IRC log for #koha, 2006-07-13

Time	Nick	Message
12:19	kados	sheesh
12:19	kados	phone ringing off the hook
12:19	kados	owen: ok, I'm back :-)
12:20	kados	so ... where shall we start?
12:21	owen	Yesterday as far as we got was, "needs to fit 800x600" monitor
12:22	kados	right
12:22	kados	so as far as the OPAC goes ...
12:22	kados	the 800 x 600 thing
12:23	kados	anything else?
12:23	kados	the faceted searches we're going to skip I think, right?
12:23	owen	Was that issue with the search form or the results?
12:23	kados	the search form I think
12:23	owen	Both look okay to me
12:23	owen	Only the simple search is a little squeezed, but that's just the big form field.
12:24	kados	I'll send you a screenshot
12:26	kados	owen: check your inbox
12:28	owen	I think yes to skipping the faceted searches, in the interests of time
12:30	owen	Wow, kados, that's remarkably different than Firefox's rendering.
12:31	kados	owen: that _is_ firefox :-)
12:31	kados	owen: it looked the same way on the circ computer at the plains
12:31	kados	owen: which I assume is win98 or 2000
12:32	owen	Just have to resize the form field a little
12:33	kados	so ...
12:33	kados	as far as maintenance goes, are you using a fresh copy of the dev-week code?
12:33	kados	cuz that's what's up on zoomopac
12:34	kados	(do you need write access to more than just the template files?
12:34	kados	(to upload changes?
12:34	owen	I don't know
12:34	kados	ok, well now you've got access to the koha-tmpl dir
12:35	kados	let me know if you need more
12:40	kados	http://zoomopac.liblime.com/cg[…]tail.pl?bib=49201
12:40	kados	I suspect these are triple or even quadrupal encoded
12:41	kados	marc8->latin1->latin1->latin1 :)
12:52	owen	Did you see the issues with "you did not specify any search criteria" with sorted search results?
12:53	kados	yea
12:53	kados	I think that's fixed in devweek
12:53	kados	but if not I'll certainly have a look
12:54	kados	(not that interesting to you :-)
12:58	kados	thd: 1106 is fixed
12:59	thd	kados: is the fix committed or where can I test the fix
12:59	kados	thd: afognak has the fix
12:59	kados	thd: and it's commited
12:59	thd	goody
13:06	thd	kados: works nicely now
13:06	kados	thd: great!
13:06	thd	kados: where is the working editor code if it is not in devel-week?
13:07	thd	kados: only on the Liblime server?
13:08	kados	thd: it is now working in dev-week and rel_2_2 afaik
13:08	kados	thd: dev-week for sure
13:09	thd	kados: you have updated devel-week since it was not working yesterday?
13:09	kados	thd: yes
13:09	kados	thd: dev-week is now completely in sync with rel_2_2 except without the MARC editor bugs :-)
13:10	kados	thd: didn't you see the 40+ commits this morning? :-)
13:10	thd	kados: I had to avoid overfilling an email quota with commits so I do not see them in real time.
13:11	kados	ahh
13:13	thd	kados: I would need to have much more time than I have to fix my local mail system again to solve that quota problem
13:14	thd	kados: what Koha SQL columns need linking to the bibliographic framework for 2.2.4?
13:15	kados	?
13:15	thd	kados: did you link only 2 yesterday
13:15	thd	?
13:15	kados	linking?
13:16	thd	kados: items.onloan etc linked to 952 $X?
13:17	kados	ahh
13:18	kados	yes, that's the only one
13:18	kados	thd: I could use your advice on a good mapping scheme
13:18	thd	kados: did you not need another for date?
13:18	kados	date is in issues, not in items
13:19	thd	kados: I thought you added date because things were not working. Not that they worked afterwards.
13:20	thd	kados: what do you want to map in a good mapping scheme?
13:20	kados	well ...
13:20	kados	thd: items fields mainly
13:21	kados	thd: we need to come up with a scheme for call number
13:21	kados	thd: that may include adding a column to items (fine by me)
13:21	kados	thd: all of the statuses need to be mapped
13:21	kados	things like datelastborrowed would be useful too
13:22	kados	datelastseen, etc.
13:22	kados	the more we can store in zebra the better
13:22	kados	because then we can search on it
13:22	thd	kados: what are all of the statuses?
13:22	thd	kados: how many are they?
13:22	kados	notforloan, itemlost, wthdrawn,binding I think
13:23	kados	thd: we also need to clean up the descriptive language of the item labels
13:23	kados	thd: so that a librarian is not confused :-)
13:23	thd	kados: those are already mapped except for binding
13:23	kados	paidfor
13:23	kados	another status
13:24	thd	kados: the descriptions were written to not confuse me :)
13:24	kados	:-)
13:24	kados	yes, I understand that
13:25	thd	kados: do you think that we might run out of punctuation symbols?
13:25	kados	thd: also, barcode should not be a mandatory field
13:25	kados	thd: but I agree we need a method to alert if a field hasn't been filled in but let the cataloger proceed
13:25	thd	kados: I never set it to mandatory except by request
13:26	kados	thd: it's set to mandatory in the framework I believe
13:26	thd	kados: I think that had been your request at one time
13:28	kados	perhaps :-)
13:30	thd	kados: I have code for filling the various call number parts for use in labels etc.
13:30	kados	in perl?
13:31	thd	yes
13:32	thd	kados: however, there is no code for what we had discussed previously for the case where NPL needs a couple of system preferences for treating fiction differently etc.
13:34	kados	well ... not necessarily
13:34	kados	it can be achieved with minimal coding
13:34	kados	if we distinguish between classification and call number
13:34	kados	and local call number
13:35	kados	and define local call number as either classification or call number depending on whether it's fiction or non-fiction
13:37	thd	kados: to remind you we need a preference for call number type LCC, DDC, UDC, etc. and a preference for cutter type and rule and a preference for secondary classification defined by an exception list such as material type specifying fiction.
13:39	kados	hmmm
13:39	thd	kados: mere classification is unknown because there are many different classifications and MARC never has one and only one place to hide a given classification.
13:39	kados	couldn't we just define it in the framework instead of a preference?
13:40	thd	kados: define what in the framework?
13:41	kados	thd: define the type of classification being used
13:41	thd	kados: we could add a column to the framework to specify which locations are good for identifying particular parts of a call number.
13:42	kados	thd: ie, the mappings would define it for us
13:43	thd	kados: the type of classification that a library is using is up to the library and cannot be defined for all libraries
13:44	kados	thd: what I mean is that each library will have to define that part of their framework
13:44	thd	kados: the bibliographic framework should not be the place to define whether the library prefers DDC, UDC, LCC, etc. should it.
13:44	kados	hmmm
13:45	kados	so what is our objective?
13:45	thd	kados: you could define the standard locations in advance but they are multiple
13:45	kados	to search and display call numbers correctly right?
13:45	kados	do we need to distinguish between call numbers and classification?
13:46	thd	kados: the objective is to fill the call number by the library's preference
13:46	kados	thd: fill it with values coming from the MARC?
13:46	kados	thd: because with zebra, we don't need it to be in the koha tables to search or display it
13:46	thd	kados: call numbers are numbers of which the classification element is only one part
13:47	kados	thd: the only reason for it to be in the koha tables with zebra is if it is used by the system in managing the collection somehow (like circ rules)
13:47	kados	(or inventory)
13:48	thd	kados: it does not need to be in the SQL tables
13:50	thd	kados: MARC has provided 852 $k $h $i $m
13:51	kados	thd: http://wiki.koha.org/doku.php?[…]rchingdefinitions
13:51	thd	kados: I provided all those in 952 in the framework
13:51	kados	thd: this is a document I wrote for NPL to help them clearly define their cataloging practices
13:51	kados	thd: I think it would also be useful in this case
13:52	kados	thd: the 'organization of materials' section is most relevant
13:52	kados	thd: I could use some expansion on what I've written to more accurately describe the components of collection organization
13:56	thd	kados: it seems as if they are using 942$c for the equivalent of $852 $k call number prefix
13:57	kados	thd: that is my mistake
13:57	kados	thd: it should read 952$c
13:58	kados	fixed
14:01	thd	kados: do they have the values in 952 $c on their spine labels?
14:04	kados	yes
14:04	kados	owen: right?
14:04	owen	Sorry, what? What's 952 $c?
14:04	thd	kados: then they are using that for 852 $k
14:06	thd	owen: the MARC location defined in the framework used originally by NPL to store the contents of `items.location`
14:07	owen	Our current production database doesn't store anything in items.location.
14:08	thd	kados: I think that you had it correct previously it was 942 $c
14:09	thd	owen: do the item type codes that you use appear on the spine labels?
14:10	owen	Yes
14:10	thd	kados: you were correct originally you should revert to the previous version in the wiki
14:11	kados	right
14:11	owen	biblioitems.itemtype
14:13	thd	owen kados: NPL uses biblioitems.itemtype stored in 942 $c as the equivalent of 852 $k
14:13	kados	right
14:13	thd	kados: all our libraries do the same as far as I know
14:14	thd	s/our/your/
14:15	kados	yep
14:15	kados	but 942 $c is not where the spine labels are printed from
14:15	kados	afaik the spine labels are coming from a value mapped to callnumber
14:15	thd	kados: so you should fill 852 $o in the current standard MARC 21 framework with the contents of 942 $c
14:16	kados	but that's the itemtype
14:16	kados	not the call number
14:16	thd	kados: I thought that NPL was not yet using Koha for spine labels
14:16	kados	they aren't
14:16	kados	but they do generate spine labels
14:16	kados	from fields in the MARc
14:17	thd	kados: yes but that is way in which NPL is using the itemtypes that it has defined
14:18	thd	kados: 852 $k is usually something like JUV, REF, FIC, etc. at most libraries
14:19	thd	kados: those are item types of some sort but not a true material type
14:20	kados	right
14:20	kados	is there a standard that defines them somewhere?
14:21	thd	kados: no just common practise
14:21	thd	kados: material types have multiple standards
14:21	kados	which turns out not to be very consistant :-)
14:22	thd	kados: amazing how far a little cultural conformism will go to address the same problem in the same manner
14:24	thd	kados: oh you were saying it was not consistent. Each library has to apply the basic intent to its own needs.
14:25	thd	kados: yet English language usage provides FIC more usually as opposed to FICT, REF as opposed to REFE, etc.
