IRC log for #koha, 2006-03-08

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

All times shown according to UTC.

Time	Nick	Message
11:00	pierrick	If you knew the speed of Gna! infrastructure :-)...
11:01	paul	for CVS it's really faster than sourceforge !
11:02	pierrick	How could it be slower?
11:02		;-)
11:52		where do I set the default intranet theme ? (I brutally modified C4::Output::themeLanguage, but it's obviously not the correct method)
11:52	fsouchon	hello
11:53	paul	pierrick : koha >> parameters >> sytempreferences is your friend
11:53	fsouchon	Is it possible to run Koha via a IIS server ?
11:53	paul	fsouchon bonjour
11:53	fsouchon	bonjour
11:53	paul	plein de frenchies aujourdh'ui...
11:53	fsouchon	lol
11:53	paul	(et même que ca pour tout dire.)
11:53	fsouchon	on a uen demande
11:53		mais en IIS :/
11:53	paul	sauf si kados est levé !
11:53		c'est qui "on" ?
11:54	fsouchon	ma boite
11:54		une SSLL
11:54		donc
11:54		est-ce possible?
11:54	paul	un nom, un nom...
11:54	fsouchon	alternancesoft
11:54		c nouveau
11:54	paul	inconnu au bataillon.
11:54		bon, pour répondre à la question :
11:54	fsouchon	c né d'une spécialisation de ntsys en opensource
11:54		c petit :)
11:54	paul	c'est possible, mais ca marche fort mal.
11:54	fsouchon	arf
11:54	paul	et pas question de mettre un client z3950 dessus
11:54	fsouchon	quel est le problème?
11:55	paul	là, ca marche po.
11:55	fsouchon	non c pour du client web léger
11:55		Moz /IE
11:55	paul	perl Net::z3950 qui est tout cassé en IIS
11:55	fsouchon	oki
11:55	paul	je précise :
11:55	fsouchon	mais sur client web http
11:55		ça passe?
11:55	paul	Le client z3950 de Koha permet d'interroger les serveurs z3950 tout autour de la planète.
11:56		pour ca, un process perl sur le serveur Koha va interroger les serveurs z3950
11:56		le susnommé process utilise Net::z3950 qui est tout cassé sous IIS.
11:56		(et compliqué à installer sous windows)
11:56		mais windows/apache, ca fonctionne, moyennant de l'huile de coude.
11:57		enfin, l'interface bibliothécaire est garantie sous mozilla.
11:57		si elle marche sous IE, c'est "par hasard"
11:57		(et je ne crois pas qu'il y ait un seul développeur de Koha sous windows... donc ca va pas changer... l'OPAC est testé sous IE, mais en toute fin de développement)
12:03	kados	paul: are you still around?
12:03	paul	hello joshua
12:03	pierrick	hello Joshua
12:03	paul	yep, of course (4PM in France)
12:04	kados	great
12:04		hi pierrick
12:04		paul: I've been hacking on authorities
12:04		paul: specifically the display
12:04		paul: http://opac.liblime.com/
12:04		paul: there is now a authorities heading search on the opac
12:05		paul: do a seach for 'NAME' Lewis
12:05	paul	men, you rock !
12:05	kados	paul: the display for that is hard-coded
12:05		paul: it does not use the ISBD style as ISBD is flawed in this instance
12:05	paul	you mean the "authorized headings column ?
12:05	kados	paul: yep
12:06	pierrick	on my working copy of HEAD, many pages produce "Internal Server Error". Only very few log in koha-erro_log, just "Premature end of script headers: systempreferences.pl, referer: http://plegall:8080/cgi-bin/ko[…]min/admin-home.pl". How do I turn on log verbosity?
12:06	kados	paul: (what used to be called 'summary')
12:06	fsouchon	merci pa
12:06	paul	you could use the summary column.
12:06	fsouchon	paul,
12:06	paul	you're welcome fsouchon
12:06	fsouchon	je susi fondamentalement contre M$ et sa saube de IIS
12:06	paul	and if you need more help : 04 91 31 45 19
12:06	fsouchon	mais là le choix ne nous appartient pas
12:06	kados	paul: the problem with ISBD is it can't handle repeatability correctly
12:06	fsouchon	merci paul
12:06	paul	kados : ???
12:07	kados	paul: I will try to explain
12:07	paul	(c'est mon business Koha, donc si tu veux qu'on fasse une affaire ensemble ;-) )
12:07	fsouchon	tu bosses où?
12:07	kados	paul: look at the auth heading for 'Twain'
12:08		paul: NAME heading
12:08		paul: it has the following entries in MARC:
12:08	paul	(je suis le paul poulain co-auteur et Release Manager de Koha ;-) )
12:08	kados	paul: 100 $a $c 400 $a $d 500 $a $d
12:09		paul: in the ISBD method you can say you want those to display as:
12:09		paul: [100a] [100c] [400a] [400d] [500a] [500d]
12:09		paul: it will work fine for this example because there is no repeatability
12:10		paul: however, if you have a record like the 'Lewis' record:
12:10		paul: 100 $a $d $q 400 $a $d 400 $a
12:11		paul: the ISBD entry would look like this:
12:11		paul: [100a] [100d] [100q] [400a] [400d]
12:11		paul: HOWEVER, the display will come out like this:
12:12		paul: 100 $a 100 $d 100 $q 400 $a 400 $a 400 $d
12:12		paul: [400a] will repeat right next to itself
12:12		paul: it does not understand that this is a hierarchy
12:12	paul	kados : right.
12:13		pierrick : read private channel.
12:13	kados	paul: the hard-coded example on opac.liblime.com will work with all MARC21 authorities records
12:13		paul: I could commit it as an alternative without touching 'summary'
12:13		paul: with a new syspref for 'authoritiesheadings'
12:14	paul	* how will it work if there is no "see also", or "see" ?
12:14		* what append if there is another field/subfield that is not in your "template" ?
12:14	kados	paul: look at the 'Twain' example
12:15	paul	otherwise, I think it's really a good hack if it work for all MARC21. It should work for UNIMARC as well if I don't mind, as, i've been told that authorities are closed between UNIMARC And MARC21
12:15	kados	paul: the code follows MARC21 practice for determining the authorized heading (ie 1XX)
12:16		paul: and for the see (4XX) and seealso (5XX)
12:16		paul: it will append any subfields that exist correctly
12:16	paul	so, it's a great feature !
12:16	kados	ok ... I'll commit it along with some documentation
12:16		sent to koha-devel
12:17		there are still some problems with the authorities editor
12:17	paul	(that just require a link to be added to fulfill UN requirements ;-) )
12:17	kados	paul: :-)
12:18		paul: can you modify the auth editor to work like the bib editor (with tabs, etc.)
12:18		paul: may require some modif to the db as well ?
12:20	paul	modify auth editor to look like bib editor : of course. (except for tabs, that are useless I thought, an authority record is never so long that it needs tabs)
12:20		and it don't need any DB modifs unless i'm missing something
12:20	kados	paul: for some reason, the default auth records are quite long in fact
12:21	hdl	kados : yes
12:21		because they display every field.
12:21		and not only the fields that you needfor your authority.
12:22	kados	right
12:22	hdl	But paul, with MSC, I have some exemples where you have 5 or 6 subfields on the same page.
12:22	kados	in fact, MARC21 authority records can be quite long
12:22	hdl	Maybe would it be useful to have 3 or 4 tabs :
12:22	kados	for instance, authorities from OCLC
12:23	hdl	Main form
12:23		Associate from
12:23		parrallel form
12:23		and rejected form.
12:23	paul	ok, ok, I resign ! you won !
12:23		(gg)
12:23	kados	hmmm
12:23		I also have a question about intended uses of the 'frameworks' in authorities
12:23		are they to specify different authorities headings?
12:24		or different authorities types?
12:24		(hint, a type of authority can hold several different headings)
12:25		this reference is quite useful:
12:25		http://www.loc.gov/marc/uma/
12:26		specifically this section:
12:26		http://www.loc.gov/marc/uma/pt1-7.html#pt4
12:26		"All headings are one of the following types: names, name/title combinations, uniform titles, or subjects."
12:26		[snip]...
12:26		The variable data fields are grouped into blocks according to the first character of the tag, which identifies the function of the data within the field. The type of information in the field is identified by the remainder of the tag. The blocks are:
12:26		0XX  Standard numbers, classification numbers, codes
12:26		1XX Headings (authoritative and reference)
12:26		2XX Complex see references
12:26		3XX Complex see also references
12:26		4XX See from tracings
12:26		etc.
12:27		there are also content designation in the remainder of the tag
12:28		ie X00 == personal names X10 == corporate names, etc.
12:28		and finally, there are designations in the 008 fixed field
12:28		specifically position 9:
12:28		a
12:28
12:28		Established heading
12:28		b
12:28
12:28		Untraced reference
12:28		c
12:28
12:28		Traced reference
12:29		d
12:29
12:29		Subdivision
12:29		e
12:29
12:29		Node label
12:29		f
12:29
12:29		Established heading and subdivision
12:29		g
12:29
12:29		Reference and subdivision
12:29		paul: so which of these levels of the hierarchy are the 'Thesaurus Frameworks' designed to handle?
