IRC log for #koha, 2005-12-01

All times shown according to UTC.

Time S Nick Message
11:27 kados paul around still?
11:27 paul yep, it's 3PM in france ;-)
11:27 paul hi kados
11:27 kados hi there :-)
11:27 kados how soon will 2.2.5 be out?
11:28 paul christmas gift. or new year gift.
11:28 kados do you have a date set?
11:28 kados ok ... good enough
11:28 paul did you reach my mail to ineo ?
11:31 kados yep
11:31 kados it's very exciting
11:32 kados I hope their developer will touch base with us
11:32 kados instead of just 'doing his own thing'
11:32 paul yep. is your roadmap for 3.0 finished ?
11:32 kados not completely
11:32 kados but it's getting there
11:32 paul i'll take care of this, don't be afraid.
11:32 kados :-)
11:32 paul and they said and said again they didn't plan to fork.
11:32 kados right
11:46 kados paul: what is the upgrade path to HEAD these days
11:46 paul ???
11:46 kados paul: can I skip 2.2.2 and go 2.2.0 to 2.2.4 to HEAD?
11:47 paul yep.
11:47 kados k ... I'm installing a HEAD server today
11:47 kados to help me finalize the roadmap, etc.
11:48 paul did you bug hdl about what he is coding atm ?
11:48 paul (i mean new features, not bugfixes on 2.2)
11:48 kados no I didn't get a chance
11:48 kados what is it?
11:48 _hdl_ yes
11:48 paul joshua, you've got the chance now;-)
11:49 paul explain joshua what you are coding on Head (opac alayahoo...)
11:51 kados _hdl_: yahoo features?
11:52 _hdl_ (tel)
13:09 paul hi owen
13:09 owen Hi paul
13:34 hdl kados
13:34 hdl are you there ?
13:35 kados hdl: yep
13:36 hdl About the OPAC "alayahoo"
13:36 hdl It consists in a Opac Categories browsing.
13:37 kados cool
13:37 hdl It displays number of biblios in relation with this category.
13:38 hdl And when clicking on this number, it would display all the biblios.
13:39 kados is it tied to subject headings? or does it consist of local tags?
13:40 hdl No it is tied to whatever field you would like, for instance itemcallnumber or dewey or lccn.
13:44 kados I mean, how do you pick 'category'
13:44 kados what are the categories?
13:47 paul kados : categories will be 1 MARC field (choosen in systempreferences). usually it will be dewey, sometimes CDU.
13:48 paul and it should work with any classification.
13:48 kados interesting ... so it adds a 'related items' to the OPAC eh?
13:48 paul related items ?
13:48 kados 'items related to the one I'm looking at'
13:49 hdl Yes and no.
13:49 paul mmm... good suggestion, I did not had this idea, but it's a good one !
13:50 hdl It allows to settle sets of biblios and levels between sets.
13:50 hdl That is you can have say ten levels of categories...
13:50 hdl 1st level :
13:50 hdl History
13:51 hdl Religion
13:51 hdl Science
13:51 hdl Second level :
13:51 hdl American History
13:51 hdl French History
13:51 hdl Bouddhism
13:51 hdl Catholicism
13:52 kados ? isn't that the same as subject classsifications as defined by LOC?
13:52 hdl s/Catholicism/Christianism/
13:52 hdl kados : could be but not necessary.
13:52 kados (only Koha's current MARC support doesn't support the ordering of them )
13:53 kados how does a library define them?
13:53 hdl It is based on a MARC field.
13:53 hdl In a systempreference : opaccategories
13:54 hdl one enter the field 100f
13:54 hdl 100f is an instance, not quite relevant.
13:54 hdl In france, dewey would be 686a
13:55 kados but dewey is just a number ... not 'History' or 'Religion' right?
13:55 kados so how does a library define 'this dewey number = 'History''
13:56 hdl (coming...) then, at first, categories would be divided by the n first characters into an authorised_value list.
13:56 hdl Say OPAC1 is for 1st character
13:56 hdl OPAC1 is for the 2 1st character
13:56 hdl OPAC2*** is for the 2 1st character
13:56 hdl OPAC3 is for the 3 1st character
13:57 hdl the value would be the characters
13:58 hdl and the lib... (don't know the english term) would be the "name of the categories plus [number biblios]
13:59 hdl The number of biblios would be regularly updated by a croned task
13:59 kados hmmm ... couldn't the same thing be acomplished by using properly ordered subject (600 fields in MARC21, dunno in UNIMARC)?