14:26	kados	:-)
14:26	thd	kados: and then there is the need to economise the space on the small spine label.
14:27	kados	right ... any suggestions on that?
14:28	thd	kados: the 852 $h $i tend to be more standard than 852 $k and $m which depends on how the library wants to organise its collection
14:30	thd	kados: spine labels should be 852 $k $h $i $m with the addition of c. $t where $t > 1 .
14:32	thd	kados: in the current MARC 21 framework that is 952 $o $h $i $t
14:34	thd	kados: so 952 $o could be filled from the values contained in 942 $c for NPL.
14:37	thd	kados: 952 $h would come from either 082 $a, 092, 055 $a, or 852 $h depending on what was filled.
14:38	thd	kados: 055 and 852 should be tested for having the correct indicators or the correct number type when indicators were not set.
14:42	thd	kados: 952 $i would come from either 082 $b, 092, 055 $b, or 852 $i depending on what was filled except that libraries may have their own preferred cuttering scheme and the supplied item part of the number may not be unique in the library's own collection.
14:43	thd	kados: 952 $m is seldom used but would come from 852 $m only from previous records of the same library.
14:44	thd	kados: 952 $t would come from 852 $t only from previous records of the same library.
14:45	kados	hmmm
14:45	kados	that's a lot to parse
14:46	kados	it occurs to me to ask why we're trying to map 852 to 952
14:46	thd	kados: I already have most of the code for parsing that neatly written for you n Perl
14:47	kados	thd: but why do we need to parse it at all?
14:47	kados	thd: with zebra?
14:48	thd	kados: well one reason for continuing to use a Koha specific items field is that you have the data in the wrong locations in existing records
14:48	kados	thd: rebuildnonmarc can fix that up
14:48	kados	thd: given a proper framework
14:49	thd	kados: another reason would be that you do not have enough subfields for everything that you need to track in one field unless you use capital letters.
14:50	kados	capital letters--
14:50	thd	kados: 852 $A $B etc.
14:50	kados	capital letters--
14:54	kados	thd: how many items do we need to track?
14:54	kados	thd: have you considered writing a 'standaard MARC holdings' framework?
14:54	thd	kados: there are not enough lowercase letters and numbers and punctuation to hold all the predefined or reserved 852 subfields in addition to things that are not in 852 such as the things in 876-878
14:55	kados	thd: I've been thinking it might be useful to separate the holdings definitions
14:55	kados	thd: from the standard MARC framework
14:55	kados	thd: to make it easier for customization (ie, a smaller file)
14:56	thd	kados: I did write a more standard MARC holdings framework in the recommendations comment within the standard MARC 21 framework
14:56	kados	ahh
14:56	kados	thd: perhaps I'll just pull that out then
14:56	kados	thd: add the onloan stuff
14:56	kados	thd: does it include call number goodness?
14:58	thd	kados: It cannot be really standard unless you change all the code that needs one and only one MARC field defined for holdings.
14:59	thd	kados: my recommendation is just like what I had specified for 952 except that their is a greater alignment between MARC 21 subfield names and what I specified.
15:00	kados	thd: I think we could create a holdings framework
15:00	kados	thd: and link holdings records to bib records
15:01	thd	kados: oh you mean really standard
15:02	thd	kados: do you mean with multiple fields for the holdings instead of everything in one field?
15:02	kados	thd: well ... for 2.4 I mean just 852
15:02	kados	thd: for 3.0 I mean a really standard one with multiple fields for holdings
15:04	thd	kados: so 3.0 could work with the needed code changes but there is not enough room in 852 for everything needed unless you can use capital letters for the subfields
15:06	thd	kados: would you change the data type as needed to allow capital letters for the subfields for 2.4. ?
15:07	kados	thd: no
15:07	thd	kados: otherwise you would need to remove standard subfields from MARC 21 to use 852
15:07	kados	thd: too many components rely on aA equivilence
15:08	kados	thd: you mean that not all koha item fields map to standard 852?
15:08	thd	kados: the best I have then is what I recommended for 85k in the framework comments with the addition of punctuation symbols for status etc.
15:10	thd	kados: the problem is that they map to 852, 876-8 and a few others including one or two that are no part of standard MARC.
15:10	kados	hmmm
15:11	thd	kados: which is why it had filled all th letters and numbers when you looked
15:11	kados	tumer is removing that limitation for 3.0
15:11	kados	right
15:13	thd	kados: 876 has some additional medium or long term status indications, where purchased, for how much, replacement price etc.
15:14	thd	kados: those are all mapped to existing Koha items columns and are not in 852
15:17	thd	kados: 852 has a very few things which I did not map to 952 such as the address of the library
15:23	thd	kados: 952 as I defined it is basically 852 combined with 876-8 with a few minor additions and omissions. Obviously, the subfield names cannot match perfectly so I wrote the verbose librarian text to keep everything straight.
15:23	thd	by subfield names I mean $a $b etc.
15:25	kados	thd: I don't think it matters if they are all mapped to 852
15:25	kados	I'm not sure what is preventing us now from being completely compliant with MARC Holdings
15:25	kados	other than a fully defined framework
15:27	thd	kados: you have a fully defined framework but you do not have code allowing the use of more than one field for holdings.
15:27	kados	thd: code meaning the editor?
15:27	kados	thd: what code is needed?
15:29	thd	kados: wherever there is code that falls down when biblioitems.itemnumber is linked to multiple fields.
15:29	thd	s/biblioitems/items/
15:30	kados	hmmm
15:32	thd	kados: because standard MARC needs 852 and 876-8 not just 852.
15:33	thd	kados: you need to be able to distinguish between 852 $c shelving location and 876 $c cost
15:35	kados	I see
15:35	thd	kados: Koha also needs to distinguish different types of cost which MARC stores yet again in other locations acquisitions cost from replacement cost
15:36	kados	Koha does too
15:36	kados	cost and replacementcost
15:36	kados	ahh, but they aren't in 852
15:36	kados	in standard MARC
15:36	thd	kados: but Koha has them both in the items table
15:36	kados	but we aren't limited to using the items table now
15:36	thd	kados: nothing about acquisitions is in 852
15:37	kados	I see no reason to map cost to items at all
15:37	thd	kados: do you not want to know what your items cost?
15:37	kados	thd: with zebra we can still see that data
15:38	kados	thd: we are no longer limited to searching/seeing data in the koha tables
15:39	thd	kados: yes exactly but if acquisitions cost is associated with an item and not only in SQL then you need it stored in MARC which means that you need more than one field.
15:39	thd	for standard MARC 21.
15:40	kados	thd: so I suppose we should create holdings based on more than one field
15:41	thd	kados: well you were proposing to do that for 3.0 not 2.4 a few minutes ago.
15:43	kados	thd: I didn't understand the issues fully
15:44	kados	it's hard to decide what is best in this case
15:45	thd	kados: do you have a good concept of how much work that would be to fix on top of encoding?
15:47	kados	thd: no
15:47	thd	kados: you had said that tumer was going to do that for 3.0 does tumer have an idea?
15:48	kados	thd: maybe
15:48	kados	:-)
16:00	thd	kados: some problems are amazingly simple to solve if you look at them closely.
16:01	thd	kados: other problems like encoding rely on external software libraries that you cannot have time to rewrite yourself if they do not work.
16:01	thd	kados: obviously, many uses of Encode::decode are needed for encoding where that can work.
16:01	thd	kados: Debian stable may be at a special disadvantage with the out of date MySQL libraries that do not understand UTF-8.
16:02	thd	kados: it was issues like that which drove me to switch to Debian testing much to my later regret.
16:02	chris	thats what backports are for :)
16:03	thd	chris: is there a backport of MySQL for upgrading that library?
16:03	kados	thd: yes
16:03	kados	thd: backports.org
16:03	kados	hey chris
16:03	chris	morning
16:03	kados	chris: I'm ready to do a 2.3.0 release
16:04	kados	chris: and I'd like to re-tag dev_week as rel_2_4
16:04	kados	chris: if that's possible and you have time to show me how :-)
16:04	chris	its pretty easy
16:04	chris	easiest way
16:04	dewey	easiest way is to try
16:04	chris	do a clean checkout of dev_week
16:05	thd	kados: what is 2.3 as distinct from 2.4? Isa it the unstable version of 2.4?
16:05	kados	thd: yes
16:06	thd	kados chris: at the time I switched to testing I could net get the updated applications I needed from backports. I should have been more patient.
16:07	chris	then cvs tag -b branchname
16:07	kados	can I then delete dev_week tag?
16:07	chris	nope, you cant delete tags
16:07	kados	ahh
16:08	chris	you'll just be branching a new branch at that point
16:08	thd	chris: what can be deleted in CVS?
16:08	chris	nothing, thats pretty much the point of version control
16:08	chris	doesnt hurt kados
16:08	kados	k
16:08	chris	you could just tag it
16:09	chris	not branch .. but then youd still be working in the dev_week branch
16:09	chris	thats what the script taht creates the tarball does
16:09	kados	hmmm
16:09	chris	tags up the files when you make a release
16:09	chris	so at any point you can check out those files using that that tag
16:09	chris	depends on what you want to do
16:10	kados	well ...
16:10	kados	what I think I want
16:10	kados	is to have a branch that will become 2.4.0
16:10	kados	I suppose I can just use dev_week for that
16:10	chris	yep
16:10	kados	poorly named, but fine
16:11	kados	ok, I"ll just do that then
16:11	chris	thats what id do, and use the script to tag the files as 2.3.0
16:11	kados	no sense making it more complicated
16:11	kados	ok
16:11	kados	so when I run the script will it show up on savannah?
16:11	chris	if you tell it to
16:11	kados	cool
16:12	chris	hmm did i commit my fixes to the script to dev_week?
16:12	chris	lemme check
16:14	chris	looks like its all set up for savannah yep
16:14	kados	sweet
16:14	chris	misc/buildrelease
16:14	kados	I'll give it a shot then
16:16	kados	chris: any specific way I should run it?
16:16	kados	chris: and from a specific directory?
16:17	chris	id run it from the base
16:17	chris	in ./misc/buildrelease
16:17	chris	in= ie
16:17	chris	sorry prolly need
16:18	chris	perl misc/buildrelease
16:18	kados	perl misc/buildrelease
16:18	kados	yea
16:18	kados	ok
16:18	chris	it should ask you
16:18	chris	it makes a good guess
16:18	chris	so if you are in the base .. then the guess will be right
16:18	chris	(making sure you are in the dev week checkout of course :-))
16:19	kados	chris: should it be a fresh check out?
