified subfield repeatability to be useful in the record editor for 2.2.6?
12:35 thd       paul: Is that prose too turgid for understanding?
12:36 kados     paul: I have beeen playing with subfield reordering
12:36 kados     paul: I think I have devised a better scheme that prevents ordering outside of a tag
12:36 kados     paul: my demo has it currently turned off because I have a web demo to a client later this afternon and it's not quite stable
12:37 kados     paul: but I"ll show you what I have later (or tomorrow most likely)
12:38 thd       paul: I know you had expected to have subfield reordering working for 2.2.6, however, I am asking about adding additional repeated subfields.
12:39 thd       kados: did you see my comments about your approach to subfield ordering late last night?
12:40 kados     thd: no
12:40 thd       kados: I had been disconnected and then reposted them.
12:42 kados     thd: didnt see them
12:43 thd       23/02/06 20:04:09+-5	<thd:#koha>	kados: In addition to +- signs for subfield repeatability there should be an add more subfields link that brings up a pop-up to add a group of repeatable subfields selected from a checkbox according to a sequence specified in a framework parameter.
12:43 thd       23/02/06 20:04:30+-5	<thd:#koha>	kados: Subfield order changing should allow the insertion of any subfield before an already completed subfield. I am uncertain that the example you showed for order changing was heading in that direction for already completed subfields.
12:43 thd       23/02/06 20:04:49+-5	<thd:#koha>	kados: actually a pop-up insertion form accessible from a link before and after every subfield is probably easier than moving arrows around. Moving arrows is how it is commonly done in user interfaces but not necessarily the most efficient.
12:43 thd       23/02/06 20:10:15+-5	<thd:#koha>	kados: If you do use arrows for moving any completed subfield content should also move.
12:43 thd       23/02/06 20:13:37+-5	<thd:#koha>	kados: The issue is not only about creating content from a completely empty record but adding additional in the midst of an already existing record. in the midst of an already existing set of subfields.
12:44 kados     thd: the last design I worked up moved the content as well as the designator, etc.
12:45 thd       s/additional/additional subfields/
12:46 thd       kados: That is fine then as long as we can get more new repeatable subfields in there somehow.
12:46 kados     thd: explain why we need a pop-up
12:46 kados     thd: because sometimes you would want:
12:46 kados     $a
12:46 kados     $x
12:46 kados     $a
12:46 kados     $b
12:46 kados     ?
12:47 kados     not always a duplicate of the same one?
12:47 thd       kados: It is not a need but just what I thought might be more efficient.
12:49 thd       kados: you have the basic concept right about why I thought a pop-up might be good or an insert more link but not limited to inserting another copy of the single adjacent subfield.
12:50 thd       kados: you often want supplementary groups of subfields that work together in a particular order.
12:51 thd       kados: look at the multiple publishers 260 and the formatted table of contents 505 examples in bug 997.
12:53 thd       kados: especially the 505 example would be rather tedious to create one repeated subfield at a time and then moving it into the correct relative position.
12:58 thd       kados: The most efficient way to create the 505 example would be to populate the field with many author title pairs ordered by a framework ordering default for that field and then cut and paste text from an OCR scan of the table of contents.
13:01 thd       kados: you should create anything that is easy enough to create valid records.  A more efficient scheme can always be added later.
13:02 kados     thd: paul just showed me how to update information from one field to another using a plugin
13:03 thd       kados: just try and avoid blinders about how every friendly user interface reorders elements in a list.
13:03 kados     thd: lets pospone the repeatability problem and figure out which fields/valus should be auto-set based on other fields
13:04 thd       paul: I have been urging kados and owen to look at that and authorities handling since many months ago but they had no time then ;)
13:05 thd       paul: It is certainly the single best feature of Koha.
13:06 thd       paul: It is something that every system should have yet without almost a single exception none do.
13:07 kados     my opinion is that it's a nice feature, but realistically, for us libraries, Koha's editor isn't up to par yet
13:07 kados     using an external editor is still a good idea
13:07 kados     if you want to ensure good data in your db
13:07 kados     now ... hopefully, with 2.2.6 or soon after, that will change
13:08 kados     but right now, i think Koha's editor is 'entry-level'
13:08 thd       paul: If we could get element repeatability and ordering fixed in the record editor working well then Koha would be a generation ahead of other record editors.
