Time  Nick  Message
15:22 kados morning owen
15:23 kados well ... afternoon I guess
15:23 thd   good afternoon kados
15:28 thd   kados: are you awake now?
15:42 kados thd: yep
15:44 thd   kados: I have supplied you with a more complete and accurate ISBD setting at http://kohatest.liblime.com/cgi-bin/koha/admin/systempreferences.pl?tab=Catalogue  You should use that one with your customers.
15:47 kados thd: yes I noticed thanks!
15:47 thd   kados: Although, maybe few care about ISBD in the webified world.
15:47 kados actually, I've had some folks very interested in it
15:47 kados especially since it is so easy to change the way it displays on the page
15:47 kados so they can easily edit the OPAC display (if ISBD was the default)
15:50 thd   kados: There are still some problems with repeated subfields in the 2.X design that prevent correct operation of the ISBD rules when repeated subfields are present.  Also, ISBD has nothing specified for holdings so that is not present in what I supplied.
15:50 kados thd: ooh, I didn't realize that
15:51 kados thd: I just read over your SQL email in more detail
15:52 kados thd: I'm going to wait a few days, to see if anyone else has comments
15:52 kados thd: and then I'm going to recommend adoption for version 3.0
15:52 thd   kados: The repeated subfields problem is seen where for example there are multiple publishers for the same book.
15:52 kados right
15:53 kados thd: did you notice that some of the subjects aren't searchable in the new opac
15:53 thd   kados: There have been a couple of comments on koha-devel that I have not had time to reply to yet.
15:53 kados http://opactest.liblime.com/cgi-bin/koha/opac-detail.pl?bib=22762
15:54 kados neither of those subject links work
15:54 kados I think I need to change the default MARC21 settings to fix this
15:56 thd   kados: yes, a bibliographic framework change is required for searching all the relevant fields subfields that may appear in 6XX
15:58 thd   kados: Also, I never committed the code that I had suggested to correct the problem you had where the indicator had been insufficient to define the thesaurus used in 6XX
15:58 thd   as in the case of Sears.
16:23 kados thd: could you at least send it to me?
16:23 kados thd: and do you happen to know what subfields should be included in the framework?
16:23 kados thd: I could commit them right away so they will be in 2.2.6
16:25 thd   kados: I would need to search for the code.  I had a better revised form that I posted on #koha after you had lost attention.  I would need to search for that.  I will commit it when I find it so that t can appear in 2.2.6
16:29 thd   kados: I am working now to extract all the needed field subfield pairs for use in the bibliographic framework 6XX.  My present setting only covers the 650 subfields.
16:43 kados thd: thanks
16:59 kados thd: should I commit your ISBD modifs to cvs?
16:59 kados thd: as the default ISBD settings?
17:00 kados thd: or are those usmarc specific?
17:05 thd   kados: yes, I was thinking of doing many things like that.  There are UNIMARC defaults that set many bibliographic framework settings and other preferences.  There should be some MARC21 defaults starting with a more complete bibliographic framework that includes 000 - 008.  I had been working on setting something up to do that for USMARC but had only been studying how it was done with UNIMARC and became overwhelmed by the nonsense in Newark Mun
17:08 thd   kados: looks like my last post was truncated at "Newark Municipal Court".
17:12 thd   kados: but the answer is that my ISBD encoding is MARC21 specific and takes some small readability liberties with the ISBD format for newlines where multiple subject fields are present.
17:12 kados ahh
17:13 thd   kados: There should, however, be MARC 21 defaults just as there are UNIMARC defaults.
17:13 kados yep
17:14 thd   kados: There is no default UNIMARC ISBD encoding.  It took me a day's work to make the one for MARC 21 thoroughly.
17:43 thd   kados:  I have updated the bibliographic framework at http://kohatest.liblime.com/cgi-bin/koha/admin/marc_subfields_structure.pl?tagfield=650&frameworkcode= .
17:44 thd   kados: However, the example that you gave for subject matching still has no matches.  Do you have another example?
17:45 thd   kados: Do you have another example record with subject subdivisions?
18:12 kados I don't off hand
18:25 thd   kados: what data set do you use to stock a demo?
18:29 kados thd: in this case, the bib data is a subset of some public library data from one of our clients
18:32 thd   kados: what general subject area do they have with a large quantity of material?
18:33 thd   kados: maybe something regional realating to there location.  Where are they?
18:34 thd   s/there/their/
18:38 kados thd: you should be able to tell by looking at the MARC view
18:52 thd   kados: The special Liblime template you have has a problem that was fixed for the NPL templates after I told owen to change the search method.  Your template needs updating.
18:53 kados thd: could you be more specific?
18:59 thd   kados: actually, that was a guess and it seems I guessed wrongly.  I will try changing the OPAC template to CSS just to be certain.
19:06 kados k
19:17 thd   kados: run rubuildnonmarc.pl .  The indexing is not working correctly on you test demo.
19:18 kados k ...
19:20 thd   kados: If I am correct you did not run rebuildnonmarc.pl after setting 650a to be the values stored for subjects in koha SQL.