12:29	paul	wow... /me never asked myself this question ;-)
12:30	kados	I think we are very close to supporting full MARC authorities -- closer than I thought
12:30	paul	I imagined thesaurus frameworks just to be able to have whet is in the 2nd digit : personal//corporate...
12:31	kados	as currently designed, the frameworks are used as 'search points' so they most closely resemble the 'headings' (ie, names, name/title combinations, uniform titles, or subjects. )
12:32		at least to the patron looking for items
12:32		however, for editing purposes
12:32		I'm not sure what function they serve
12:32		(I'm quite excited about our level of conformance to MARC in 2.2.6)
12:34		(subfield reordering and repeatability within a tag for both authorities editing and biblios editing will be a giant step as well)
12:35		finally, the plugins do not seem to work in the auth editor ...
12:35		but I have not investigated why
12:38		(also, authorities editor seems to truncate the 008 field)
12:39		hdl: did you see my bug report on serials creation?
12:39	hdl	Seems I read it.
12:39		Can you remid me ?
12:39		remind
12:40	kados	serials creation don't work anymore in rel_2_2 :-)
12:41		http://bugs.koha.org/cgi-bin/b[…]w_bug.cgi?id=1048
12:41		hdl: error codes in the report
12:42	pierrick	(I made my first commit, yeah :-)
12:42	kados	pierrick: woohoo! :-)
12:42		pierrick: congrats
12:43	pierrick	what a great modification I made !!!
12:43	kados	pierrick: what was it? :-)
12:43	pierrick	something huge
12:43	kados	heh
12:43	pierrick	typo fixed: a ";" was missing at the end of Zebra connection retrieving in C4::Search->get_record
12:44	kados	ahh ... quite huge in fact :-)
12:45	paul	champagne for pierrick 1st commit !
12:45	kados	paul: (since you will not be attending the meeting tonight ... do you have any results from your queries into where the non-profit should be located: france, us, or elsewhere?)
12:45	paul	http://www.box-evenement.com/s[…]lle_champagne.jpg
12:45	kados	hehe
12:46	paul	my conclusion : it's a very very hard question.
12:46	kados	wow very large champagne :-)
12:46	paul	for a large commit
12:46	pierrick	I can't drink so much ! (I don't drink alcohol at all in reality)
12:47	paul	that must be handled carefully, with an international lawyer. I asked a french from mozilla fundation (Tristan Nitot), but didn't get any answer yet
12:47	pierrick	Tristan is a very busy person
12:47	paul	I suspect it would be easier to do something in US and "localize" it in France, than the opposite.
12:48		(I know, but I tried ;-) )
12:49	hdl	kados : seems it work for me.
12:49		Did you change something ?
12:50		Or have I forgotten a commit ?
12:51	kados	hdl: maybe forgot a commit?
12:51		paul: OK ... I will mention this at the meeting
12:52		now ... unfortunately I must go :/
12:52		as I have several appointments today
12:52		paul: hdl: will you be around tomorrow?
12:52	hdl	yes.
12:52	kados	I would like to discuss authorites and serials in further detail
12:53		bye all
12:53	pierrick	see you tonight
12:53	kados	pierrick++
12:53	paul	kados : yep, i'll be here tomorrow
12:57	hdl	kados : I donot have the same problem. Is it with npl templates ?
12:57		And it seems thet all my files were committed.
12:59		owen : can you answer ?
13:06	pierrick	where are stored members in DB ? I see table "users" was dropped by updatedatabase
13:06	paul	borrowers
13:07	pierrick	OK
13:08	paul	hello owen. kados leaves 10mn ago.
13:09		[16:52:09] <kados> now ... unfortunately I must go :/
13:09		[16:52:19] <kados> as I have several appointments today
13:09	owen	He's probably out conquering North America in the name of Liblime :)
13:10	paul	lol
13:10		(and he is doing a pretty good job it seems !)
15:21	thd	paul: are you there?
15:28		I will find you tomorrow paul_away
16:46	kados	T-MINUS 15 MINUTES TO KOHA MEETING
16:47		chris: you here?
16:47		chris: I don't have oper privs
16:47		chris: so can't change the topic
16:47	chris	only pat can
16:47	russ	pate-lurk is the only one who does
16:47	mason	:)
16:47	kados	what!
16:48		isn't this a katipo box?
16:48		i get it now ... he was first on
16:48	thd	kados: I have his address if you still need it
16:48	kados	chris: is this running on a katipo box though?
16:49		chris: ircd?
16:49	chris	hybrid-ircd
16:49	kados	if so, you should be able to manually add operators
16:49	chris	no
16:49		doestn work that way
16:49	kados	and reload the config without killing everyone
16:49		ahh
16:50	chris	you can add irc operators .. but not channel operators
16:50		irc operators can ban, gline etc .. but they dont have any special rights in a channel
16:50	kados	in Unreal ircd you can add channel opers too
16:51		I'm pretty sure anyway ...
16:51	chris	ill have to go read hybrid-ircd docs more
16:51	kados	well ... not a big deal really
16:51	chris	but i couldnt see any way to do it
16:51	kados	MEETING AGENDA: http://tinyurl.com/ztkar
16:52		if anyone has stuff to add, please do
16:52		in the next ten minutes or so
16:56	thd	kados: Have you fixed original cataloguing subfield repeatability in the biblo record editor some where?
16:56	kados	thd: the underlying marc structure and routines support it now
16:57		thd: but the actual editor hasn't been modified to support it
16:57	hdl	hi
16:57	chris	hi hdl
16:57	thd	kados: So the template work to support the underlying remains?
16:57	kados	hi hdl
16:57	chris	just quickly before our meeting hdl
16:57	kados	thd: we just need some javascript to insert a new subfield below the existing one
16:58		thd: and we need some javascript for reordering of subfields within a tag too
16:58	chris	for your zebra problem .. how did you create your kohazebra db ?
16:58	kados	actually chris ...
16:58		I think I know the problem
16:58		I suspect it's shdow registers
16:58	chris	joshua and I were having this problem
16:58		ahh
16:58	kados	you have to run a 'commit' after you create a new db
16:58	chris	ahh
16:59		ill let you answer hdl on the list then
16:59	thd	kados: I was forgetting that this was almost a JavaScritp template :)
16:59	kados	I ran into the same prob with setting up owen's test site
16:59		and that seemed to be the only solution
16:59	chris	cool, its good that we know what it is now
16:59	kados	yea ... quite a relief when i figured it out
16:59		it was driving me batty :-)
16:59		T-MINUS 2 MINUTES TO KOHA MEETING
17:00		AGENDA: http://tinyurl.com/ztkar
17:00	hdl	ow do you run a commit ?
17:00	kados	hdl: zebraidx commit
17:00	hdl	is that all ?
17:00	kados	hdl: so the steps are:
17:00		zebraidx create kohadb
17:01		zebraidx commit
17:01		zebrasrv localhost:2100
17:01		yep, that should do it
17:01		if not let me know
17:01	hdl	ok.
17:01	kados	ok ... it's 20:00 GMT
17:01		lets get our mtg started ... so first ... roll call
17:01		who's here?
17:02	thd	still her
17:02		s/her/here/
17:02	pierrick	I'm her
17:02	kados	paul's not going to make it
17:03		so first things first
17:03		welcome to pierick!
17:03	pierrick	thank you Joshua
17:03	kados	pierrick even :-)
17:03		pierrick did his first commit today
17:03		he's a fast learner apparantly :-)
17:03	pierrick	(bwouhaha)
17:03	chris	:-)
17:03	russ	most excellent, welcome aboard pierrick
17:04	kados	here's our mtg agenda:
17:04		http://tinyurl.com/ztkar
17:04		anything to add before we start taking about perl-zoom?
17:04	thd	Koha is intuitive only cvs is not
17:04	pierrick	(I'm CVS expert, but not Koha yet)
17:05	kados	I'll take that as a no
17:05		so perl-zoom integration
17:05		chris: want to summarize where we're at with the plugin?
17:06	chris	ok
17:06		with a few database changes, and copying some scripts and some modules from HEAD to a 2.2.5 koha install, you can get zebra playing nice with 2.2
17:06	kados	(I wrote a mail to koha-devel about how to get the plugin working but it hasn't arrived yet)
17:07	chris	there is more work to be done, and some more testing ... but its all work useful for 3.0 also
17:08		there is no added functionality, BUT searches are now done on zebra
17:09		so whats left to be done?
17:09	kados	hmmm
17:09	chris	well, currently there is no easy way to upgrade from a 2.2 install
17:09	kados	right ... so maybe a shell script or two to help with that
17:09	chris	ie, it works fine if you install clean, make the db changes and then load your data in
17:10	kados	i forwarded the description email to meadville so they should start testing soon
17:10	pierrick	personnaly, HEAD installation was not so easy, mainly due to environnement issues
17:10	chris	theres still bunches of bugs that need to be ironed out
17:10	thd	kados: What is meadville?