13:08 kados     thd: don't other editors do this already?
13:08 thd       kados: no others do not to the extent that I have investigated.
13:10 thd       kados: the best that I have seen have context sensitive pop-up help files for the field that you are on.
13:10 kados     TLC's editor is pretty nice
13:10 kados     and I'm pretty sure it handles subfield repeatability
13:11 thd       kados: I just described the TLC editor but it is mostly a free text editor.
13:11 kados     yep
13:12 thd       kados: I told them what they ought to do but they had not even thought of that.
13:12 kados     thd: OK ... just read 997's 505 paragraph
13:13 kados     thd: of course, your criticism of that record isn't quite on, since Koha would preserve subfield order in that case
13:13 kados     thd: even if not displaying it as expected
13:14 thd       kados: the preservation and display issues for 997 have been mostly corrected thanks to paul.
13:14 kados     thd: i think some judicious use of javascript could really make the process of editing a record much easier
13:15 kados     thd: 'mostly' ?
13:16 thd       kados: The detail view and the ISBD view have problems with repeated subfields outside of 6XX which has been fixed.
13:17 kados     thd: there are no preservation issues to my knowledge
13:17 kados     thd: we should do a test
13:17 thd       kados: Also creating original record content and editing existing content are still issues.
13:17 kados     thd: import a record using bulkmarcimport
13:17 kados     thd: then export again and do a 'diff' on it
13:18 kados     thd: (use of an external cataloging client solves that problem)
13:18 kados     thd: (until we fix the framework)
13:19 thd       kados: I attached a comment to 997 long ago qualifying the original statement which had been underinformed.
13:19 kados     yep
13:20 thd       kados: one record is not a sufficient test.  Many varied and troublesome records might be.
13:21 kados     right
13:22 thd       kados: I need to write the sole remaining LC CDS technical services employee back to obtain a set of bibliographic test files.
13:22 kados     thd: can we get a case scenerio for auto-updating a fixed field from another field's value?
13:23 kados     thd: ie, an example of when this should happen and how?
13:23 thd       kados: you had two from yesterday.
13:24 kados     I dont' remember them ...
13:24 thd       kados: illustrations and content elements for books to use in 008.
13:26 thd       kados: illustrations would be coded in 300
13:27 thd       s/coded/abbreviated/
13:28 kados     right
13:29 thd       kados: 300 $b
13:32 thd       kados so you regular expression search of 300 $b would be looking for ~=/ill[.a-z]/  and many other things.
13:35 thd       s/ill\[\.a-z\]/ill[.u]/
13:36 kados     right
13:36 thd       kados: also ~=/col[.o]/ for variations of colour etc.
13:37 thd       kados: ill. and col. are the standard AACR2 abbreviations
13:38 thd       kados: although you may not want to miss the typos :)
13:41 kados     is there a comprehensive list of possible values for 300$b?
13:42 kados     I can't find one ... would that be in AARC?
13:43 thd       kados: it is rule 2.5C1-2.5C2.
13:43 kados     is AACR3 out yet?
13:43 kados     ahh ... coming in 2007
13:45 thd       kados: when you were becoming very fatigued yesterday I explained that AACR3 is a complete rewrite where the committee threw out the first draft and had to start over because of too many objections.
13:46 kados     heh
13:47 thd       kados: Whatever the virtues and vices of AACR3, it will not cure the legacy records.
13:55 thd       kados: common 300 $b usage may not always be sufficient for good completion of BOOKS 008/18-21 but it can succeed most of the time.
13:56 kados     thd: found a complete list:
13:56 kados     http://www.itsmarc.com/crs/Bib0379.htm
13:56 kados     thd: this will be quite easy to do
13:56 kados     thd: though I think I'll need to rest a bit first
13:56 kados     thd: still feeling pretty sick
13:57 thd       kados: 008 was a relatively recent addition to USMARC in the grand scheme of MARC time and initially had to use record parsing of textual data to apply it to old records retrospectively.
13:59 thd       kados: try to avoid exacerbating your condition when work is too much fun.
13:59 thd       :)
14:19 kados     thd: I found and fixed the problem with | btw
14:19 kados     thd: can you construct a default leader value string?
14:19 kados     thd: I'll reintroduce | into the plugins as the default value where applicable
14:21 thd       kados: I gave you default values for minimal level books records yesterday.  Do you want them again?