16:19	chris	doesnt have to be
16:19	chris	but is probably safest
16:19	kados	k
16:20	chris	hmm interesting synchronicity
16:20	chris	http://www.bigballofwax.co.nz/[…]07/11/eclipse-ide
16:21	kados	neat
16:21	kados	nice new website chris :-)
16:21	chris	id encourage you to take a look at eclipse sometime .. with the epic plugin
16:21	kados	will do
16:22	chris	the syntax highlighting, and completion ... plus cvs integration
16:22	chris	make it pretty cool
16:22	kados	cool
16:22	chris	ie you type " and it does the matching "
16:22	chris	{} etc
16:22	kados	sweet
16:22	chris	plus you type $
16:22	chris	and you get a list of variables you have already used
16:23	chris	so its really good writing new big chunks of code
16:23	kados	Would you like to tag the CVS repository?
16:23	kados	I'm guessing Y
16:23	chris	yes
16:23	kados	well here it comes :-)
16:23	chris	kados: that blog is my first real play with ruby on rails
16:24	kados	cool
16:24	kados	yea, I browsed some books at B&N last week about ruby
16:24	kados	ok, the tarbal is in my dir now
16:24	kados	is it also on savannah?
16:24	chris	sweet
16:24	chris	no
16:24	chris	its not that cool :-)
16:24	kados	:-)
16:24	chris	you will have to go to savannah and do the release thingy
16:25	chris	hmm lemme look
16:25	chris	i havent done a savannah one before
16:25	chris	but it should be similair to sourceforge
16:27	chris	hmm maybe not
16:27	chris	http://download.savannah.nongn[…]rg/releases/koha/
16:27	chris	i think you just have to stick it in here, somehow
16:27	kados	their searches suck
16:28	kados	ahh
16:28	chris	then write some release notes for the news bit
16:28	chris	https://savannah.nongnu.org/projects/koha/
16:28	chris	for on that page
16:28	kados	yea
16:28	kados	I've done news before
16:28	kados	that's pretty simple
16:30	chris	https://savannah.gnu.org/faq/?[…]o_I_add_files.txt
16:30	chris	so we need a gpg key
16:30	kados	k, I can handle that
16:37	chris	:)
17:27	chris	is half an hour up yet? :-)
17:31	kados	yea, it's uploaded
17:31	kados	and I did a news item on it
17:31	chris	excellent
17:31	kados	http://download.savannah.nongn[…]rg/releases/koha/
17:31	russ	news item ? where?
17:32	chris	on savannah
17:32	kados	russ: probably not newsworthy in a general sense
17:32	chris	http://koha.org/download/ <-- should we write something here
17:32	chris	to make sure ppl know that 2.3.0 is unstable
17:32	kados	definitely
17:32	kados	but I plan on doing 2.3.1 in a day or two
17:32	kados	so I'm not sure just how much news we want to generate :-)
17:32	chris	otherwise they will click on the 2.2.5 link
17:33	kados	yea ... 2.2.5 is still good I think
17:33	chris	and see 2.3.0 sitting there
17:33	chris	and grab it cos its newer
17:33	kados	can't we just link directly to the download ?
17:33	russ	two minutes
17:33	kados	ie, not the file list?
17:33	chris	probably the best bet
17:34	kados	too bad you can't put text in the list
17:34	kados	names and such like on sourceforge
17:34	chris	yep
17:35	chris	but then you also have to click 54 times to actually download the file
17:35	kados	yea :-)
17:35	chris	i think linking to the file and having good text around the link is the way to go
17:35	kados	yea I agree
17:35	russ	kados - can we move those wiki notes into the en section?
17:36	kados	russ: sure
17:36	chris	ill have a go with 2.3 after work, on my koha at home
17:36	kados	russ: be my guest :-)
17:36	kados	chris: I haven't touched the installation scripts yet
17:36	kados	chris: but I can send you a tarbal of my migration stuff
17:36	chris	but if i follow your docs on the wiki it should work eh?
17:36	kados	yea
17:36	chris	ill have a go by hand first
17:36	kados	cool
17:37	chris	so i have an understanding, then ill try out the automatic way
17:37	kados	yep
17:37	kados	it's no walk in the park :-)
17:37	chris	:)
17:37	kados	and there are some hard-coded things still
17:37	kados	that need to be moved to the config file
17:37	chris	yep i figure ppl will be asking questions sooner or later
17:37	kados	yea
17:40	kados	nothing like a release to make you feel like you've acomplished something in a day :-)
17:40	kados	even if I know it's not a stable release :-)
17:41	chris	:-)
17:53	thd	kados: have you solved the problem of migrating a rel_2_2 installation with UTF-8 data?
17:53	kados	thd: not yet
17:53	kados	thd: it's more a 'how to migrate from mysql 4.0 to 4.1' problem
17:53	kados	thd: for NPL it's hopeless
17:54	kados	thd: because they had marc-8 data that was interpreted as latin1 data and I think it's been completely mangled in some cases :-)
17:54	thd	kados: so you did not worry about the issue for NPL because NPL had relatively little MARC 8 data?
17:54	kados	right
17:55	thd	kados: If you had a script that searched for encoding problems you could concentrate on upgrading those records.
17:57	kados	thd: I've got one
17:57	kados	thd: it just needs to be updated a bit to deal with these specific probs
17:58	thd	kados: really, would it also work for Afognak double encoding, and other similar problems that you have seen?
17:58	kados	yes, given enough time to troubleshoot
18:00	thd	kados: I think tumer has many records that he needs to find before his records including the problem ones become canonical for Turkey
18:01	thd	kados: nothing like a new national authority file complete with encoding problems :)
18:02	kados	:-)
18:03	kados	that wouldn't bode well for Koha :-)
18:04	thd	kados: no, having the national library of Turkey using Koha should be all to the good as long as the records can be shared without problems.
18:06	russ	philip you there?
18:06	thd	russ: I do not see philip logged in?
18:07	russ	oops wrong channel
18:07	russ	:-)
18:16	thd	kados: looking at http://wiki.koha.org/doku.php?[…]tion_of_materials. Does NPL not have full call numbers in 952 $k?
18:16	kados	they do
18:16	thd	kados: also, I have noticed that you have been using the old WIKI namespace
18:17	thd	kados: the new namespace from devweek has en for English
18:18	kados	yea, russ is gonna fix that for me :-)
18:18	thd	kados: That way Pascal or whomever can make a French translation under /fr/ instead of /en/
18:18	kados	yep
18:19	thd	kados: russ had been cleaning the whole thing so that you would not have made the mistake in the first place.
18:19	kados	:-)
18:20	thd	kados: pierrick had also been working on that but he is no longer participating :(
18:22	thd	kados: if using multiple holdings fields is discovered to be too much work for 2.4 do you know of any problems using punctuation for subfields all in one items field?
18:25	kados	thd: looking at the framework right now
18:25	kados	thd: got some questions
18:25	kados	INSERT INTO `marc_subfield_structure` VALUES ('952', '2', 'Source of classification or shelving scheme (similar to 852 $2)', 'Source of classification or shelving scheme', 0, 0, '', 10, '', '', '', NULL, 0, '', '', '');
18:25	kados	is that really necessary?
18:26	kados	some of these seem like they won't ever be used
18:26	thd	kados: it is necessary for designating the call number for fiction in the case of NPL.
18:26	kados	how so?
18:27	kados	(btw: items.wthdrawn not items.withdrawn ... a minor misspelling in the framework due to a poorly named column name)
18:28	thd	kados: indicators are used to specify if the call number is LCC, DDC, NLM. or SUDOC, not that Koha is providing access to the 952 indicator currently.
18:29	thd	kados: anything that uses some other classification is supposed to specify the classification in $2
18:29	kados	specify what about it?
18:29	kados	the name of it?
18:29	thd	kados: the classification code and version number
18:30	thd	kados: kados has not seen nearly enough correctly encoded records
18:30	kados	:-)
18:31	thd	kados: NPL's private scheme may not have a standard classification code but they can pretend if there is not one for generic local scheme
18:31	kados	they aren't really up to doing that :-)
18:32	kados	and it doesn't really serve any purpose
18:32	thd	kados: no of course not but you have the system do it for them
18:32	thd	kados: it allows classification searches to use information in the records to do an intelligent search
18:33	kados	ok so I'll call it NPL v1
18:33	kados	:-)
18:33	kados	thd: do you know NPL's MARC ORG code?
18:34	thd	kados: no but classification codes in MARC 21 are all lower case
18:34	kados	thd: it's 'ONe' :-)
18:34	thd	kados: do you not have the code list bookmarked
18:34	kados	thd: 'ONe' :-)
18:34	kados	thd: that's it ... pretty cool, eh?
18:35	thd	kados: does LC do vanity classification codes for a special fee? :)
18:35	thd	s/classification/organisation/
18:39	thd	kados: the only thing that I remember having put in 952 that is not most likely to be useful is article '952', 's', 'Copyright article-fee code (similar to 018 $a, 852 $s)'
18:40	thd	kados: although WIPO might care very much about that
18:41	thd	kados: do you think they observe the rules they establish for everyone internally?
18:41	kados	thd: so ... question
18:41	thd	yes
18:42	kados	if we created a full MARC authorities file
18:42	kados	with MARC holdings
18:42	kados	852, etc.
18:42	kados	how many fields would need to contain itemnumber?
18:43	kados	and could we acomplish the same thing with a link?
18:43	kados	ie, link the itemnumber field to a field elsewhere in the record
18:43	thd	kados: changing the code to use a link would be more MARC like but may not be the most efficient
18:44	kados	with zebra I don't think it matters
18:44	kados	in terms of searching we have no problem
18:44	kados	of course, display is not a problem either
18:44	thd	kados: do you mean full MARC holdings file not full MARC authorities file
18:44	thd	?
18:45	kados	correct
18:46	thd	kados: quite a few fields
18:50	thd	kados: some are more important than others
19:04	thd	kados: perhaps about 25 fields
19:04	thd	kados ...