13:59 kados ie, if you have subject authorities
14:00 kados and order is preserved when importing 600 fields (which it is now IIRC now)
14:00 kados then you automatically have the categories right?
14:01 hdl But the fact is that then, one need authorities and subject authorities to acheive this.
14:03 kados hmmm ... but right now, one can't achieve it at all because 600 field order is not preserved (correct me if I am wrong)
14:03 hdl Our proposition is to base things on a well known classification. or on itemcallnumbers, which are commonly used in our customer library.
14:03 paul kados : the ordering of authorities is in CVS
14:03 kados paul: ok ... thanks for that clarification
14:04 paul mmm... not sure we speak of the same ordering...
14:04 paul i speak of the result page you get when you search an authority.
14:05 kados I'm talking about subjects ordering (fields in tag 6XX) when importing into Koha
14:05 kados so for example
14:05 kados I have a subject that catalogers setup:
14:05 kados (false example because I'm not a cataloger)
14:06 kados 600 $bHistory $aUnited States $fOhio
14:07 kados Koha will reorder this :
14:07 kados 600 $aUnited States $bHistory $fOhio
14:07 kados thereby losing the category levels
14:07 kados (I believe this only happened with Koha 2.2)
14:07 paul right. there is no simple solution to this. Except that in CVS there is a new button in MARC editor to reorder the subfields.
14:08 kados (as I believe koha 2.0 used Chris's original subjects code)
14:08 paul thus you get $a$b$f and you can reorder manually to $b$a$f
14:08 kados right ... for all 130,000 records :-)
14:08 paul (dirty hack, your expert MARC editor will be the best solution)
14:08 kados well ... here's my proposal
14:08 kados we fix the ordering problem on import
14:09 paul "the koha team agains unemployement : 1000 catalogers hired by Koha libraries !" ;-)
14:09 kados hehe
14:09 kados if we fix it, we will automatically have a very nice complex list of categories with levels built in
14:09 kados even without authorities
14:10 kados and we will have better support for MARC :-)
14:10 paul mmm... I think we should separate classification and authorities.
14:10 paul classification : hierarchic numeric/alpha, like dewey / CDU ...
14:10 paul authorities (constructed authorities) : hierarchic through MARC.
14:11 paul subfields.
14:11 paul hdl works on navigation on classification.
14:11 kados ahh ... so maybe a 'browse the same shelf' feature
14:12 kados 'items near this item'
14:12 kados (I realize location is different from classification ;-))
14:13 kados am I correct?
14:14 hdl you could do it with itemcallnumber.
14:14 hdl is that not quite like shelf browsing ?
14:14 kados I think so
14:14 kados at least at NPL
14:15 kados though I think some libraries distinguish between 'location' and 'classification'
14:15 hdl what does itemcallnumber look like ?
14:15 kados and that would be similar but not the same as shelf browsing
14:15 kados hdl: it depends on the type of the item
14:16 kados hdl: for fiction:
14:16 kados hdl: SF Stephenson
14:16 kados hdl: is the callnumber for 'cryptonomicon' as well as all of neal stephenson's other books
14:17 kados for non-fiction:
14:17 kados hdl: 641.5 SAI
14:17 kados hdl: "The art of Greek cookery, based on The Grecian gourmet"
14:17 hdl dewey space Author ?
14:17 kados sometimes yes
14:18 kados to be honest, NPL isn't a great example
14:18 kados because our catalogers aren't really professionals
14:18 kados oops ... did I say that out loud?
14:18 kados :-)
14:18 hdl It surely IS. since it is the case most of the time....
14:19 hdl would it be beacuse there were MANY catalogers with MANY experiences and way of mind.
14:19 hdl or becaus noone ever came across what real classification was.
14:19 hdl :)
14:19 kados partly ... also because they have chosen local classification schemes that don't really follow any known standard for classification
14:21 hdl ok.
14:24 hdl Are you OK with Opac "alayahoo" feature ?
14:28 kados yes ... it sounds exciting
14:28 hdl paul thanks for your help.
14:28 kados sorry I missunderstood at first what it was :-)
14:29 kados (I may want to change the name from 'alayahoo' to 'browse shelf' or something)
14:29 kados (if that's OK)
14:29 paul that's OK to me
14:29 hdl OK
14:29 paul + hdl, & kados :
14:29 paul if we use itemcallnumber, then we may have letters (3 first of the author)
14:30 paul but if we use dewey code (676 in unimarc) we should not have it
14:30 kados paul: at NPL you would have both cases
14:30 paul note we may have it sometimes, so we have to solve the problem...