17:11	pierrick	I mean libyaz/yaz/ZOOM/zebra
17:11	kados	thd: one of my clients who sponsored perl-zoom development
17:11		pierrick: yep, installation's never been easy in Koha
17:11		so chris, what search points are working now?
17:12		author, title, subject I think
17:12		any others?
17:12	chris	isbn
17:12		itemtype will when we add it to our collection.abs and pqf.properties
17:12	kados	so that's a todo
17:12	chris	basically we just have to add more data to collection.abs and pqf.properties
17:13		then another elsif in SearchMarc.pm
17:13		its very easy
17:13	kados	sweet
17:13		series title, format, location, barcode, call #, publisher, published between dates
17:13	chris	yeah
17:13	kados	adding biblios works eh?
17:14	chris	yep
17:14	kados	and items as well ...
17:14	chris	yep
17:14	kados	how about deletions?
17:14		and edits?
17:14	chris	dunno havent tested
17:14	kados	k ... two more todos then
17:14		what else?
17:14		bulkmarcimport and rebuild_zebra been tested with the plugin?
17:14	chris	no
17:14		if rebuild_zebra works, then that will allow us to upgrade
17:15		which would be good
17:15		cant see why they wouldnt
17:15	kados	anyone think of anything else?
17:15	pierrick	I don't
17:15	kados	order by
17:15	chris	the only thing i can think of
17:16	kados	also, stopwords
17:16	thd	upgrading is a problem for holdings
17:16	kados	for searching
17:16	chris	stop words should just keep working
17:16		they should be stripped before the search is called
17:16	kados	zebra's got a much better system for handling stop words that we should definitely use
17:16	chris	3.0 will
17:17	kados	ok ... sounds reasonable
17:17	chris	returns .. if we search over branch by zebra
17:17		returns has to update zebra
17:17	kados	it does?
17:17	chris	yes
17:17		if we are searching the branch location
17:18	kados	to change holdingbranch?
17:18	chris	in zebra, then zebra has to know what the currentbranch is
17:18	thd	code needs to change to preserve holdings ID from rel_2_2 to 3.0
17:18	chris	this isnt upgrading from 2.2 to 3.0 .. its slapping zebra into 2.2 .. but yes that will be needed in the future thd
17:19	kados	right
17:19		so ... who's gonna work on that task list?
17:19		for perl-zoom?
17:19	chris	me
17:19	kados	woot :-)
17:19		ok then we can move on eh?
17:20		paul's not here so I think we'll have to skip authorities discussion in detail
17:20	thd	kados: I understand authorities fairly deeply
17:20	kados	thd: great!
17:20		first let me show off what thd and I did
17:21	thd	kados: just that most of what I saw was in a compressed time a few months ago
17:21	kados	http://opac.liblime.com
17:21		there is now an authorities search in the OPAC
17:21		a NAME search on 'Lewis' or Twain
17:21	thd	kados: If you read a little French then you can read authorities :)
17:21	kados	will give you a good feel for what the modifications we made will allow
17:22		mainly they were display issues
17:22		though we also hacked on the authorites editor as well as the framework
17:22		I plan to talk at detail tomorrow with paul about some other ideas we have
17:23	thd	s/read authorities/read authorities in Koha/
17:23	kados	I think we can move on ... unless anyone has something to add to the auth discussion
17:24	hdl	I wonder
17:24	thd	UNIMARC and MARC 21 authorities are very similar
17:24	kados	hdl: I'm listening
17:25	hdl	If ppl wil be able to have different summary for 1 auth type or what use we could do for different summary (headings)
17:25	kados	in MARC21 all authorized headings are in the 1XX field
17:25	thd	There is much less divergence than with the bibliographic standards except for subject subdivisions which are much more troublesome in MARC 21
17:26	kados	all see and see also are in 4XX and 5XX respectively
17:26		so it's quite easy in fact to categorize the hierarchy correctly for display
17:26		there are several other types I'd like to add to the summary
17:26		for one, the 'from source'
17:27	thd	hdl: what you may see has not been corrected for all template parts
17:27	kados	the problem with the old ISBD style of building the summary was that it could not represent the true hierarchy of the record
17:28	thd	hdl: the OPAC advanced search pop-up lacks template corrections
17:28	kados	$400 $a $b $400 $a $b was represented as $400 $a $a $b $b
17:28		thd: I think it's working now for all but author searches
17:29		thd: though as we've discussed, the behavior after selection is not as expected
17:29	thd	kados: you are a little mistaken about the hierarchy issue in Koha ISBD
17:29	kados	thd: please explain
17:30		thd: actually ... lets postpone that until after the mtg
17:30	thd	kados: you should check my comprehensive MARC 21 ISBD system preference to see how it is done
17:30	kados	thd: lets move on to discuss the Koha non-profit
17:30	thd	kados: yes please do as I have no idea what that is even
17:31	pierrick	thank you thd for asking a brief explanation :-)
17:31	kados	I spoke to paul about the non-profit briefly this morning
17:32		I'll paste in the conversation
17:32		07:43 < kados> paul: (since you will not be attending the meeting tonight ... do you have any results from your queries into where the non-profit should be located: france, us, or elsewhere?)
17:32		07:44 < paul> my conclusion : it's a very very hard question.
17:32		07:45 < paul> that must be handled carefully, with an international lawyer. I asked a french from mozilla fundation (Tristan Nitot), but didn't get any answer yet
17:32	russ	i concur :-)
17:32	kados	07:46 < paul> I suspect it would be easier to do something in US and "localize" it in France, than the opposite.
17:32		07:50 < kados> paul: OK ... I will mention this at the meeting
17:33		so ... now I have :-)
17:33	russ	:-)
17:33	thd	kados: what purpose does the the non-profit have?
17:33	kados	thd: several purposes that we can discuss after the meeting
17:33	thd	kados: please explain the basic conception behind the non-profit.
17:33	kados	in detail
17:34		1. we can go after grants for development works
17:34		work even
17:34		2. the non-profit will own the Koha brand as well as copyrights on the code
17:34		(think 'Apache foundation'
17:34		)
17:34		3. it will provide a measure of protection for companies selling services on Koha
17:35		(against suits, etc.)
17:35	chris	2. im not sure about
17:35	russ	to me the purpose of the non profit discussion at this stage is to ensure that the project carries on
17:35	chris	i think its safer/good for developers own copyrights
17:35	thd	kados: Then it should be based or have qualified offices in every country where grant money is available but no country with software idea patents.
17:35		:)
17:36	russ	i think the details need to be worked on and discussed
17:36	hdl	India ?
17:36		:D
17:36	thd	chris: assignment is an absolute requirement to upgrade the license
17:36	chris	i have no plans to upgrade the license
17:36	kados	chris: I disagree, though I'm not a lawyer ... I think the fsf always recommends turning over the copyright of a GPL work to them
17:37	thd	s/assignment/copyright assignment/
17:37	kados	chris: because any legal disputs over patents then have to be filed against the corp and not the individual
17:37	chris	yes, but in order to change the license you must get agreement of all the copyright holders
17:37	thd	chris: yes that is the virtue of assignment
17:37	chris	so having more copyright holders
17:37		is safer
17:38		than having one copyright holder
17:38	kados	safer for who?
17:38	chris	the project
17:38		koha
17:38	kados	ahh :-)
17:38		we had different assumptions then :-)
17:38	thd	chris: now the community is small and all assignment is to Katipo so only one entity can upgrade to GPL V3
17:38	chris	no its not
17:38	hdl	chris : yes, but developers can be at risk.
17:38	chris	there is some code copyright to others
17:39	thd	chris: which code is that?
17:39	chris	plus i dont even want to start on the whole gpl v3
17:39	russ	we wouldn't lay claim to writing all the code
17:39	kados	thd: Amazon.pm is copyright LibLime IIRC
17:39	chris	im sure there is some skemotah in there too
17:39	thd	russ: claim is not about righting it is about administering the license
17:40		s/righting/writing/
17:40	chris	what i see the foundation doing
17:40		is fulfilling the kaitiaki role
17:40		which is currently not being done
17:41	thd	what is the meaning of kaitaiki?
17:41	chris	guardian
17:42	thd	chris what has the kaitiaki done in the past?
17:43	chris	they are the central point of contact .. overall guidance, press, etc
17:43		it has always been a volunteer role
17:43		and it takes a huge amount of work
17:43	kados	yep
17:43	russ	it also takes a special kind of person
17:44	kados	pate did an excellent job while we were transitioning from 1.2 to 2.0
17:44		there's a ton of press from that time period
17:44	thd	I thought rach was kaitiaki currently
17:44	chris	a not for profit allows for funding to be sought to remunerate people
17:44		technically she is, but with a hmm 6 week old baby
17:45		and a company to look after .. its a pretty unrealistic thought that she can devote the time necessary to the role
17:45	kados	right
17:45	chris	the way i see it
17:45		its a process
17:45	thd	Therefore rach has a small kaitiaki assistant :)
17:46	chris	1 get some kinda not for profit organisation
17:46		2 slowly evolve/expand organisations role
17:46		i dont think we want/need to do everything at once
17:46	kados	good point
17:47		so the first step is to get an org going
17:47	russ	well i have been looking into that
17:47	kados	cool ... what'd you find out?