19:04	thd	541 - IMMEDIATE SOURCE OF ACQUISITION NOTE (R)
19:04	thd	561 - OWNERSHIP AND CUSTODIAL HISTORY (R)
19:04	thd	562 - COPY AND VERSION IDENTIFICATION NOTE (R)
19:04	thd	563 - BINDING INFORMATION (R)
19:04	thd	583 - ACTION NOTE (R)
19:04	thd	841 - HOLDINGS CODED DATA VALUES (NR)
19:04	thd	842 - TEXTUAL PHYSICAL FORM DESIGNATOR (NR)
19:04	thd	843 - REPRODUCTION NOTE (R)
19:04	thd	844 - NAME OF UNIT (NR)
19:04	thd	845 - TERMS GOVERNING USE AND REPRODUCTION NOTE (R)
19:04	thd	850 - HOLDING INSTITUTION (R)
19:04	thd	852 - LOCATION (R)
19:04	thd	853 - CAPTIONS AND PATTERN--BASIC BIBLIOGRAPHIC UNIT (R)
19:05	thd	854 - CAPTIONS AND PATTERN--SUPPLEMENTARY MATERIAL (R)
19:05	thd	855 - CAPTIONS AND PATTERN--INDEXES (R)
19:05	thd	856 - ELECTRONIC LOCATION AND ACCESS (R)
19:05	thd	863 - ENUMERATION AND CHRONOLOGY--BASIC BIBLIOGRAPHIC UNIT (R)
19:05	thd	864 - ENUMERATION AND CHRONOLOGY--SUPPLEMENTARY MATERIAL (R)
19:05	thd	865 - ENUMERATION AND CHRONOLOGY--INDEXES (R)
19:05	thd	866 - TEXTUAL HOLDINGS--BASIC BIBLIOGRAPHIC UNIT (R)
19:05	thd	867 - TEXTUAL HOLDINGS--SUPPLEMENTARY MATERIAL (R)
19:05	thd	868 - TEXTUAL HOLDINGS--INDEXES (R)
19:05	thd	876 - ITEM INFORMATION--BASIC BIBLIOGRAPHIC UNIT (R)
19:05	thd	877 - ITEM INFORMATION--SUPPLEMENTARY MATERIAL (R)
19:05	thd	878 - ITEM INFORMATION--INDEXES (R)
19:06	thd	kados: do you think that would be enough?
19:07	kados	enough for what?
19:08	thd	linking to items.itemnumber ?
19:09	kados	hmmm
19:09	kados	we need to link all of those?
19:13	thd	kados: the only scary ones are 853-855 and 863-868
19:14	thd	kados: 866-868 are not scary for humans, they are merely scary for computers.
19:34	thd	kados: you would not need to link specifically to 841 but the shiny forest in SQL would be much more efficient to parse for the scary cases.
19:36	thd	kados: I can present a simplified case for the only users that you can obtain now in any case
19:39	thd	kados: acquisitions would need 541, 583, and 876 that is just 3 fields with 877-8 for very special cases
19:43	thd	kados: cataloguing would need 852 or 856, and sometimes 863-868
19:44	thd	tumer: is that really you?
19:44	tumer	yep awake and zebraing
19:45	thd	tumer: kados had said that you were intending to eliminate the dependence that Koha had on using one and only one MARC field for holdings.
19:45	tumer	i already did
19:45	thd	for 3.0
19:46	tumer	yep
19:46	thd	tumer: really?
19:46	dewey	really are quite different
19:46	thd	tumer: and it works?
19:46	tumer	yes i have separate marc record for holdings niw
19:47	tumer	thd: which field is best to hold the biblionumber in holdings record?
19:47	thd	tumer: and do these separate holdings records use more than one of the holdings fields?
19:48	tumer	they have their own 001 005 007 852- to 999 all
19:48	tumer	a complete marc record
19:49	tumer	any field to put anything in
19:49	tumer	no restrictions
19:49	tumer	well within reason
19:50	tumer	if kados spares same space i will commit the whole lot to the cvs
19:50	thd	tumer: well if you are doing something non-standard for Koha I like some field with a letter in it like 90k but I realise that may require some code change
19:51	tumer	i used to now i stick with standard LC nummbers
19:52	tumer	separate holdings,bibblios and authorities records you do not require fields with letters there is abundance of fields
19:52	thd	tumer: well the standard place for the link number is 004 - CONTROL NUMBER FOR RELATED BIBLIOGRAPHIC RECORD
19:53	tumer	great thanks
19:53	thd	tumer: that would be the place for storing the 001 from the bibliographic record
19:53	tumer	i hope to get a full koha intranet working on this model by weekend
19:54	thd	tumer: are you no longer using SQL code for tracking the items?
19:54	tumer	i already have biblionumber in 001 in biblios itemnumber 001 in holdings
19:55	tumer	there is nothing in sql any more only blobs of marc records
19:55	tumer	except issues and borrowers and subscript ofcourse
19:55	thd	tumer: do you mean you are creating a separate holdings record for each item even if you have duplicate items for the same biblio?
19:57	tumer	yes because even a second copy is a different entity which has copy 2 in it
19:57	thd	tumer: do you have multiple copies of the same biblio in your library?
19:58	tumer	yes but all call numbers end with c1 c2 etc so they are different and unique
19:58	thd	tumer: you do not attempt to make use of the repeatable fields so that you have for example 852 $t1 and 852 $t for the first and second copies?
19:59	thd	tumer: you do not attempt to make use of the repeatable fields so that you have for example 852 $t1 and 852 $t2 for the first and second copies?
19:59	tumer	its easeier to manage them this way
19:59	thd	tumer: yes that is very evident
19:59	tumer	easier on indexing releases the starin on zebra
20:00	tumer	s/starin/strain
20:00	thd	tumer: and there I was thinking about how to do it the most difficult way
20:00	thd	kados: are you there?
20:04	thd	tumer: that is brilliant to have done it the way that is both easiest and most efficient
20:04	tumer	well thanks i dont know whether correct but thats how i designed it
20:05	thd	tumer: that is perfectly in conformance with the standard which is the only important obstacle
20:06	thd	tumer: I would like to persuade you to proceed with authorities in a standards compliant manner
20:07	tumer	thd: i have to redesign authorities, the way they are now is not working
20:07	thd	tumer: It may be a little painful in the short run but you will be much happier if they interoperate well with other systems
20:08	thd	tumer: what about them is not working?
20:08	tumer	thats my intention
20:08	thd	tumer: specifically is not working?
20:08	thd	tumer: what specifically is not working about authorities?
20:09	tumer	all linking of authorities etc i have to study it a bit more
20:09	tumer	now we have thd producing some things i did some other all non complimenting each other
20:10	tumer	the editor does not work with them either
20:10	thd	tumer: the editor code was never fixed for authorities
20:11	tumer	i know
20:11	thd	tumer: the authorities editor code needs to be copied from the bibliographic editor code
20:12	tumer	yes but try to add a authors name to a bibliographic record from authorities. Its now a mess
20:12	thd	tumer: do you mean with encoding?
20:12	tumer	no
20:13	tumer	you try filling 700 it fills 690
20:13	tumer	all the java script messing it all up
20:13	thd	tumer: ooooh that is bad
20:13	tumer	i reported a bug but in vain
20:15	thd	tumer: unfortunately, the only Koha libraries actually using authorities merely use the script designed to fill existing values in the bibliographic records into the database.
20:16	tumer	its not obvious untill you use subfield cloning -- the bug i mean
20:17	tumer	whats the use of creating authorities that you do not comply with?
20:17	thd	tumer: those libraries are all using UNIMARC
20:17	thd	tumer: they comply if they comply by using already compliant records.
20:19	tumer	well its not good for me
20:19	tumer	we are creating 90% of our records
20:19	thd	tumer: subdivided subjects will seem to need repeated $9s for most free floating subdivision $z $x $y $v
20:20	tumer	thd i am so far away from the subject, also they have to be freely ordered
20:21	thd	tumer: I suspect that having repeated $9 in the same way that BnF has repeated $3 for subject subdivisions is the only way to have a manageable authority file.
20:22	tumer	thd:ok
20:22	thd	tumer: repeatable $9 should allow for freely ordering the subfields
20:23	thd	tumer: your library is translating LC authority files, is it not?
20:23	tumer	but i dont even know whether we actually fill $x $y $z from differnt authorities or one record that has all fields in it
20:24	tumer	currently its hybrid. LC subject headings beung translated but no actual records
20:24	thd	tumer: decades ago LC stopped creating authority records with all the subdivisions in the authorised form
20:24	thd	for subject
20:25	tumer	BnF do dont they?
20:26	thd	tumer: LC subject authority records only comprise about 13% of subject authority strings
20:26	thd	13% of subject bibliographic strings I mean
20:27	thd	tumer: BnF bends the UNMARC rule that specifies $3 as not repeatable for subject subdivisions
20:27	tumer	thats a relief, I thought i was going to have them all like that in millions
20:29	thd	tumer: subject authority files would be millions of authorised forms if they had to be complete subject strings with all the subdivisions in one record
20:30	thd	tumer: LC only has circa 125,000 authorised subject headings which can be used to build all the others.
20:31	tumer	i am in more clear now
20:31	chris	is there authorities for other languages thd?
20:31	thd	yes it is very easy
20:32	thd	chris: although tumer says there are none yet for Turkish so he must create them
20:32	chris	right
20:32	chris	im betting there are none for maori too
20:33	chris	which would be needed if the total emmersion schools were to use them
20:33	chris	immersion even
20:34	chris	maybe thats something the future foundation could do, help hold/build authority records for other languages
20:34	thd	tumer chris: preserve the original LC record you may be translating into Turkish or Maori by giving the original heading in a 7XX linking field in the authority record
20:35	chris	ahh good idea
20:36	tumer	i give it in 750
20:36	tumer	i even linked them so searching in one language brings the other as well
20:37	tumer	but thats not standart i know
20:38	thd	tumer: that is standard
20:39	thd	chris: national libraries usually do that or some really large university library
20:39	thd	chris: although most counties lack authority records
20:41	thd	chris: authority records are found in Latin America even if under used but they are largely absent even in some relatively rich countries like Italy
20:42	thd	tumer: you have done it precisely correctly as long as you are not still using $6 to link outside the record
20:43	thd	$6 is for linking fields within the same record
20:44	tumer	which subfield should it be 8?
20:44	thd	tumer: $8 is also for linking within the record
20:44	tumer	or a specific field?