19:20 kados right
19:20 kados don't remember setting that actually
19:22 kados thd: running now
19:22 kados thd: it's taking forever
19:22 kados about 5 per second
19:23 kados prob about 50K items in this db
19:23 kados :-)
19:24 thd   kados: maybe 650 linked with bibliosubject.subject is the default for MARC 21.
19:24 thd   if you had not changed that value.
19:25 thd   s/650/650a/
19:25 kados I'm not sure I touched the default values
19:25 kados don't remember doing that for this collection
19:27 thd   kados: did you import the records before checking that the MARC Check was OK?
19:27 kados thd: don't remember :-)
19:27 kados thd: why?
19:28 thd   kados: I believe there maybe unknown problems with the system if you import records without having a validated MARC check.
19:30 thd   kados: Whenever I discovered that I had done that I always used bulkmarkimport.pl to purge and reload my records :)
19:45 thd   kados: Is this the same library that you were having trouble with for some of its usage of Sears subject headings?
19:46 kados probably
19:46 kados we fixed that right?
19:46 thd   kados:  I mean the same records :)
19:46 kados yea
19:46 kados same records
19:47 thd   kados: We fixed that with code that was neve committed.  I will search for that fix but that is not the problem at the moment.
19:48 thd   kados: Has reuildnonmarc.pl finished running?
19:48 kados thd: not even close :-)
19:48 kados about 1/5 done
19:49 thd   kados: what is the specification of your server.  I use such small sets of records I cannot see what a real load might be.
19:50 kados it's a pIII 2.4 Ghz with 2 gigs of ram
19:51 thd   kados: how many records are in this data set?
19:52 kados 50K
19:54 thd   If rebuildnonmarc.pl reports a running time, please let me know what that actually is when it is finished.
19:54 kados sure
19:57 thd   kados: presently, all subject searches fail, even the simplest of them directly from the search form.  I hope rebuildnonmarc.pl cures that problem.
20:32 kados thd: 23722 MARC record done in 3609.88483405113 seconds
20:32 kados still failing
20:33 kados I'm guessing Owen didn'g commit the opac subject search changes
20:33 owen  Or else I overwrote them by mistake
20:39 kados hmmm, templates are identical
20:39 kados very strange
20:41 thd   kados: owen committed the changes within a couple of days after I had told him about them.
20:41 kados ahha
20:41 kados they're not identical
20:41 thd   kados: what is not identical?
20:42 kados hmmm
20:42 kados didn't make a difference
20:42 kados operator=starts vs operator=contains
20:44 thd   kados: that what I had guessed but that seemed OK when I had actually checked earlier
20:46 kados didn't seem to make a difference
20:47 kados in my test case
20:47 kados did we strip out the - or something ?
20:47 thd   kados: a simple subject search for a single term appearing in 650 $a such as history or whatever fails from the advanced search.
20:48 thd   kados: As, I said, I have not installed 2.2.5 yet myself.
20:49 thd   kados: I had been too busy looking for a lawyer to keep myself out of gaol.
20:50 thd   kados: Do you know of any installation using 2.2.5 successfully.  Could there be a big bug?
20:53 thd   kados: If you query the database directly do you find valid data in bibliosubjects.subjects?
20:55 kados I can check
20:59 thd   kados: I have to help a friend move shortly.  I want to help you solve this quickly but I may not have more time when we are both awake today.
20:59 kados I'm working on this until I fix it :-)
20:59 kados it's a major problem :-)
21:00 kados first I'm copying the known-good koha filesystem over to the test box
21:00 kados to test if it's in the files or the db
21:00 thd   kados: other suggestions of mine will include purging the records and reloading with bulkmarcimport.pl.
21:02 kados yep, though I use mysqldump to load this data
21:02 kados so I'd be surprised if that wias it
21:02 kados unless you changed something in the marc framework?
21:02 kados before I ran rebuildnonmarc?
21:02 thd   kados: If that were to fail and lacking any other obvious reason, I would suggest droping the whole database and starting over in case there is some corruption.
21:05 thd   kados: I only changed the search also for 650 $a, management in tab 3 for all 650 subfields, and repeatability for the repeatable 650 subfields.
21:07 thd   kados: I could try a more conservative search also setting using only fields from 650.
21:08 thd   kados: my comprehensive search also value for 650 $a had been newly created today and was untested.
21:14 kados must be that then
21:14 kados :-)
21:15 thd   kados: I just noticed that biblioitems.dewey had been set to 650b
21:16 kados er?
21:16 kados did you do that?
21:24 kados thd: I'm going to change things back to the way they were
21:27 thd   kados: I fixed them already
21:27 kados um ... really?
21:27 kados when?
21:27 thd   yes
21:27 thd   just now
21:27 kados not really :-)
21:28 thd   sorry, I had been on telephone.
21:29 kados now, they should be fixed
21:31 thd   kados: a least basic searches are working now
21:33 kados I see the problem
21:33 kados I'll fix it so you can see it too
21:33 kados don't do anything on the marc framework till I'm done ok?
21:34 thd   kados: is it the template issue for contains instead of stars?
21:34 thd   kados: I actually need to go and help move my friend very shortly.
21:35 kados ok
21:36 thd   kados: what did you find was the problem?
21:46 kados thd: template problem
21:46 kados I committed the change and it's working now
05:59 |hdl| hello world
10:54 kados morning #koha