17:47	russ	and the thing is that you need to have a clear idea of what you want to do
17:47		i have looked a little bit at us, but mainly nz not for profit setups
17:47		but with both you need a constitution
17:48		or a definition of purpose
17:48		i think we need to work on that first
17:48	kados	yep
17:48	russ	and then decide where
17:48	kados	in the US it's usually just a two-three sentence mission statement
17:48		something like:
17:48		"to promote the Koha integrated Library system"
17:49	russ	i found an example of another software project
17:49	kados	doesn't need to be _too_ specific, though I think it would be a good exercise to put down in writing our goals
17:49	russ	who have published their cert of incporp
17:49		kados: exactley
17:49	kados	cool ... got urls?
17:49	russ	lets find out if they are even remotley aligned
17:49		before we start down the process of creating an entity
17:49	chris	i think thats best discussed on the devel list
17:49	russ	know you were going to ask that
17:50		chris : me
17:50		oops
17:50		chris: i agree
17:50	thd	There are different costs and liability issues for various jurisdictions
17:50	chris	i think what has to be decided first, and then where after
17:51		so lets start a discussion on what, on the list and see if some nice goals coalesce
17:51	kados	sounds good
17:52		anything else to discuss on that topic?
17:52	russ	no i think that covers it for me
17:52	kados	cool
17:52		last item on the agenda is the Koha Con
17:52	russ	i have some urls on a scrap of paper that i will dig out when i see the post on the list
17:52	kados	russ: can you summarize where we're at with that?
17:53	russ	Koha Con - um not really :-)
17:53		i have some dates from Paul
17:53	kados	chris: ?
17:53	russ	i had hoped he would be here today
17:53	thd	what is Koha Con?
17:53	kados	thd: we're organizing a Koha Conference to be held in France in May
17:53	pierrick	Koha Con(ference)
17:54	kados	Paris in fact
17:54	russ	i'll email paul again
17:54	hdl	and Marseille
17:54	kados	right
17:54	russ	cos we really need to get something out to all the lists and the site soon
17:54	chris	we have some dates
17:54	kados	so maybe hdl should summarize what our plans are thusfar :-)
17:54		hdl: can you?
17:54	hdl	yes
17:54	pierrick	1 day of general presentation
17:54		(sorry HDL, go on)
17:55	hdl	1 day gl presentation
17:55	russ	sorry i have another meeting, i must go, catch you all later
17:55	hdl	2 days with users
17:55		And 4 days for developers in Marseille
17:55	chris	so
17:56		may 2 general koha presentation, everyone interested come along
17:56		may 3 - koha users - for existing koha users
17:56		may 4-7 free time (explore paris)
17:56	hdl	They could then gather and share experiences.
17:56	chris	may 8-12 developer week in marseille
17:56	pierrick	INEO plans to call some french libraries for the first day presentation
17:57	chris	may 2 and 3 in paris
17:57		cool pierrick
17:58		i plan to pay pierrick or hdl or paul with chocolate to sit next to me and translate :-)
17:59	hdl	:)
17:59	thd	spoken French, I would have no chance
18:00	hdl	For users' conf, will be needed.
18:00		But not for dev conf.
18:00		Since we are mostly english speaking persons.
18:00	pierrick	Joshua will speak english on day 2
18:01		and someone will translate to french
18:01	kados	heh
18:01	thd	Will there be a video archive?
18:02	pierrick	video ? we didn't speak about it? It's a good idea. I'll ask my boss.
18:02	kados	so basically about 10 days
18:02		well ... 11 :-)
18:02		plus some time off inbetween
18:03		do we need to talk about anything else related to the con?
18:03		obvoiusly we need to start promoting it as soon as possible
18:03	chris	:)
18:03	thd	pierrick: A video archive should be posted somewhere when it is all done unless it is too poorly organised.  It ought to be an enduring archive of good publicity.
18:04	kados	thd: good idea
18:04	hdl	maybe on koha-fr.org
18:04		or koha.org ?
18:05	chris	yep
18:05		or both :)
18:05	hdl	:D
18:05	chris	we are working on getting a US based mirror for koha.org
18:05	kados	sweet
18:05	chris	should be done soonish
18:05	thd	However, someone needs to have the right well tested equipment.  Poor quality sound is poor publicity.
18:06	chris	we should get people to transcribe also
18:06	hdl	If EMN is involved, thay can have a good auditorium room.
18:06	pierrick	MN ?
18:06	chris	having the transcriptions of the talks are valuable
18:06	pierrick	EMN?
18:06	thd	Someone needs to know how to make the equipment work.  Good equipment run by an unprofessional user might as well be poor equipment.
18:06	hdl	Ecole des Mines de Nantes (User's conf)
18:07		I shall remind paul.
18:07	pierrick	Nantes ??? Paris, you mean
18:07	hdl	Sorry EMP.
18:08		you're right pierrick....
18:08		Stupid of me.
18:09	pierrick	making great quality video seems to be quite complicated, not so sure it will be possible
18:09	kados	anything else to discuss?
18:09	chris	not from me
18:10	hdl	And transcription is a real job.
18:10	chris	yes, its hard work
18:10	kados	k ... meeting's closed then ... thanks everyone!
18:10	hdl	Nope.
18:10		Thx
18:10	thd	pierrick: If you have good sound and tolerable video it can be acceptable.  Poor sound is bad.
18:10	kados	I'll post a summary later today
18:10	chris	thanks kados
18:12	pierrick	thanks Joshua
18:12	thd	kados: There is much to discuss.
18:14		kados: I worked on the MARC 21 default framework all day yesterday and will be done after another day except for corrections and a proper import script.
18:15	kados	thd: excellent
18:15		thd: give me just a sec and I"ll be with you
18:15	thd	s/corrections/careful proofreading/
18:16		s/corrections/careful verification/
18:17		In summary still a few more days required.
18:32		kados: I have been so careful that I have even found a typo in an official MARC 21 document.
18:32	kados	thd: nice :-)
18:32		thd: I'm back now
18:32		thd: if you need to discuss anything with me
18:33		thd: good news about the framework
18:33		thd: I'm guessing paul can help us to distribute it in the best way
18:33		thd: I'll add that to my list of questions for him tomorrow
18:34	hdl	kados : i will have to work hard on authority soon.
18:34	kados	thd: can you expand on what you said earlier about ISBD?
18:34		hdl: I committed my changes
18:35		hdl: so you can see what I did
18:35		hdl: the best resources are the following (for marc21)
18:35		http://www.itsmarc.com/crs/Auth0001.htm
18:35		http://www.loc.gov/marc/uma/index.html
18:35	hdl	can you give me details for you bug #10008 paul and I couldnot see it happen.
18:35	kados	right ...
18:36		you can try it on liblime demo
18:36		adding any serial fails
18:36		and the errors are in the log as i posted to the bug report
18:37		hdl: I tested with default templates even
18:37		hdl: (btw: are there any changes to make to a template to support serials itemization?)
18:38	hdl	SURE!!!!
18:38	kados	hdl: http://koha.liblime.com/cgi-bi[…]bscription-add.pl
18:38	hdl	there are.
18:38	kados	hdl: try with default templates
18:38		supplier 1 is EBSCO
18:39		hdl: use 9903 as bib record (magazine)
18:40	hdl	can you give the name of the serials ?
18:40	kados	any search for 'magazine'
18:40		try 'american' with itemtype magazine
18:41	thd	kados: The existing Koha ISBD model can represent repeated fields acceptably but has a problem with repeated subfields within the field where the repeated subfields are not consecutive.
18:41	hdl	internal server error?
18:42	kados	hdl: did you get an internal server error?
18:42		hdl: I just attempted and it seems to have worked
18:42	hdl	yes.
18:42	kados	hdl: when did you get it?
18:42	hdl	I tried Iowa journal.
18:42	kados	hdl: [Mon Mar 06 13:39:36 2006] [error] [client 62.212.120.129] HTML::Template->output() : fatal error in loop output : HTML::Template::param() : attempt to set parameter 'itemlocationloop' with a scalar - parameter is not a TMPL_VAR! at /usr/lib/perl5/site_perl/5.8.5/HTML/Template.pm line 2963, referer: http://koha.liblime.com/cgi-bin/koha/bull-home.pl
18:42		[Mon Mar 06 13:39:36 2006] [error] [client 62.212.120.129]  at /var/www/liblime.com/koha/intrane​t/cgi-bin/bull/statecollection.pl line 160, referer: http://koha.liblime.com/cgi-bin/koha/bull-home.pl
18:42	thd	kados: Orderliness is not read from the record in the existing Koha ISBD model it is instructed in the Koha ISBD configuration supplied.