20:45	tumer	as i say i know zilch about authorities
20:46	thd	tumer: $0 is used for linking from the authority record to the original record although you should also preserve the original record control number in 035
20:47	thd	tumer: an example is 750 #0$aSummer resorts$0(DLC)sh#85130430#$wna
20:48	thd	tumer: http://www.loc.gov/marc/author[…]link.html#mrca750
20:49	tumer	i will look into it
20:49	tumer	i have to check this zebra which is refusing to start
20:50	thd	tumer: be careful not to be kicked from behind by your zebra
21:41	kados	Burgundavia: still waiting for that email :-)
21:41	kados	Burgundavia: I'll hold you to your purpose :-)
21:41	Burgundavia	kados: it is stuck in the loop of no time
21:42	kados	:-)
21:42	Burgundavia	I am about to run out to the local LUG meeting
21:42	kados	I hear you
21:42	kados	cool
21:42	Burgundavia	it is mostly written, just missing the mockup
21:42	kados	excellent
21:42	kados	can't wait to read it
21:42	kados	I'll be working pretty much all week on 2.3.x
21:43	Burgundavia	I just saw that annoucement
21:43	kados	so now would be a really great time to have the ideas/mockup
21:43	Burgundavia	what is policy on posting stuff to koha.org?
21:43	kados	posting what?
21:43	kados	announcements?
21:43	Burgundavia	yep
21:43	Burgundavia	the new development announcement
21:44	kados	do you mean to savannah? or koha-devel? or on the website koha.org?
21:44	kados	well I'm the release manager
21:44	kados	so I can basically do anything :-)
21:44	kados	muhahaha
21:44	kados	if you have something you need announced you can generally send it to russ or someone
21:44	Burgundavia	the website, in the little news box which has nothing in it
21:44	Burgundavia	nothing new, that is
21:44	kados	right
21:45	chris	yep, ppl have to tell us news, then we'll put it there
21:45	kados	well basically if someone has time to write something they submit it to russ
21:45	kados	or someone
21:45	kados	the project is really loosely defined
21:45	kados	there are not that many core developers
21:45	kados	so we basically know what's going on :-)
21:46	Burgundavia	indeed
21:46	kados	and news and stuff is loose as well
21:46	kados	plus, koha.org is really just the project page ...
21:46	chris	pretty much anything thats interesting or new, and someone tells us about can go up
21:46	Burgundavia	the start of a major new development cycle is pretty good news, I think
21:46	kados	the liblime team has enough trouble keeping our marketing website up to date :-)
21:46	Burgundavia	gets people all excited, reminds them the project is not dead
21:46	kados	yep
21:47	kados	but it's really the tail end of the dev cycle
21:47	chris	yep
21:47	kados	we've been working on this release for about two years now :-)
21:47	Burgundavia	2.3?
21:47	kados	yea
21:47	Burgundavia	see, that is not how (as an outsider), interpreted your email
21:47	kados	right
21:47	chris	work leading up to 2.4 and 3.0 has been going on since 2.2.0 was released
21:48	kados	yea
21:49	chris	when 2.4 comes out there will be tons of fanfare on koha.org .. and on the main koha-devel list
21:49	Burgundavia	yep
21:49	kados	so 2.3.0 is basically where we've said "ok, stuff is actually working ... now lets test and bugfix, etc. and in a month or so, we'll have a stable product"
21:49	chris	but for now 2.3 is so bleeding edge we dont really want too many people trying it
21:49	chris	cos their will be blood everywhere :-)
21:50	Burgundavia	no, but you do want people to be excited about it existing
21:50	chris	there even
21:50	kados	yea, bleeding edge is key
21:50	chris	its hard to get them excited without them asking where is it, can i have it
21:50	kados	Burgundavia: not really, because the core developers don't even know how to install it :-)
21:50	Burgundavia	heh
21:50	chris	its a tricky one
21:50	kados	Burgundavia: only tumer and I have got it going so far :-)
21:50	chris	maybe in a week or so
21:50	kados	yea
21:50	Burgundavia	ok
21:50	chris	when we are at 2.3.2 ish
21:50	kados	yep
21:51	kados	I'll do 2.3.1 tomorrow afternoon
21:51	Burgundavia	just don't want to miss an opportunity to get us talked about
21:51	kados	already spotted some probs
21:51	kados	also I want to get some public demos going
21:51	kados	so folks can try it out with real data
21:52	kados	Burgundavia: well, when you get to that email, I'm all eyes :-)
21:52	Burgundavia	ok
21:52	chris	yep its true, more publicity would be good, but id hate to have ppl see it, and then try it out and say it sucks because they cant install it :-)
21:52	Burgundavia	too many projects, so little time
21:53	kados	heh
21:53	chris	but there are few new libraries running koha
21:53	Burgundavia	somewhere in there I also need to finish that edubuntu case study
21:53	chris	which we should do some news about
21:54	Burgundavia	chris: just to introduce myself, I am Corey Burger. I work for Userful in the day and volunteer with Ubuntu/Edubuntu at night
21:54	chris	ahhh ive seen your name around :)
21:54	Burgundavia	nothing bad I hope :)
21:54	chris	im Chris Cormack, i work for Katipo Communications in the day .. and some of the night too, and Koha all over the place
21:54	Burgundavia	cool
21:55	chris	trying to remember where ive seen it
21:55	Burgundavia	I am co-author on the Official Ubuntu Book?
21:55	chris	ahh
21:55	chris	that'll be it
21:55	chris	planet ubuntu
21:56	Burgundavia	ubuntu is everywhere
21:56	chris	yep
21:57	chris	i met mark at linuxconf.au this year
21:57	Burgundavia	ah yes
21:57	Burgundavia	I was supposed to be at the Paris development conference, but it clashed with ALA in New Orleans, sadly
21:57	chris	ahh yes
21:57	chris	how was ALA?
21:57	Burgundavia	big
21:57	chris	it sounded big
21:58	Burgundavia	a big, giant Windows world, with a few beacons of hope: our booth and Index Data's
21:58	chris	:-)
21:58	Burgundavia	our booth = Userful's
21:58	chris	right
21:58	chris	i think liblime was sharing with index data?
21:59	Burgundavia	had a laugh at SirsiDynix claiming to built on "open standards"
21:59	Burgundavia	yep
21:59	chris	yeah
21:59	chris	interestingly enough sirsidynix are taking up the first 2 hours of the first day of lianza (the nz equiv of the ala conference) here this year
21:59	Burgundavia	interesting. What with?
22:00	chris	the keynote address
22:00	Burgundavia	ah
22:00	chris	and then chairing a panel
22:00	Burgundavia	"how libraries can spend more money on us"
22:00	russ	http://www.lianza.org.nz/event[…]06/programme.html
22:01	chris	im picking that will be the angle they will take
22:01	Burgundavia	library I know is migrating Dynix Classic --> Sirsi Unicorn. Loosing 900k in fines, due to Unicorn not understanding that fines should persist after the book/thing has been removed from the collection
22:01	chris	roll up roll up, get locked in to our solution, ull never escape
22:01	chris	yikes thats a lot to lose
22:02	Burgundavia	not too mention the headache of the migration itself
22:02	russ	oooh that would cripple a public library in our region
22:02	chris	2 libraries in nz have migrated from old dynix to koha in the last year
22:02	Burgundavia	they said that getting support for classic has basically been impossible since the merger
22:02	chris	yeah
22:03	chris	the ones who migrated really had no choice, migrate or upgrade
22:03	Burgundavia	it is important to publicize any libraries that do migrate
22:03	Burgundavia	the koha website gives the impression that it is a one library trick
22:03	chris	yep, we are working with the libraries to get them to do some publicity
22:03	Burgundavia	well, two, if you county athen county
22:04	Burgundavia	count, that is
22:04	chris	yeah, its more like a 200 library trick
22:04	chris	but its hard to show that in a nice way
22:04	chris	we used to have a map
22:04	chris	on the old website ... and we need something like that again
22:04	Burgundavia	a little news piece on the main website
22:04	Burgundavia	X library has migrated from Y to Koha
22:05	chris	http://old.koha.org/about/map/index.html <-- old map
22:05	Burgundavia	I have noticed libraries hate being pathfinders on technology
22:05	russ	i guess we have steered clear of that in the past
22:05	chris	we try to get the libraries to give us the permission for that
22:05	Burgundavia	ok, that is an amazing map
22:06	chris	thats old too, theres lots more than that now
22:06	Burgundavia	who does Koha marketing?
22:06	chris	no one :)
22:06	Burgundavia	hmm
22:06	chris	katipo markets its services ... which semi markets koha
22:06	chris	liblime does the same
22:07	chris	paul in france does the same
22:07	Burgundavia	but you have no Koha case studies, in printed form
22:07	Burgundavia	nor a product guide
22:07	chris	in printed, nope
22:07	chris	we have lots in html
22:07	chris	in the wiki
22:07	chris	and at www.kohadocs.org
22:07	chris	hehe
22:07	russ	lol
22:08	chris	have you looked at wiki.koha.org ... and www.kohadocs.org ?
22:08	Burgundavia	yep I have dug through
22:08	chris	thats pretty much what we have
22:08	chris	we'd love printed material
22:08	Burgundavia	most oss projects are bad on this
22:08	chris	but suffer from the same only so many hours in the day proble
22:08	chris	m
22:09	Burgundavia	hence the edubuntu case study
22:09	chris	right
22:09	Burgundavia	http://ubuntu.ca/Edubuntu-casestudy.png
22:09	chris	we have some printed material
22:10	chris	some brochure stuff etc, that we (katipo) have used at ALIA online, and lianza
22:10	Burgundavia	ah
22:10	chris	and i think liblime has some
22:10	Burgundavia	let me get this edubuntu case study out the door and the OPAC UI critique and I will see if I can take stock of what is available
22:11	chris	but there is no really general koha ones
22:11	chris	excellent :)
22:11	Burgundavia	ideally you wanat something you can co-brand: liblime/katipo/koha-fr and koha
22:11	chris	yeah that would be ideal
22:12	Burgundavia	does koha have a palette? (aside from purple)
22:12	chris	koha the project?
22:12	Burgundavia	yep
22:12	chris	hmm ill defer to russ on that one
22:13	chris	you dont want me doing anything to do with design :-)
22:13	Burgundavia	heh
22:14	russ	ahh not really
22:14	Burgundavia	ok
22:14	Burgundavia	the blue and green on the website are not bad
22:14	Burgundavia	anyway, I have to run
22:14	russ	it is mainly the green and the purple
22:14	Burgundavia	I will get my work IRC client on this channel as well
22:14	russ	catch you later
22:15	chris	cya later
23:16	thd	kados: are you still awake?