14:22 thd       kados: the existing leader string default is already what I had suggested to paul.
14:25 thd       kados: he made a leader plugin for MARC 21 using the default I suggested because the leader is almost identical between MARC 21 and UNIMARC so the extra MARC 21 plugin was a trivial amount of extra work.
14:33 kados     thd: the default leader plugin makes no use of |
14:34 kados     thd: the default value is:
14:34 kados     "     nam         7a  4500"
17:13 kados     thd: BWS Johnson just wrote:
17:13 kados     The 650 x problem is a big deal....Now, if there is away to make the subject subheadings play nice with their bigger brothers, let
17:13 kados     me know
17:14 chris     is that what i fixed the other day?
17:15 chris     http://opac.koha3.katipo.co.nz:81/cgi-bin/koha/opac-MARCdetail.pl?bib=29
17:15 chris     that bit?
17:15 kados     dunno
17:16 kados     didn't that just put the '650' back in?
17:16 chris     could be
17:16 kados     or did it affect ordering too
17:16 chris     probably not
17:16 chris     couple of nice emails from stephen and brooke though
17:17 chris     nice way to wake up
17:23 thd       kados: what is meant by the 650 X problem?  Is that the economical MARC view that you made the preference with a standard value to avoid? Therefore the same problem to which chris is referring?
17:25 thd       kados: What you had created the preference for was presentation only where the underlying data is valid.
17:26 thd       kados: Do you want my SQL dump of marc_subfield_structure so far?
17:28 thd       kados: I am having to add most of the fields in the MARC 21 specification.  Most of the fields were in fact missing from the default framework.
17:30 thd       kados: I still have much more to do.  I will need to actually rest soon. :)
18:02 kados     thd: did you just try to call?
18:11 thd       kados: It was not I.
18:12 chris     not me either
18:12 chris     :)
18:32 kados     hehe
00:02 kados     woo ... back online
00:02 chris     yeah we think bad ram
00:42 thd       kados: I have thought that filling every unused field and subfield in unmanaged fields and subfields will preserve all obsolete and local use data that people might have.  However, I am uncertain if that may introduce a noticeable performance drag on the framework.
00:44 thd       kados: This would be to protect against data loss when adding items or otherwise editing records.
00:50 kados     thd: yep
00:51 kados     thd: it becomes quite a large form element
00:52 thd       kados: I do not understand your last sentence.
06:15 hdl       hello
06:28 osmoze    bonjour hdl :)
06:29 hdl       bonjour osmoze.
06:29 paul      hello les frenchies
06:30 osmoze    :)
07:01 hdl       Someone using mediatypetable around ? chris kados ???
07:02 hdl       paul ?
07:02 paul      nope
07:03 osmoze    paul, il y a combien d inscrit sur la liste fr maintenant ?
07:03 paul      123. Ca stagne un petit peu
07:04 paul      (on voit quelques désabonnements, et quelques abonnements, qui s'équilibrent)
07:04 osmoze    la derniere annonce sur la liste etait de 60, donc c est toujours ca :)
07:04 paul      181 sur la liste annonce.
07:04 osmoze    et sur la devel, tu peux le savoir ?
07:05 paul      tu retardes, il me semble que j'ai récemment souhaité la bienvenue sur la liste à la personne qui a écrit juste après être inscrit et qui étit la 120°
07:05 osmoze    roh, aurais je loupé un mail ^^
07:05 paul      non (c'est joshua l'admin). quelques dizaines
07:05 osmoze    ok merci :)
07:06 paul      dans la liste koha anglaises, il y a plusieurs centaines d'abonnés.
07:06 paul      genre 500
07:06 osmoze    ca fait pas enorme si c est la liste internationale :(
07:07 osmoze    en comparaison
07:08 paul      (pas sûr du chiffre)
07:08 paul      cela dit, ce n'est pas anormal. Nous nous adressons à une niche très spécifique.
07:09 paul      et nous n'aurons jamais 50 000 abonnés !
07:09 paul      (et puis je vais être mauvaise langue : contrairement à d'autres, nous n'annoncons que des chiffres vérifiables et minimums)
07:10 paul      (pas comme certains autres SIGB libres qui annoncent des milliers d'installation)
07:14 osmoze    cad ?
07:14 osmoze    du style pmb ?
07:15 paul      c'est pas moi qui l'ai dit !