18:42	kados	[Mon Mar 06 13:39:36 2006] [error] [client 62.212.120.129] Premature end of script headers: statecollection.pl, referer: http://koha.liblime.com/cgi-bin/koha/bull-home.pl
18:43		thd: right
18:43		thd: I don't think it can represent repeated tags acceptably either
18:43	hdl	Strange.
18:44		Are your templates up to date ?
18:44	thd	kados: so you can make 400 $a $d 400 $a $d work
18:44	kados	thd: you'll have to show me how
18:44		hdl: I'll check
18:45		hdl: try now
18:45		hdl: (I updated all templates in default/en)
18:45		hdl: (but no changes were listed)
18:45	thd	kados: However, 260 $a $b  $a $b comes out as 260 $a $a $b $b
18:46	kados	thd: do you have an example of the previous one working?
18:46		hdl: could french templates be different than english templates?
18:47	thd	kados: yes just look at any record with multiple repeated subject fields using the ISBD configuration that I supplied to you
18:47	hdl	I work on english templates.
18:47	kados	hdl: strange then
18:48	hdl	and paul has the same behavior as I.
18:48		it works.
18:49		would you send me your statecollection.pl and statecollection.tmpl ?
18:49	kados	sure
18:49	hdl	henridamien at koha-fr dot org
18:49	thd	hdl: Does it work with the latest rel_2_2 or are you and paul using the 2.2.5 release?
18:49	kados	hdl: in fact, they are stock cvs
18:49		hdl: i will delete and reload from cvs
18:50	hdl	thd: we are both on latest rel_2_2
18:57	thd	kados: Your ISBD system preference is empty.  I will fill it
18:58	hdl	kados : I know I first had this error when I first commited. But should have been fixed for more than one week now.
18:59		The first version I commited had to be worked upon.
19:01	kados	hdl: I deleted statecollection.pl/tmpl and grabbed latest cvs copies
19:01		hdl: try to add the serial again please
19:04	thd	kados: you can see both the success and the failure of the Koha ISBD configuration in this record http://opac.liblime.com/cgi-bi[…]detail.pl?bib=109 .
19:06		kados: This record has multiple publishers in 260 $a $b $a $b $c repented as 260 $a $a $b $b $c.
19:08		Which is unavoidable in the current Koha ISBD design.
19:08	kados	right
19:09	thd	kados: However it does have repeated subjects correct.
19:09	kados	it's possible the ISBDdetail code works better than the ISBD code that was included in the authorities display
19:09		I haven't looked at it
19:11	thd	kados: you are missing the field separator from the code that you had used
19:14		kados: In either case the Koha ISBD design needs to change to read field and subfield order from the record rather than from the supplied configuration.
19:18		kados: ISBD configuration should inform the system only which fields and subfields to return at what points as well as what joining punctuation to use.  The order of repeated fields and any specified intra field subfields should be whatever order appears in the record while now it is the order that they appear in the ISBD configuration specified.
19:20		kados: I fixed authorities author searching from the LibLime demo OPAC for the advanced search pop-up last night.   Maybe you missed my announcement in the log.
19:21	kados	thd: I did notice that though I'm not sure what you did ... could you expand?
19:22	thd	kados: I had meant to make the change while you were still awake but was then distracted when I realised that you may be reading too much from a non-working bulkauthimport.pl design.
19:24		kados: All I did was add all the author subfields contained in existing 100 authority records.
19:26		kados: So I added 100d which had been missing from the see also for 100 $a in the default bibliographic framework.
19:28		kados: the default bibliographic framework already had 100q.  I added whatever I could within 255 bytes to the see also for 100 $a in the default MARC bibliographic framework.
19:32		kados: The completed default MARC 21 bibliographic framework will increase the size from 255 to marc_subfield_structure.seealso VARCHAR(2047).
19:34		kados: That is needed to support seealso searching of all relevant subfields from all related fields.
19:37		kados: are you still there?
19:42	kados	thd: sorry ... got disconnected
19:43		thd: so the 255 byte limit for the 'search also' feature is a problem, eh?
19:43	thd	kados: I have fixed it.
19:43	kados	thd: do we need to increase the size of that field?
19:43	thd	kados: Yes
19:43	kados	thd: I'll bring that up to paul as well tomorrow
19:44	thd	kados: of course I have not fixed it at LibLime because I do not have that level of access.
19:44	kados	right
19:44		thd: what was involved in fixing it ... do you have the SQL?
19:45	thd	kados tests my instant SQL  coding
19:45	kados	:-)
19:46		thd: don't you have it in your mysql client history?
19:46	thd	ALTER `marc_subfield_structure.seealso` VARCHAR(2047);  or something like that
19:47		kados: I have been dro[pping the table and recreating it from an SQL file filled with insert statements.
19:49		kados: I could find the confirmed exact answer to your question in a minute but I did it differently,
19:49	kados	right
19:49		tell me what you have in your SQL file
19:49		and I'll commit it to rel_2_2
19:52	thd	kados: `marc_subfield_structure`(`seealso` varchar(2047) default NULL ); is in there but that is not the code you would want ot commit.
19:57		kados: it has been changed to type text in rel_2_2.  What are the limitations on text?
19:58	kados	dunno
20:00	thd	kados: here is the ALTER command:  ALTER TABLE `marc_subfield_structure` CHANGE `seealso` `seealso` VARCHAR( 2047 ) NULL DEFAULT NULL;
20:01	kados	I'll ask paul whether text or varchar would be better there
20:01		thd: thanks for the command :-)
20:02	thd	kados: varchar was the original when I had started working on it.
20:04		kados: a rel_2_2 update must have changed it unless my dump was constrained for compatibility with multiple versions of MySQL.
20:05		kados: Did you see my private messages from an hour and a half ago?
20:07		kados: We do not have to discuss the possibility at this moment but there is a time limitation factor.
20:09	kados	thd: I didn't see it
20:10		thd: time limitation factor?
20:16	thd	kados: did you see my messages now?
20:17	kados	thd: no ... strangely
20:17	thd	kados: Are you connecting from an unusual place?
20:18	kados	thd: no
20:18	thd	kados: have you changed your IRC client configuration?
20:18	kados	nope
20:18	thd	kados: Do you have private messages blocked?
20:18	kados	nope
20:18		:-)
20:18	thd	kados: Or me blocked :)
20:19	kados	nope :-)
20:19		thd: just sent you a private message ... could you see it?
21:06	thd	so the bibliographic framework design does not have a problem but it does have inadequacies for expressing what is mandatory.
21:07		kados: there is no way to express mandatory if applicable.
21:08		kados: So that some 1XX is always filled.  Any cataloguer worth his pay understands that.
21:09		kados: However, issues outside of fill some 1XX may be more likely missed accidentally.
21:10		kados: Also, as you correctly point out it would be helpful to add a check for 6XX as good to have even if not mandatory under some LC standard.
21:11	kados	what about the 'mandatory' flag?
21:11		(phone call)
21:12	thd	kados: It should also be possible to create incomplete records that require further editing.
21:13		kados: signify when your are off phone he said as he stepped out of the room for a moment.
21:35		kados: so If incomplete records were submitted there should be a warning to accept an incomplete record instead of refusal, although I do not remember what the message is now when leaving out something mandatory.
21:36		kados: mandatory needs gradations at some point.
21:37		kados: more fine grained aspects for individual indicator and fixed field positions
21:39		kados: there is something that ought to be changed now for the frameworks though.
21:39	kados	what's that?
21:39	thd	kados: a new column is needed for obsolete fields and subfields
21:40	kados	could you expand on what those are?
21:42	thd	kados: a modest code change in a few places would allow legacy fields and subfields which should no longer be used to create current records but still bring them up in the correct place if filled with legacy data.
21:43		kados: changes over time have moved the correct place to encode data from one place to another.
21:44		kados: Libraries who cared and had money sent their records off for conversion to the newer format.
21:46		kados: However, some obsolete fields were not even moved but simply abandoned with changes in AACR.  Those fields still have data of value if they are populated.
21:48		kados:  There tends to be a correlation between libraries with money and libraries with the oldest MARC records but there is a big gap where records may have been copy catalogued by poor libraries from rich libraries with old records and then never updated at the poor libraries.
21:49		kados: so my first purpose with the default frameworks has been to ensure that users do not loose data that they already have.
21:50		kados: However, the options to support visibility of content are very limited in the Koha frameworks.
21:51	kados	that's true
21:52	thd	kados: No one wants their record editor forms supplying them with many obsolete fields and subfields merely to support the small but valuable set of records that use obsolete fields.
21:52	kados	good point
21:53		but I think that's a version 3.0 discussion
21:53		paul and I will be working on a new XUL-based design
21:53	thd	kados: Although, I suspect that in the real world at least some obsolete subfields do not have small use.