14:30 paul maybe drop trailing letters ? or anything after the space ?
14:31 kados it might need to be cusomized by each library ...
14:32 kados because classification varies greatly from library to library
14:32 kados (at least in US libraries)
14:33 kados hmmm ...
14:33 kados here's an idea ... instead of building a new table
14:33 kados why not just query Zebra for the data?
14:34 kados and dynamically build the list of 10 or so 'nearby' items
14:34 kados ie, 'items before' and 'items after'
14:34 kados they you don't have to worry about a batch
14:34 kados and we don't replicate our data yet again :-)
14:34 paul the new table could be useful because in my idea, we should show the code + the text attached to it.
14:34 paul something like :
14:34 paul "browse the catalogue :
14:34 paul 1- XXXX
14:35 paul (nnnn items)
14:35 paul 2- religion
14:35 paul (ppppp items)
14:35 paul 3- history
14:35 paul (qqq items)
14:35 kados I see
14:35 paul when the user clic on 3, he get :
14:35 paul 30 - XXX ...
14:35 paul but you idea of seeing nearby items is very good !
14:35 kados ok ... here's my idea
14:36 kados what you are describing above would work better with subject headings IMO
14:36 kados whereas a 'nearby items' feature woudl be good after I have already found a book
14:36 kados then I can click 'see items near this item'
14:37 kados let me find an example
14:37 kados
14:38 kados search for 'neal stephenson'
14:38 kados then bring up first record
14:38 kados and click on the 'classification code'
14:38 kados you will see a list of items on the same shelf
14:38 kados but also ... note the 'show similar items' button at the bottom of the detail page
14:39 kados it uses subject headings to display related items
14:39 kados (I think ...)
14:39 paul subject headings : there are 6
14:39 paul (in your sample)
14:39 kados yep
14:40 paul do you mean a "similar item" has the same 6 subjects ?
14:40 kados according to this OPAC :-)
14:40 kados try the linnk
14:40 kados link I mean
14:40 kados not exactly the same subjects ... but it seems to use the subject terms as a keyword search
14:42 kados IMO classification is not the best way to do categories
14:42 kados especially not categories that look so much like subject categories
14:43 kados but classification is a great way to do 'nearby items'
14:43 kados if we want subject categories then I suggest we fix koha's ordering problem wtih 600 fields
14:43 kados paul: btw: do you have any ideas for how to do this for Koha 3.0?
14:44 paul it's almost 7Pm in france, i have to leave.
14:44 paul i'll investigate joshua.
14:44 kados ok ... thanks
14:44 kados I hope I make some sense :-)
14:44 paul let's add this to the long long long list of my idea ;-)
14:44 kados hehe
14:44 paul (just a last word : i've added some webservices for a customer (work on 2.2.4))
14:45 paul I have some ideas with webservices...
14:45 kados for acquisitions?
14:45 paul (including catalogue in a CMS through webservice)
14:45 kados nice
14:45 paul no, the webservice was for catalogue querying
14:45 kados I see
14:45 paul we will speak of this tomorrow ;-)
14:45 paul (if you're there)
14:46 kados ok ... talk to you tomorrow
14:46 kados I'll be here all week
14:46 paul_away bye world
14:50 hdl bye paul
16:09 rach morning
17:34 kados morning
17:34 kados I think the lists are all moved
17:35 kados except for manage, advisors, and testing which I'm proposing we get rid of
18:03 kados chris:[…]w_bug.cgi?id=1033
18:03 kados russ:[…]w_bug.cgi?id=1033
18:03 kados when you get a chance, update with new mailing list info
18:04 kados (not sure who's doing that these days)
18:10 rach will be russ
19:59 thd kados: are you present now that NZ has returned?
19:59 kados on phone
19:59 thd kados: beep when off phone
20:34 thd kados: are you still on phone?
00:02 thd +
01:59 rach For ayone who wants to see "bump" progress -
01:59 rach well and of course rosalie being sworn in as  a JP (Justice of the Peace)
06:22 Sylvain hi all
06:33 hdl hi alone
06:33 Sylvain :)
06:45 Sylvain Little request from ENSMP, in normal view, it is shown if an item is on loan but in marc it isn't shown, has something been developped to change  that ?
07:10 paul sylvain : nope, ENSMP is right
07:10 paul (still have to answer infos mailing list on this)
07:10 Sylvain I was hoping another answer ;)
07:10 paul sorry.