02:39	ToinS	hello world
02:40	chris	hi toins
02:40	ToinS	hello chris
03:43	chris	toins: with the work you are doing with C4::Koha ... are you planning to use get_itemtypeinfos_of instead of GetItemTypes ?
03:44	chris	or keep using GetItemTypes ?
03:44	ToinS	i've not really worked on koha.pm
03:45	ToinS	i've just meet getitemtypes in a script and renamed it
03:45	ToinS	so i don't knw
03:45	chris	ahh ok
03:45	btoumi	hi all
03:45	ToinS	hi btoumi
03:45	btoumi	hi toins:
03:45	btoumi	hi chris
03:46	chris	hi btoumi
03:46	chris	toins: i just noticed the FIXME but im not sure who wrote that
03:46	btoumi	chris i have a question for u
03:47	btoumi	can I?
03:47	chris	i think they do different things, so i think keep using GetItemTypes for now
03:47	chris	sure btoumi
03:50	ToinS	chris: ok !
03:53	btoumi	i work on fines
03:54	btoumi	but i can't see all value for accounttype can u help me?
03:55	chris	right
03:55	chris	there is Pay
03:56	btoumi	i find fu o f m lr lc status but not all
03:56	chris	as well as those ones
03:57	chris	L, F, Rep, F, FU, A, Pay
03:57	chris	are the ones i know
03:57	btoumi	ok
03:58	chris	plus the
03:58	chris	CS, CB, CW, CF and CL
03:58	btoumi	can u configure them or not?
03:59	chris	no, they arent configurable
03:59	chris	you can add more if you need more tho
03:59	btoumi	now i learn more about the fines functionnement
04:00	chris	all the C ones are credits
04:01	btoumi	because for my library we must manage fines
04:01	chris	right
04:02	btoumi	but now the calcul of fines is do by fines2.pl ?
04:02	chris	yes
04:02	chris	and Accounts2.pm
04:03	btoumi	i find a lot of problem with this file
04:03	chris	and Circulation/Fines.pm
04:03	chris	fines2.pl ?
04:03	btoumi	yes
04:03	btoumi	it uses a table whio not exist
04:04	chris	ahh 2 seconds
04:04	chris	ill update it
04:04	btoumi	ok thanks
04:04	chris	hm which table does it use?
04:06	chris	maybe im looking at the wrong file
04:06	chris	is it misc/fines2.pl ?
04:08	btoumi	just one minute i think i make confuse with file i think i must sleep now
04:08	btoumi	;=)
04:08	chris	:-)
04:09	btoumi	i confirm
04:10	chris	in head?
04:10	btoumi	not fines2.pl but fines.pm because in fines2.pl u use fines.pm
04:10	chris	ah ha
04:10	btoumi	yes i work only in head
04:11	btoumi	line 161 in fines.pm
04:12	chris	ahh yes i see it
04:12	chris	categoryitem
04:12	chris	that should be issuingrules
04:12	btoumi	yes
04:12	chris	2 mintues ill fix that sql up
04:13	btoumi	i do it in my repository but i wasn't sure
04:13	btoumi	that why i ask u
04:14	chris	back before issuingrules existed, the issuing rules where in a table called categoryitem (borrower categories, plus itemtypes = categoryitem)
04:14	chris	but issuingrules replaced that
04:15	chris	there we go
04:16	chris	damn
04:16	chris	i edited that file just a while ago and didnt even notice that :-)
04:18	btoumi	i'm happy to help u ;=)
04:19	chris	:)
04:21	btoumi	perahps i have another question for u but not for the moment
04:21	btoumi	ty
04:22	chris	no problem
04:25	btoumi	another probleme with fines
04:25	chris	yep?
04:26	chris	hi arnaud
04:26	alaurin	hi, how are u .????
04:26	btoumi	calcfines was not called in export
04:27	chris	ahh, it should be, can you fix that?
04:27	btoumi	yes i do it
04:27	chris	arnaud: im good thanks, and you?
04:32	alaurin	fine, i'm near of my holidays, so , I feel good !!!!!
04:33	ToinS	hello arnaud
04:36	alaurin	hi Toins
04:37	btoumi	chris: ok for fines.pm i have some probleme with my repository because call to calfine is ok in fines.pm
04:37	chris	ahh ok
04:37	btoumi	need one week for sleep
04:38	btoumi	;=)
04:39	chris	hehe
04:40	paul	hello world
04:40	chris	hi paul
04:41	chris	oh no
04:41	paul	so works from Antoine home !!!
04:42	chris	up near the Notre Dame de la Guarde ?
04:42	paul	yep
04:42	chris	there must be a great view for Antoine's home
04:42	paul	my new home being close to ND
04:42	chris	ohh cool
04:43	paul	just on the other side of the hill (not sure of the word hill)
04:43	chris	i think hill is right ... its a big hill though :)
04:49	chris	will you get cable or adsl in your office paul?
04:49	paul	adsl
04:49	paul	something like 4MB dn / 800KB up
04:49	paul	s/B/b/
04:49	chris	cool, thats not bad at all
04:50	chris	i have 2mb dn, 2mb up
04:50	chris	on a cable modem
04:50	paul	I had 6Mb previously, but my main concern is the up bandwidth
04:50	chris	yeah
04:51	paul	to investigage zoom & unimarc things...
04:51	chris	excellent
04:51	chris	did you see there is a 2.3.0 ?
04:51	paul	ToinS will start cleaning acquistion module this afternoon
04:51	paul	yep. it's a dev_week snapshot I bet ?
04:51	chris	yep
04:52	chris	cool, we merged all our fixes in the acquistion module in head a few weeks ago
04:52	chris	so it probably needs some cleaning :)
04:52	paul	did you look at ToinS code cleaning on suggestions ?
04:53	chris	yes its really good
07:43	kados	hi paul
07:44	paul	hello kados
07:44	paul	good morning to you
07:44	paul	(working from ToinS home, because I still don't have any connexion at my new home :( )
07:45	paul	do you have a few seconds for some questions ?
07:45	kados	sure
07:45	paul	http://wiki.koha.org/doku.php?[…]ingzebraplugin226
07:45	ToinS	hi kados
07:45	kados	hey ToinS
07:45	paul	I'm not sure to understand :
07:45	paul	- run rebuild-nonmarc from rel_2_2 if your framework has changed
07:45	paul	#
07:45	paul	run missing090field.pl (from dev-week)
07:45	paul	#
07:45	paul	run biblio_framework.sql from within the mysql monitor (from dev-week)
07:46	paul	#
07:46	paul	run phrase_log.sql from within the mysql monitor (from dev-week)
07:46	paul	5.
07:46	paul	export your MARC records
07:46	paul	6.
07:46	paul	run them through a preprocess routine to convert to utf-8
07:46	paul	7.
07:46	paul	double-check again for missing 090 fields (very critical)
07:46	paul	rebuild non marc is OK I think
07:46	kados	except it should say from dev_week :-)
07:47	paul	i have completed this page & will continue as well, when I encounter something I don't understand well
07:47	kados	have you done those steps? or don't understand how?
07:47	paul	I haven't done anything yet
07:48	paul	I just want to know what does those scripts ;-)
07:48	kados	http://cvs.savannah.nongnu.org[…]_with_tag=R_2-3-0
07:48	kados	they are in the zebraplugin dir
07:48	paul	(+ learn what you mean by "preprocess routine to convert to utf-8 ? don't you have one already ?)
07:49	kados	well ... preprocess includes many things
07:49	kados	every previous version of Koha produced improper MARC records
07:49	kados	so all of those marc records must be preprocessed
07:49	kados	change encoding is just one step
07:49	dewey	kados: that doesn't look right
07:50	kados	other steps are to add a new leader
07:50	paul	dewey : do back to bed please. It's time to sleep in new zealand...
07:50	dewey	paul: excuse me?
07:50	kados	add all of the fixed fields
07:50	kados	(used for searching by date and format/content/audience)
07:51	kados	plus ...
07:51	kados	koha 2.x doesn't export items properly
07:51	kados	so it's necessary to query items table to get the right values for items fields
07:51	kados	all of that is covered under preprocessing
07:52	ToinS	have someone ever seen this error : HTML::Template->param() : You gave me an odd number of parameters to param()! ???
07:52	paul	yes ToinS : you have a 'XX' => $variable and $variable is probably empty.
07:52	paul	try 'XX' => "$variable"
07:52	ToinS	ok
07:52	paul	kados: is there a script that does all of this for you somewhere ?
07:53	kados	paul: it's not completely written yet
07:53	kados	paul: and every client has different needs
07:53	paul	yes, but that would be nice to have a starter ;-)
07:53	kados	right ... but MARC21 is quite different than UNIMARC
07:53	kados	I'm afraid it won't be useful to you
07:54	paul	because I can accept that " so it's necessary to query items table to get the right values for items fields" but I don't know what it means exactly & what to code to fix this !!!
07:54	kados	well ...
07:54	kados	what it means is that somewhere in rel_2_2 code
07:55	kados	when items are updated, MARC isn't
07:55	paul	on transfers, right ?
07:55	kados	also, somewhere in rel_2_2 code, 090 fields are being deleted or are never added for some records
07:55	kados	NPL has several thousand!! missing 090 fields
07:56	paul	wow ! strange, I never had this problem in France !
07:56	kados	have you run missing090.pl before? :-)
07:56	kados	you may be surprised :-)
07:56	paul	I'll run & let you know if there is the problem
07:57	kados	paul: also, we spoke briefly about acquisitions bugs before
07:57	kados	http://wiki.liblime.com/doku.php?id=koha226bugs
07:57	kados	a client has explained the bugs
07:57	kados	and I have submitted bug reports for each confirmed bug
07:57	kados	(welcome back :-))
07:58	paul	ToinS will start code cleaning on head in the afternoon
07:58	mason	hiya guys
07:58	kados	excellent
07:58	kados	hey mason
07:58	kados	mason: kinda late, eh? :-)
07:58	kados	paul: I hear your new home is very beautiful
07:59	mason	i have some acqui. changes done in that last couple of weeks that ill commit to head in the next hour, too
07:59	mason	tis only 1am :)
07:59	paul	great mason, we will wait for them then
07:59	mason	cheers paul
08:00	kados	paul: it's good to have you back :-)
08:01	kados	paul: been very lonely in #koha in the morning for me :-)
08:02	kados	paul: http://liblime.com/public/roundtrip.pl
08:02	kados	paul: there is a sample of a preprocess script that only converts to UTF-8
08:02	kados	paul: for MARC21 only
08:03	kados	paul: it also keeps track of problem records and saves them in XML and iso2709 form in separate dump files
08:06	paul2	ok kados, i'll investigate in the next minutes, i'll let you know if I have another problem
08:41	kados	paul: I just investigated whether it would be possible to keep the rel_2_2 style API for searching, and I think it could be done with some minor changes
08:41	kados	paul: changes only to the routines in SearchMarc.pm
08:42	paul	good news...