21:53	kados	so when we're ready to spec that out we'll definitely discuss it with you
21:53		my goal with 2.2.6 is to be MARC compliant
21:54		I'm not as concerned about usability
21:54		because the current design is deficient
21:54		and it's not worth spending too much energy patching a sinking ship :-)
21:55	thd	kados: But this is a problem now because the same column that is used to place the subfields into the correct tab for viewing in the right sequence is also the same one for setting them to not viewable.
21:56		kados: I do have an experiment to conduct with the hide option but that certainly does not help visibility..
21:57		kados: The values for subfield management are -1 - 10.
21:58		kados: -1 preserves content but hides, 0-9 are tabs, and 10 is the items tab.
22:01		kados: I thought of changing behaviour by extending the value range so that obsolete would be shown only if present where -7 is obsolete for tab 7 but then -0 for tab 0 is not a numeral.
22:01		kados: Therefore another column is required.
22:03		kados: I do like paul's frameworks very much.  I see no reason to abandon them for 3.0.  They might be extended or transformed.
22:04	kados	why do we need another colum?
22:04		why can't we just extend it as you propose?
22:05	thd	kados: Because we have no column to support hide if empty because this field/subfield is obsolete but show if data present.
22:06		kados: The problem is that -0 is the same number as 0 for obsolete fields/subfields that should appear in tab 0.
22:07	kados	thd: it should automatically hide if empty in the record display
22:08	thd	kados: I am also concerned with the record editor being filled with obsolete fields and subfields.
22:10		kados: No one will want those in the way if empty but everyone will want to see them when editing old records if they have content.
22:10	kados	we could create a tab '11'
22:10		for obsolete fields
22:11	thd	kados: Then they would not appear in the correct logical place in relation to the rest of the record in either the record editor or the OPAC.
22:12		kados: If I am not mistaken tab assignment sets order of appearance even in the OPAC MARC view.
22:13		kados: and their is not even a tab 10 for the record editor.  Tab 10 is for items.
22:14	kados	thd: i think it's ok if they don't appear in the logical place
22:15		thd: the goal is compliance, not usability for 2.2.6
22:15		thd: there could easily be a tag 11 for the record editor
22:15	thd	kados: Adding a new column and modifying code to support it is also much less coding than managing than whatever would be needed to extend the original column.
22:15	kados	thd: it would not take much time to add that
22:15		thd: maybe ... I will bring it up to paul tomorrow
22:16	thd	kados: new column is much less time I believe.
22:16		kados: Especially as until something is added to support it nothing would break.
22:17		kados: another problem is the lack of any way to separate OPAC from patron visibility.  There are fields and subfields which are private or non-public either absolutely or if an indicator is set.
22:18	kados	I think if you leave out the OPAC description for a given subfield it will not display but I haven't tested that theory
22:19	thd	kados: Anonymous donor may stop giving if his name appears in the OPAC as one example of the consequence.
22:20		kados: But that information can never be entered in the record editor in the first place or changed for existing records if the subfield is hidden.
22:22		kados: The possibility that a blank value for the OPAC label hiding the content is interesting.
22:23		kados: That does not cover the equally common case where the indicator is supposed to be read to determine privacy.
22:24	kados	thd: we need to account for that I agree
22:25		thd: if you write a brief description of the feature and what indicators use it i will gladly code it
22:25	thd	kados: I will definitely experiment with a blank OPAC value for the subfields labelled non-public.
22:25		s/value/label/
22:26		that would at least cover the explicit case if it works.
22:29		kados: Another framework column would also be good for collapse or hide in JavaScript editor unless explicitly opened or expanded.
22:30		kados: This would be for the uncommonly used but occasionally needed fields and subfields.
22:31	kados	right, that woudl be nice
22:32	thd	kados: No one will be happy with a default framework that preserves all data but always has uncommonly needed fields and subfields in the way.
22:34		kados: As we discussed before, the current option is only to hide them where they cannot be recalled without changing the framework rather than as the need arises for the material being catalogued.
22:36		kados: Another great inefficiency I have seen with the record editor is that context is not preserved when adding repeated fields.
22:37		kados: If I add a second 650 I am thrown back to the beginning of tab 0.
22:38		kados: This is usually manged with internal anchors while the page refreshes.
22:40		kados: I do not know what extra state information would be needed to preserve tab location in JavaScript.  I have never looked at the record editor code.
22:41		kados: A page refresh is also required and absent to change labels after reordering subfields.
22:42	kados	I've got to head out
22:42		thanks for your explainations thd
22:42	thd	kados: OK
22:42	kados	I'll discuss this with paul tomorrow
22:43	thd	circa what time if I rise :)
22:43		?
22:43	kados	undetermined
22:43		but it will probably be a pmesg
22:43		as we may discuss money things :-)
22:43		ok ... I've really got to go now
22:44		ciao
22:44	thd	good evening kados
22:45		kados: you should try to keep the generally useful things on the open channel as I just did for logbbot
00:17		chris: I have found a problem on #koha
00:17		chris: are you there?
00:18		or anyone from Katipo who knows the IRC system
00:18	Jo	Mason: can you come to hlt
00:18		testing your bug fix..
00:19	thd	russ si: do you know the IRC system?
00:20		the one running #koha?
00:24		sorry chris russ si: I had used syntax that was not parsed correctly
00:24		I was certain that I had always done it the same way but I must be mistaken
00:25	russ	thd: so you are all good now?
00:26	thd	russ: yes sorry to have disturbed you with user error issuing IRC commands
00:27		russ: unless I am now living in an alternate universe :)
00:29		back in the correct universe now
00:39	russ	thd :-)
04:27	hdl	kados : I tried npl and default templates :
04:29		npl donot propose data input for barcodes or branches and so on and so forth...
04:31		And default templates ends in Internal server error.
04:31		I shall try and delete mine that works, to see if I get the same error.
04:37	thd	hdl: Is kados awake?
04:37	hdl	thd: seems not.
04:37		But I tried anyway.
04:37		:)
04:39	thd	hdl: he was awake long past now Friday night and Saturday morning when he could not stop working on authorities but he is not as crazy as me ordinarily.
04:48	russ	hi hdl
04:48	hdl	hi russ
04:48	russ	do you know if paul is going to be around soon?
04:48	hdl	in about half an hour.
04:48	russ	cool
05:00	osmoze	hello
05:41	paul	russ, i'm here.
05:42	russ	hi paul
05:42	hdl	hello osmoze
05:42	russ	its census day in nz
05:42		just finishing my form :-)
05:44		the conference - how is that being organised?
05:44	paul	census day ?
05:44		throw the questions.
05:45	russ	http://www.stats.govt.nz/census/default.htm
05:45		ok here we go
05:45		do people need to pay? do they need to register?
05:46		this is for the events on the tuesday and the weds
05:46	paul	are you speaking of the day for public demo (2-3) or the developper meeting ?
05:46	russ	public
05:46	paul	ok
05:46		we think we will ask for a small fee (something like EUR15) for the 1st day.
05:47		librarians will be OK to pay a small fee, and, once they have pay, they come. Otherwise, we may have zillions of registering, but only a few people.
05:47	russ	yep
05:47		it is a good idea to ask for a small amount
05:48	paul	for day 3, we will probably ask a little bit more. the idea being to have some funds to give to the future NPO. that could be something like EUR50
05:48		(May, 3 I mean. It's day 2)
05:48		;-)
05:48	russ	yep
05:49		are you thinking of hiring a professional conference organisor - or doing it yourselves?
05:50	paul	we will do it ourself (Ineo & me)
05:50	russ	bit of background: i have lost count of the number of conference websites we have done for vairous things
05:50		we = katipo
05:50	paul	we already have a room (= ENSMP, still to be confirmed, but seems OK)
05:51	russ	great
05:51	paul	we have a mail address conference.mai06koha-fr.org for registering.
05:51	russ	ok cool - so if was to do a form on the koha.org site, it could submit to that address?
05:51	paul	Ineo will manage mailings to invite pple widely (they already have the coordinates of the 1000 main libraries in France + we have 2-3 major mailing lists)
05:51		yep.
05:52		and Ineo will manage registering too.
05:52	russ	right - so they will take the payments?
05:52	paul	I just manage the "who speaks of what"
05:52		payments : still undefined.
05:52	russ	ok
05:53	paul	(if osmoze reads this thread, he will see very interesting things. pls jean sebastien, don't speak publicly of this yet. announcement in a week or 2)
05:53	russ	ah
05:54	paul	for the developper meeting, we already have all what is needed :
05:55		* the room (in CMI, a french library of mathematics, in Marseille, in one of Marseille University)
05:55		* a good net connexion
05:55		That's enough.
05:55	chris	:)
05:55		and a cafe nearby ?
05:55	paul	+ some restaurants close. Just have to check if we can get some vegetarian food.
05:56		hello chris.
05:56	chris	hi paul
05:56	paul	I'll take care of coffee, stil have to speak with anna on those details.
05:56		we also should have 2 librarians :
05:56		* pascale nalon, from ENSMP
05:57		* Anna W..., CMI librarian (she's polish, her name is a nightmare for a frenchy, that's why I write W...)
05:57	russ	:-)
05:57	chris	:-)
05:57		ive been practising henri damien
05:57	osmoze	ok paul :)
05:57	chris	paul is quite easy :-)
05:58	paul	lol (and available is most languages i've been told)
05:58	chris	yep, its pretty common
05:59	paul	russ : another question about the meeting ?
05:59	russ	umm
05:59		it sounds like it is still a bit loose
05:59		as in loosely organised at the moment
06:03	chris	paul, when its a little bit closer to the time, do you think we could/should contact some media in paris and marseille, see if they are interested
06:03		?
06:04	paul	I think/hope Ineo will do. They plan to do something with specialized press at least (there are something like 5 librarian oriented news papers)
06:04	chris	cool
06:05	paul	I'm not sure TV and non-specialised press could be interested.
06:07	chris	i would think only small non-specialised papers might be .. if you have them (here we have a few small free newspapers, who run pretty much any articles people send them)
06:07		but the big papers are only interested if its something big, or new
06:10	russ	paul - i will start making some pages to go up on koha.org that can be used for reference
06:10		and promoting the event
06:11		is that ok with you?
06:12	paul	yep
06:13	russ	cool - i'll use the info i already have, and i will forward you a login to the cms for the website so you can see it before it goes live
06:13		but now i need to go to bed
06:13		catch you later
06:14	paul	have a good night
07:57	thd	paul: are you there?
07:57	paul	leaving for lunch in 10mn...
07:58	thd	paul: you mean I have ten minutes to ask you something before you leave for lunch :)
07:58		?
07:58	paul	yep.
07:58		but i'll be back in one hour and a half.
07:58		(say in 2, to be sure)
07:59	thd	paul: authority types are stored in two tables.
07:59		auth_header and auth_types.
08:00	paul	yes
08:00	thd	paul: what is the function of each. Why two tables.
08:00		?
08:00	paul	auth_types just contain authority definition (head of structure)
08:00		1 row for 1 auth type
08:01		auth_header contains 1 row for each auth type.
08:02	thd	paul: would that same content not function in one table?
08:02	paul	nope.
08:02	thd	paul: I miss something.
08:02	paul	(look at the DB scheme, on koha.org/wiki)
08:03		auth_type : PERSONAL NAMES
08:03	pierrick	I'm looking for sampledata compatible with current HEAD
08:03		misc/sampledata-1.2 does not work, due to DB scheme modification
08:03	paul	auth_header : authority number 1353968, that is a PERSONAL NAMES
08:03	pierrick	(hello, by the way :-)
08:03	paul	authority number 1244535, that is also a PERSONAL NAMES
08:03		hello pierrick
08:05	thd	paul: Are there also two tables for bibliographic frameworks?
08:05	paul	yep :
08:06		biblio_framework and marc_biblio
08:06		(I mean in 2.2.x, as marc_biblio will die in 3.0. and the link to framework will be/is stored in biblio table)
08:07	thd	paul: So you mean the framework types could be stored in one instead of two tables :) ?
08:08	paul	nope.
08:08		the framework attached to a biblio is in marc_biblio in 2.2
08:09		as this table is dropped, we have to put it somewhere else.
08:09		that will be in biblio !
08:09	thd	paul: OK now the function is clear.
08:13		enjoy your lunch paul
08:24	pierrick	Paul: do you have sample data to give me ?
08:25	paul	mmm... c'est sur channel privé !
08:26	pierrick	bon alors je quitte irssi que je ne maîtrise pas du tout, je reviens avec Xchat, c'est plus intuitif
09:41	osmoze	c est bien mieux xchat non pierrick  ;)
09:50	kados	hi guys
09:51		pierrick: I'd suggest you load in data elements separately so you can get familiar with the process
09:51	pierrick	(xchat, ça dépanne, mais je compte bien maîtriser irssi)
09:51		hi joshua
09:51	kados	pierrick: so you could, for instance, create a CSV file with some fake borrower information
09:51		and use LOAD DATA INFILE to  load it in
09:52	paul	hello kados. nice sleep ?
09:52	pierrick	paul gave me a dump of a 2.2 installation from EMN (Ecole Mines Paris)
09:52	paul	full of authorities dreams ?
09:52	kados	you could also obtain some MARC and use bulkmarcimport ...
09:52		paul: yes in fact :-)
09:52		pierrick: that'll work too :-)
09:53		paul: so I actually have several questions
09:53		paul: looks like thd may have already asked some of them ...
09:53	paul	and i think it's better when starting : he will have enough problems with HEAD bugs to avoid DB ones ;-)
09:53		throw them, i'm listening
09:53	kados	my first question is whats' the best way for thd to distribute the new MARC21 bib framework he's been working on?
09:54		In fact, what is the best way to upgrade a framework (bib or auth)
09:54	paul	a sql dump of the 2 tables. then i'll include them as replacement in marc_datas/sql/marc21_en
09:55		delete * from marc_*_structure where frameworkcode='';
09:55		then mysql -uX -p XXX <new_framework.sql
09:55		(I suppose the framework is the default one.
09:55		)
09:55	kados	that first query scares me :-)
09:55		yes, default
09:56		ok good ... next questions
09:56		the current size of the 'search also' field is 255 chars
09:56	paul	(i've a quick question too)
09:56	kados	(or actually, it might be of type 'text' now)
09:56	paul	it could be TEXT, I agree
09:57	kados	ok ... what is the upper limit on TEXT?
09:57		it should be enough I think
09:57		your turn to ask ... :-)
09:58	paul	I don't understand what LabelMARCView does
09:58		(your new systempref)
09:58	kados	right ...
09:58	paul	(i've another question in fact, so 2. But the 2nd one is not a quick one ;-) )
10:00	kados	LabelMARCView: the previous code for display of MARC in OPAC results
10:00		was customized for your clients to be non-standard IIRC
10:00		so for example
10:00		say you have two 650 tags like this:
10:00		650 $a $b $c
10:00		650 $a $b $c
10:01		your clients prefer the MARC labels to look like this:
10:01		650
10:01		a
10:01		b
10:01		c
10:01		a
10:01		b
10:01		c
10:01		but my clients prefer:
10:01		650
10:01		a
10:01		b
10:01		c
10:01		650
10:01		a
10:01		b
10:01		c
10:01	paul	ah, OK got it !
10:01	kados	make sense?
10:02	paul	my second question then ?
10:02	kados	sure
10:02	paul	or another one from you ?
10:02	kados	ok ... one from me :-)
10:02		what is the 'itemcallnumber' setting in sysprefs?
10:03	paul	it's used to define the DEFAULT biblio subfield that will be automatically reported as itemcallnumber when creating a new item.
10:03		for example : 676a if a library has physical shelves based on dewey
10:03	kados	why is it defined in sysprefs rather than the MARC framework?
10:04	paul	it's the default value. Thus, when you edit an item, it's automatically filled with 676$a, but you can overwrite it.
10:04		because it was simple, quick & not too dirty.
10:04	kados	ok
10:04		your turn :-)
10:04	paul	(+ I don't think we need a different value for each framework)
10:04		(but maybe i'm wrong)
10:05	kados	(only if a library uses more than one classification system I suspect)
10:05		(this may apply to NPL for instance)
10:05	paul	should we call the next release 2.2.6 or 2.4.0 ? because there are a LOT of new features, even if the DB schema did not change.
10:05	kados	(they use Dewey for non-fiction and a internal classification for Fiction)
10:05	paul	ok, for NPL, it would make sense to have itemcallnumber based on framework.
10:06	kados	2.4.0 perhaps
10:06		it's ok with me to be 2.4.0
10:06	paul	i'm afraid we could introduce new bugs with what is in rel_2_2, thus I think it's better to move the 2nd digit to warn the libraries.
10:06		ok, i'll write something to koha-devel.
10:06	kados	true
10:06	paul	expect it in april in your mailbox ;-)
10:07	kados	heh
10:07	paul	(unless savannah goes faster suddenly)
10:07	kados	I hope so
10:07		I have two questions left ... both quite long
10:07		shorter one first:
10:08		I'm having trouble with rel_2_2 serials
10:08		sometimes I cannot save a newly created serial
10:08		http://bugs.koha.org/cgi-bin/b[…]w_bug.cgi?id=1048
10:08	paul	I saw (if you speak of adding a serial)
10:08	kados	I also don't understand how to add an item with a serial
10:08	paul	but i'm like hdl : I never had this problem.
10:08	kados	right ...
10:09	hdl	kados
10:09		did you update C4/*
10:10	kados	I will check right now
10:10	hdl	(I don't remember which C4 file I modified to add serialsadditem Koha.pm or Bull.pm)
10:11		(paul : kados problem results in a 550 Internal server error)
10:11	kados	I just did cvs update from C4 dir ...
10:11		no reported updates
10:11	hdl	Really Strange.
10:11	kados	hdl: maybe I need to run updatedatabase?
10:12	hdl	you may try.
10:12	kados	I run it now again
10:12	paul	there is a new systempref, so maybe yes.