08:42	kados	paul: the CCL style of searching is very similar to the old API
08:42	kados	:-)
08:43	kados	the only tricky part is how to integrate the results
08:43	kados	because with zebra, we get back full MARC records
08:43	kados	not bibids
08:44	kados	(in fact, we don't get the whole set of bibids at all ... that is kept in zebra)
08:44	kados	so maybe each MARC record in the result set must be processed to split it into Koha fields and MARC fields for display
08:49	paul	if you have the MARC record, MARCmarc2koha sub in Biblio.pm will create the Koha field easily
08:50	kados	cool, thanks
09:12	paul	kados : what do you mean by "update to the latest bib framework"? do you mean "update to the latest marc21 frameworks from thomas" ?
09:13	paul	if yes, what is specific with them (as I can't do it for unimarc)
09:15	paul	other question : I think convert_to_utf8.pl is redundant with updater/updatedatabase (see updatedatabase line 1404+ : it's the same alter table)
09:22	paul	kados : another problem/question : i try to run missing090field.pl, but it fails, as it requires zebra working. are you OK if I move this part AFTER Starting zebra ?
09:25	paul	kados : another one : biblio_framework.sql is partially already in updatedatabase. Are you OK if I put everything in this script ? (should be easy)
09:25	paul	same for phrase_log.sql
09:25	paul	hello tumer/
09:25	tumer	hi paul sorry about france
09:26	paul	the priest of my church is italian, so we all had a very large problem on sunday : could we have been happy with a victory from anyone.
09:27	paul	in fact, this end was the worst possible, but also the best possible : it means nothing to win with penalties.
09:27	paul	so it also means nothing to loose with penalties ;-)
09:27	tumer	not if you ask italians
09:28	paul	not all italians : our priest was happy, but only half happy.
09:28	paul	and in 2 or 3 days, their joy will end with the justice decision, probably.
09:28	paul	did you hear of this scandal in cyprus as well ?
09:28	tumer	paul i am almost finishing a complete new api for record handling. 1 question?
09:29	tumer	do you think we shouldmanage itemtype at biblio level or item level?
09:29	paul	mmm... this question should be asked to koha-devel. In UNIMARC, itemtype is a biblio level information.
09:30	paul	the best being to have something like FRBR, that is incompatible with unimarc.
09:30	paul	(and maybe with marc21 as well)
09:30	tumer	so you cannot hve a CD version of a book under same biblio
09:30	paul	no you can't, because it's a different intellectual object.
09:31	paul	otherwise, we could, for example, put all books from the same author in the same biblio.
09:31	tumer	a book with a cd causing us problems-- CD gets loaned out for shorrter
09:31	tumer	and sometimes separately
09:31	tumer	but they are in sme biblio
09:32	tumer	well may be we should change library policy tehn
09:33	paul	note that we could have itemtype at item level & still have the same behaviour as previously, the library just would have the same itemtype for each item.
09:34	paul	but I think it must be asked to koha & koha-devel ml
09:34	tumer	i'll ask it there
09:34	tumer	here is what i have changed:
09:34	tumer	biblio table has only marc,biblionumber and itemtype
09:35	tumer	items has itemnumber,biblionumber,marc,barcode
09:35	kados	paul: just make sure you have two conf files
09:35	tumer	no biblioitems
09:35	tumer	biblio also has frameworkcode
09:35	kados	paul: and you are using rel_2_2 bases when running missing090
09:35	kados	hi tumer
09:35	tumer	hi kados
09:36	paul	ah, ok , it means you must have 2 DB during migrations.
09:36	kados	tumer: I read your conversation with thd yesterda
09:36	kados	y
09:36	kados	paul: not quite
09:36	kados	paul: you run missing090 in the database before you mysqldump it
09:36	kados	paul: then you must run it again when you have the MARC data
09:37	kados	paul: (because sometimes there are still missing 090 fields and zebra indexing will crash if they are missing)
09:37	kados	tumer: please proceed with your plan and feel free to commit to HEAD
09:38	tumer	kados:it will break everything itsa a complete new api
09:38	kados	tumer: that's the point
09:38	kados	tumer: there is no official HEAD api yet
09:38	kados	tumer: what you are doing is closest to my vision of it
09:39	tumer	ok but head has lots of junk old script in it, i have cleaned mine
09:39	tumer	how can i delete from head
09:39	kados	good question
09:39	kados	I wonder if savannah supports subversion yet
09:39	kados	that would allow this
09:39	kados	you can also do:
09:39	kados	cvs delete file.pl
09:40	tumer	does that put them in archive?
09:40	paul	tumer: yes
09:40	kados	yes, you can't actually delete them
09:40	paul	(in Attic directory)
09:41	tumer	so what i will do is submit new biblio which will break all record handling ok?
09:41	paul	you plan to break everything ? sounds cool to me :D
09:42	kados	tumer: ok
09:42	tumer	i think thats the only way to actually get good clean code
09:43	kados	tumer: if you commit it I will take a look immediately
09:43	kados	tumer: and tell you if I don't like something :-)
09:43	kados	(note that HEAD is already completely broken)
09:43	kados	(so it won't change anything :-))
09:47	tumer	i am currently working on it hope to get it fully functional by this weekend and put it to production
09:48	tumer	then i will commit
09:49	tumer	btw it was not zebra problem zebra was crashing cuase this new M:F:X was not giving me xml_as_record
09:49	tumer	i did not understand why
09:58	kados	weird
09:58	kados	what new M:F:X?
09:58	kados	is it the newest from SF?
09:58	tumer	yep
10:00	tumer	i write $record->as_xml_record() and still get <colection> wrapper
10:00	paul	kados : I can't get missing090.pl work correctly.
10:00	paul	it calls MARCgetbiblio with a biblionumber as parameter.
10:00	paul	in rel_2_2, the parameter must be a bibid.
10:01	paul	and if I try with PERL5LIB pointing to dev_week, it fails, because the KOHA_CONF is not xml, but the old .conf file
10:02	paul	so, I have to run it from dev_week 100% :
10:02	paul	- copy rel_2_2 DB
10:02	paul	- run updater/updatedatabase
10:02	paul	- run missing090.pl
10:03	paul	the good news being that i don't have any missing 090
10:03	kados	excellent
10:03	paul	oups, no, I have a few.
10:03	kados	ahh...good for me :-)
10:04	paul	10 for a 14000 biblio DB
10:04	tumer	but should be 0
10:05	paul	right tumer
10:05	paul	it's really a lot
10:06	paul	kados : you've missed some of my questions it seems :
10:06	paul	- update to the latest bib framework
10:07	paul	means "update to marc21 from thomas" ?if yes what is interesting for me to know for unimarc ?
10:07	paul	(hello owen)
10:07	owen	Hi paul
10:08	tumer	paul i dont think you need it and infact should not otherwise you will loose your mappings
10:08	paul	- biblio_framework.sql, phrase_log.sql are already partially in updatedatabase. Is it OK if I put everything there ?
10:08	kados	yes
10:08	kados	everything should be in dev_week updatabase
10:09	kados	eventually I would have put it in but didn't get to it
10:09	paul	ok, i take care of it & update the wiki
10:09	kados	thx
10:09	tumer	biblio_framework update is not an essential part of this upgrade is it?
10:09	kados	yes
10:09	kados	it's poorly named
10:09	tumer	it merely adds more field definitions to what we already have and which we will not use
10:09	kados	biblio_framework.sql is what alters biblio and moves the frameworkcode as well as adding some columns
10:10	tumer	ahhh
10:10	paul	things that are already in udpatedabatase (like utf-8 conversion)
10:10	kados	utf-8 conversion should not be in updatedatabase
10:10	paul	why ?
10:10	kados	the wiki is not up to date with my latest tests
10:11	kados	because depending on the version of mysql you are migrating from, converting to utf-8 could mangle all your data
10:11	kados	if you are running rel_2_2 on 4.0
10:11	kados	there is no character set support
10:11	paul	right. but we require 4.1 isn't it ?
10:11	kados	but you table and database definitions are probably set to latin1
10:11	kados	I speak of migration
10:11	kados	for new installs the procedure is different
10:12	paul	so we should add a test to check for 4.1 when starting updatedatabase ?
10:12	kados	you can't
10:12	paul	I can't what ?
10:12	kados	you'd have to test the mysqldump to see if the data was coming from 4.0 or 4.1
10:12	paul	add a test or require 4.1 ?
10:12	kados	4.1 is required for dev_week
10:12	kados	but most libraries will be migrating from 4.0
10:13	kados	if you migrate from 4.0 you must do special things to preserve your character sets
10:13	paul	but I thought we have said previously that 4.1 will be required for Koha 3
10:13	kados	yes
10:13	kados	4.1 _is_ required for 2.4 and 3.0
10:14	paul	se a library migrating from Koha 2.2 will have to deal with mysql upgrade if needed.
10:14	kados	but most libraries are running 2.2 on mysql 4.0
10:14	paul	before upgrading.
10:14	kados	yes
10:14	kados	but that is a complicated process
10:14	paul	and we could check this in updater.
10:14	kados	I doubt it
10:14	kados	there are so many cases
10:14	kados	what encoding they started on, what they want to end up with, etc.
10:14	paul	so, in updater, if mysql 4.0 => stop immediatly, if 4.1 => continue
10:15	kados	if 4.1, make sure the mysqldump is from 4.1 or stop immediately
10:15	paul	but everybody & everything will be in utf8 in 3.0
10:15	kados	yes, but you need to convert it to utf-8 _and_ tell mysql that it is utf-8
10:15	paul	and mysql alter table know where it comes from isn't it ?