10:12	hdl	But normally, if you have serialsadditems
10:12	kados	I do
10:13	hdl	The problem comes from a failure in HTML::Template output calculation
10:13	kados	what version of HTML::Template are you running?
10:13	hdl	He says we are passing him a scalar in place of loop.
10:14	kados	right
10:14		$ updater/updatedatabase
10:14		connected to your DB. Checking & modifying it
10:14		Alter encoding in marc_breeding
10:14		synch'ing borrowers
10:14		synch'ing items
10:14		synch'ing biblio
10:14		synch'ing biblioitems
10:14		Checking for data required in table userflags...
10:15		Checking for data required in table systempreferences...
10:15		done
10:16	hdl	2.7
10:16		(HTML::Template version)
10:17	kados	hmmm ... I just added a subscription and it seems to have worked
10:17		hdl: can you try adding the one you did yesterday?
10:17		(IIRC yesterday, I could add one but your's failed)
10:18		(suggesting the problem is not always present)
10:18		btw: 'Waited' should be 'Waiting' in english
10:19		(in Subscription summary, 'end date' is missing)
10:19	hdl	can you show me the logs ?
10:19		Internal server error
10:19	kados	did it fail ... shoot
10:20		[Tue Mar 07 05:17:39 2006] [error] [client 62.212.120.129] HTML::Template->output() : fatal error in loop output : HTML::Template::param() : attempt to set parameter 'itemlocationloop' with a scalar - parameter is not a TMPL_VAR! at /usr/lib/perl5/site_perl/5.8.5/HTML/Template.pm line 2963, referer: http://koha.liblime.com/cgi-bi[…]?subscriptionid=3
10:20		[Tue Mar 07 05:17:39 2006] [error] [client 62.212.120.129]  at /var/www/liblime.com/koha/intrane​t/cgi-bin/bull/statecollection.pl line 160, referer: http://koha.liblime.com/cgi-bi[…]?subscriptionid=3
10:20		[Tue Mar 07 05:17:39 2006] [error] [client 62.212.120.129] Premature end of script headers: statecollection.pl, referer: http://koha.liblime.com/cgi-bi[…]?subscriptionid=3
10:20		hdl: http://koha.liblime.com/cgi-bi[…]?subscriptionid=3
10:21		hdl: is that the one you added?
10:21	hdl	This is the one I TRIED to add.
10:21		Never could get onto the page.
10:21	kados	strange that it still shows up ...
10:22	hdl	would you make me available a dump of your base and a tarball of your install ?
10:22	kados	sure
10:24		liblime.com/koha_liblime.tgz is the install tarball (well, it is the cvs repo the install is symlinked too)
10:25		dumping db now
10:26		gzipping
10:27		hdl: how do I add items to a subscription?
10:27		(while I wait)
10:28	hdl	In default template :
10:28	kados	liblime.com/koha_liblime.sql.gz is the database
10:30	hdl	you have a new line in statecollection page :
10:31	kados	same error
10:31		interesting
10:31		why does this:
10:31		http://koha.liblime.com/cgi-bi[…]?subscriptionid=4
10:31		statecollection page does not fail
10:31		but this:
10:32		http://koha.liblime.com/cgi-bi[…]?subscriptionid=3
10:32		it does ... !
10:32	hdl	Numbered   planned for   Status   Notes
10:32		Issue
10:32		Callnumber Barcode  branch itemstatus location
10:32	kados	wait ... I'm wrong
10:32		it fails on both :/
10:32		(i clicked on the wrong button)
10:32	hdl	unfortunately
10:32	kados	hdl: ahh ... I haven't seen this page yet because of Internal Server Error :-)
10:33		paul: while hdl sorts out serials problems ...
10:33	hdl	tarbal ?
10:33	kados	paul: have time for my last question?
10:33	paul	yep
10:33	hdl	ok
10:33	kados	hdl: the db is a gziped mysqldump
10:34		paul: ok ... so my last question is actually several questions
10:34		related to authorities
10:34		some of them I asked yesterday
10:34		for instance: what is the role of the 'framework'?
10:35	paul	define the MARC templates, for biblio editing
10:35		I use framework in biblios.
10:35		and authtype for authorities
10:36	kados	right, but they are also used as search points in authorites if I'm not mistaken
10:37	paul	mmm... I don't see what you're speaking of.
10:37	kados	a search point means that when you go to search
10:37		you have an option to search: 'titie', author, subject
10:37		in authorities search, you have search points defined as the frameworks
10:38		so if you have NAME, SUBJECT, TITLE ...
10:38		the user must select one before performing the search
10:38		the problem is ... if I understand correctly
10:38		in MARC21 there are only 4 valid search points
10:39		names, name/title combinations, uniform titles, or subjects.
10:39		(also called the headings)
10:40		but I _think_ there are potentially many more editing frameworks desired
10:40		but as I'm not a cataloger I can't be 100% sure
10:42		paul: does that make more sense?
10:42	paul	yes, but I don't see where the problem is
10:42		(or am not sure to see)
10:43	kados	ok ... we can skip that one :-)
10:43		until I have more information
10:43	paul	In unimarc we have a little bit more search points, as, for authors, you may have author PERSONAL NAME or CORPORATE name.
10:43		ok, let's delay this question
10:43	kados	in MARC21 PERSONAL NAME and CORPORATE NAME are grouped within NAME
10:43		for _searching_
10:43		but not for _editing_
10:44		but we will skip for now :-)
10:44		I have read through both build_authorities.pl and bulkauthimport.pl
10:45		I cannot understand fully the whattodo hash
10:45		I will paste in each element with the included explaination then explain my question:
10:45		\ttaglist : the list of MARC tags using this authority
10:46	paul	(bulkauthimport is really outdated & probably don't work at all)
10:46	kados	does this mean the list of bib MARC tags that this authority will fill?
10:47	paul	=> yes
10:47	kados	for a NAME entry, this will include: X00, X10, X11, X51 ... very many tags!
10:47	paul	in UNIMARC, there are not so many.
10:48		(at most 3)
10:48	kados	key is all subfields, I understant his
10:48	paul	NO.
10:48	kados	heh
10:48	paul	it's all subfields that you want to use as PRIMARY KEY.
10:48	kados	meaning all subfields to be used in a search?
10:49	paul	it depends on what you try to migrate. let me take some examples...
10:49		===== 1ST one ======
10:49		(using 700 as personal name, and UNIMARC)
10:49		700$31234$apoulain$bPaul$d1968-
10:49		700$1234$aPOULAIN$bPaul
10:50		oups, a 3 missing.
10:50		700$31234$aPOULAIN$bPaul
10:50		if you set key = 3 (and only 3) then those 2 lines are similar.
10:50		and build_auth consider them as only 1 authority.
10:51		if you set key=$a$b$d, then they are NOT the same authority
10:51		and you will get 2 differents authorities entries.
10:51		depending on your DB structure/quality, you may want to do one thing or the other !
10:51	kados	$3 is linkage?
10:52	paul	for every subfield you want to report in authority but NOT use as a primary key, you have the "other"
10:52		yes
10:52		there maybe more than 2 "john doe" in your DB
10:52		maybe one is 1840-1894 and the other 1920-
10:52		maybe you can't rely on birthdate.
10:53		you have to choose between quality of your data and quality of your authorities !
10:53	kados	hmmm
10:53		I almost understand
10:53		in marc21 if I have good data and authortiies
10:53		I should select $9 as the key
10:54	paul	if you put key=a|b|d and have "1840-1894" and "1840-1894 (dec)" in 2 differents biblios, then you'll get 2 authorities, even if it's the same person
10:54		you already have a $9 ?
10:54	kados	it is linkage right?
10:54		I set it up for some tags
10:54	paul	because it's supposed to be created/added by build_authorities.
10:54	kados	I don't already have $9
10:54	paul	no, the $9 is our INTERNAL-KOHA linkage.
10:54		it's filled by build_authorities
10:54	kados	ahh ... so you have cases where an ID was preserved from the old system
10:55		currently I have no such cases :/
10:55		in fact, I think MARC21 does not have this feature
10:55	paul	the diff between unimarc and marc21, is that in unimarc we have the $3
10:55		that does not exist in marc21
10:55	kados	right
10:55	paul	iirc, in marc21 "linkage is a local problem, solved by local solutions".
10:56	kados	right
10:56	paul	that's why authorities are better in unimarc imho
10:56	kados	yep
10:56		so what do you suggest I put in 'other' when importing MARC21 authority records ?
10:57		(I don't quite understand what to do)
10:57		and what should be in 'key'?
10:57	paul	it depends on what is in your datas !
10:57		if you are confident with your datas, are sure that their quality is correct, you can put every subfield in key.
10:58	kados	if I have many "john doe" some with birthdate and some without in bib records
10:58		but very good auth records (fresh from OCLC)
10:59	paul	(that's a situation build_authorities can't handle yet, but I answer and suppose you will modify the script)
10:59		then you should do :
10:59		* check with all subfields => authorities found : ok, i'm sure

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