10:16	kados	http://wiki.koha.org/doku.php?[…]ncodingscratchpad
10:16	kados	ok
10:16	kados	I will give you NPL as an example
10:16	paul	yes, i've read, but that's not enough for me
10:16	paul	ok, listening
10:16	kados	before Koha NPL had MARC_8 encoded records
10:16	kados	they imported them into a mysql 3.23 db
10:16	kados	the database was set up for latin1 defaults
10:16	kados	so ...
10:17	kados	when NPL migrated from 1.9 to 2.0 using mysqldump
10:17	kados	all of the data was mangled by mylsql
10:17	kados	so now NPL has a database that thinks it has latin1 data
10:17	kados	but actually has marc8, mangled once in a conversion
10:18	kados	to upgrade, NPL must:
10:18	paul	what does 'mangled' means ?
10:18	kados	alter the mysql tables and convert to binary or blob
10:18	kados	(so now mysql doesn't know the encoding)
10:18	kados	(mangled means the characters have been re-encoded)
10:18	kados	(as was common when moving from 3.23 to 4.0 (which started having some character set support))
10:19	kados	(so now mysql doesn't know the encoding because the data is of type blob, etc.)
10:19	kados	so now we can run mysqldump without mangling the data
10:19	kados	(again)
10:20	kados	now we need to convert everything to utf8
10:20	kados	then we need to tell mysql that the tables are utf8
10:20	kados	and in the case of the MARC data, we need to update all the leaders
10:21	kados	but every Koha migration will be different
10:21	kados	depending on the mysql version and encoding defaults they are using
10:21	kados	and the type of MARC
10:22	kados	paul: make better sense?
10:22	kados	paul: for a new install it's much simpler
10:22	paul	yes partially make sense now.
10:22	kados	paul: a library must simply make sure that mysql 4.1 is set up with the correct encoding
10:23	kados	paul: and that all their data is in utf8 before importing
10:23	paul	to rewrite it : depending on what you really have and what mysql think you have, you will be in a big pain ;-)
10:23	kados	yep :-)
10:23	kados	esp for french libraries
10:23	kados	who have many non-ascii characters
10:24	paul	to transform from one to another encoding, is iconv enough (on the dump) ?
10:24	kados	you can transform within mysql itself
10:24	kados	I don't know about iconv
10:24	paul	how do you do to know what is the real encoding ?
10:24	kados	hehe
10:24	kados	there is no way to know
10:24	kados	unless you know beforehand :-)
10:25	paul	really nice...
10:25	kados	you can examine the hex :-)
10:25	kados	the reason I know all of this
10:25	paul	but we don't have any hex in mysql
10:26	kados	is because with WIPO, I was losing hair trying to get mysqldump to export utf-8 data to their db
10:26	kados	because I didn't know that mysql thought it had latin1, not utf-8
10:26	kados	very frustrating
10:26	kados	no hex in mysql
10:26	kados	you need to export to a file
10:26	paul	ah, I understand now why you don't have your long hairs anymore...
10:26	kados	well ... even that could change it
10:26	kados	some tests will need to be done
10:27	kados	mysqlhotcopy will take a snapshot of the db
10:27	kados	without touching the encoding
10:27	kados	so you can set up a test environment
10:27	kados	on another server
10:28	kados	it's a really big problem, unfortunately, maybe too big a job for updatedatabase :(
10:28	kados	but it's also unreleated to zebra :-)
10:28	kados	it's a problem we created ourselves :-)
10:28	paul	do you know how we created it ?
10:28	kados	yes
10:29	kados	because we said 'koha can run perfectly on any version of mysql with any settings'
10:29	kados	and not 'make sure you are running mysql 4.1 with encoding set to the same as your data'
10:29	paul	mmm... I never said that... but I agree that I never checked ;-)
10:30	kados	(we of course, didn't say it, it was a sin of omission :-))
10:30	paul	my problem is that I think I alway had iso8859 datas, but I was probably wrong, and I'll discover it soon
10:30	paul	another question :
10:30	paul	if the datas where not iso8859, how is it possible to have correct MARCdetail ?
10:30	paul	(with proper accented chars)
10:30	kados	hehe
10:30	kados	that's just the best part
10:31	kados	mysql will let you store any encoding
10:31	kados	it doesn't care what you put in
10:31	kados	except when you try to mysqldump or convert to another encoding
10:32	kados	so you may have correctly encoded values
10:32	kados	but to get them out you must trick mysql
10:32	paul	but when you put it on a web page, if it's not 8859, the page should be wrong in the browser isn't it ?
10:33	kados	isn't 8859 eq latin1?
10:33	paul	yep
10:33	kados	so if your encoding is 8859, and your db is 8859, and your meta tag is 8859, it's all good
10:34	paul	but you said that I could not be sure of my real encoding. So, if mysql = latin1, meta=latin1 and the pages are OK, then I have real latin1 datas ?
10:34	kados	sounds like it
10:34	paul	if yes, then all my libraries are ok !
10:35	paul	that's what is nice with a language that uses accents : any problem is easy to find ;-)
10:35	kados	:)
10:35	kados	so you can probably convert to utf8 before mysqldump
10:35	kados	(use a test db first of course)
10:36	paul	of course.
10:37	paul	bon appetit !
10:45	paul	kados : how do you export your datas ?
10:46	paul	because export/export.pl is buggy (compilation failure)
11:07	kados	paul: in rel22 it should work
11:10	paul	forget this, my copy was wrong. i had updated from something locally modified, so I got some cvs errors. restarting from a fresh copy
11:11	thd	kados: did you understand from reading the logs what tumer had done for holdings
11:11	thd	?
11:11	paul	hello thd.
11:11	thd	hello paul
11:16	kados	thd: yes, and I approved it
11:16	ToinS	paul : updatedatabase has not been merged between rel_2_2 & head
11:17	kados	thd: it will be in head as soon as tumer deems it stabler :-)
11:17	kados	ToinS: I did it yesterday
11:17	ToinS	kados ah ok
11:17	thd	kados: he is not using it in production yet?
11:18	thd	kados: did you see my message about how there are only about 3 fields for the librarian to worry about in standard MARC 21 holdings?
11:18	kados	thd: no
11:18	kados	thd: tell me
11:18	thd	kados: or 3 at a time
11:19	thd	kados: acquisitions would need 541, 583, and 876 that is just 3 fields with 877-8 for very special cases
11:19	dewey	i already had it that way, thd.
11:20	kados	mainly dev_week and rel_2_2 are merged
11:20	thd	kados: cataloguing would need 852 or 856, and sometimes 863-868
11:20	kados	there are a few that need to be manually merged
11:20	kados	like biblio.pm, if there was anything in there
11:20	kados	thd: if they are defined in tab 10 they will show up I think
11:21	kados	even in the current scheme
11:21	thd	kados: only 952 is defined in tab 10
11:22	thd	kados: would you not use the MARC editor to load a holdings record?
11:23	thd	kados: that would give the librarian access to whatever is needed
11:24	thd	kados: a holdings framework would in the case of MARC 21 be a very trimmed down bibliographic framework
11:25	kados	ahh ...
11:25	kados	right
11:25	thd	kados: the fields are all there it is just a question of unhiding them
11:25	kados	so the regular MARC editor would be used
11:26	thd	kados: although, tumer has spotted an editor bug for filling a field from an authority field
11:26	thd	s/field/record/
11:27	thd	kados: tumer reports that if you add a repeated field the authorised heading fills the wrong field
11:29	thd	kados: in his case he found that an attempt to add an authorised heading was filling 690 instead of 700 after a repeatable field had been added
11:29	kados	thd: I'll take a look
11:29	paul	kados/tumer/thd : which templates are OK for dev_week ? npl ?
11:30	thd	kados: will there be a meeting today?
11:30	thd	now that paul is back
11:30	kados	paul: only npl are tested
11:30	paul	i'm not back thd
11:30	thd	welcome back paul
11:30	thd	no
11:30	kados	paul: and don't expect searching to work except at opac-zoomsearch.pl
11:30	kados	paul: :-)
11:30	paul	i'm at ToinS home, so I will not be here everytime
11:30	thd	paul: you are still moving?
11:31	kados	thd: no time for a meeting today
11:31	paul	my move is done. but still no DSL
11:31	thd	paul: do you have dial-up
11:31	thd	?
11:32	paul	i've DSL at ToinS home. I may be able to have a dial up, but it's very expensive
11:33	thd	paul: I have cheap dial-up because my building has the wrong too old equipment for cheaper DSL and the telephone company will not replace it.
11:35	thd	paul: I will have optical fibre or a hole in a foundation wall before the telephone company will replace a circuit box from the 60s.
11:36	thd	monopolies in action
11:39	thd	kados: why was it so difficult to think of the simplest most efficient solution for standard holdings?
11:40	kados	thd: sorry, got a lot on my plate right now
11:40	kados	thd: I really want to resolve the standard holdings issues
11:40	kados	thd: and I think we're very close
11:40	kados	thd: but don't have the time today to discuss deeply
11:41	thd	kados: tumer has it by not needing repeated fields just repeated holdings records
11:42	thd	kados: OK I have to go now in any case
11:51	paul	kados ?
11:51	dewey	kados is probably becoming a true Perl Monger...
11:52	paul	zebraidx -g iso2709 -c /koha/etc/zebra-biblios.cfg -d kohaplugin update /path/to/records => /path/to/records is the path to all records, right ?
11:52	paul	so we have to export all of them in 1 big file (using export/export.pl) or 1 for each biblio ?
11:53	kados	all in one file
11:53	tumer[A]	paul one big chunck
11:53	kados	file must be named *.iso2709
11:53	kados	and put in /path/to/records
11:53	paul	the /path/to/record being a directory with only this file, or can contains other things ?
11:53	kados	it can contain other things
11:54	kados	but they will be ignored
11:54	tumer[A]	only with marc records
11:54	tumer[A]	sometimes does not ignore them
11:54	kados	only files with an excension of .iso2709 will be noticed
11:54	kados	ahh ... tumer is probably right, i have not tested this extensively
11:55	tumer	kados:it tries to read every file and sometimes goes crazy saying they dont have id's so best to have marc records only
11:55	kados	:-)

← Previous day | Channels | #koha index | Today | Next day → | Search | Google Search | Plain-Text | plain, newest first | summary