07:11 paul that should not be to hard to fix for me
07:11 paul (but very tricky)
07:21 thd good morning paul
07:23 thd |hdl| are you logged in or your ghost?
07:23 |hdl| yes
07:23 |hdl| phoning
07:25 thd |hdl| let me know when you are off phone.  I have some legal caution for your classification hierarchy feature.
07:37 paul |hdl| : une petite envie d'aller au Maroc ???
07:52 |hdl| thd back.
07:56 thd hdl: Your proposed classification browsing feature is a very good idea.  I have code written in Python that was written for that problem domain.
07:57 hdl thd: Code used with koha ?
07:57 thd hdl: If GPL version 3 fixes the right problems I would release that code.
07:58 thd hdl: I have not mingled that code with Koha although GPL version 2 would allow it as a web service.
07:59 thd hdl: The appropriation of my code as someone else's close web service is what I guard that python code against..
08:01 paul why do you want to guard your code thomas ?
08:03 thd paul  hdl one moment.  I may need to restart my Xserver.
08:17 thd paul hdl: sorry, too much cutting and pasting between xterms causes problems for me eventually.
08:20 thd paul hdl: The problem I have with sharing my code for classification browsing or even showing the user interface just yet is that I fear losing the possible business from it to a  closed source company with much much better ability to advertise than I have myself.
08:22 paul if chris, joshua or me had had the same fear, Koha would not be what it is nowadays !
08:22 thd paul hdl: I would freely share everything else I have but that until or unless GPL 3 corrects the web services loophole.
08:24 thd paul: I have no shortage of ideas unrelated to classification browsing that I will happily contribute to Koha at least equal to what has been done so far.
08:25 thd paul: My business model for classification browsing is web services and not support for individual libraries.
08:26 thd paul: Actually, I have not worked on my classification browsing code for over a year and I have worked on other things for Koha.
08:27 thd paul hdl: I would still be happy to share some of my understanding of classification browsing.
08:28 thd hdl: There is a legal problem with using the DDC.
08:29 thd[…]ember/034527.html
08:31 thd OCLC owns the DDC and the file real lawsuits in real courts over seemingly frivolous claims which require a real defence that cost real money if you intend to defend them.
08:33 thd paul: Are all Koha copyrights assigned back to Katipo.
08:33 thd ?
08:36 thd paul hdl: Who holds the copyrights for Koha code?
08:38 thd paul hdl: are you still there?
08:45 thd paul hdl: ping
09:02 hdl thd ! back
09:03 thd hdl: Who holds the copyrights for Koha code.
09:03 thd ?
09:04 thd hdl: Is it not Katipo that has the assignment?
09:05 hdl Yes.
09:05 hdl paul settled a french ownership on Koha Trademark.
09:07 thd hdl: If you are lucky, OCLC might only threaten the offending library and Katipo, but maybe not you; if a library used Koha to represent the DDC in some manner of which OCLC disapproved.
09:09 hdl thd: the fact is that we donot want to make only DCC but also allow people to customize their classification.
09:09 thd hdl: OCLC requires a license to use the current form of the first three DDC levels.  That is the levels up to the decimal point.
09:09 hdl Maybe we should get in touch with OCLC.
09:10 hdl Do you have an idea of the License cost ?
09:10 thd hdl: Obviously there is no problem with custom classifications.
09:11 hdl Maybe with hiding the figure and displaying only classification labels, that would go.
09:11 thd hdl: The price is based on the institution size and membership arrangements and is not really expensive as I recall..
09:12 hdl kados there ?
09:12 thd hdl: I would suggest a public domain approach in relation to the original form of the DDC.
09:13 hdl Could you be more precise ?
09:14 hdl (paul and I were at lunch. 13pm in france.
09:14 hdl )
09:14 thd hdl: They refuse to license use of anything beyond the decimal point for a hierarchal classification browser as opposed to a shelf listing.
09:15 hdl thd : just precise a little more what you understand with public domain approach.
09:15 thd hdl: I thought that you had taken grave offence at my concern over protected my code.
09:16 thd when I had no reply
09:16 hdl No. For sure.
09:16 hdl Wa are not THAT touchy. although we are frenchy ;))))
09:17 hdl Nowadays, we know that laws and copyrights HAVE to be abided by.
09:19 hdl In France, the french governement wants to vote a thing that would oblige softwares to use DRM.... Working with opensource softwares leads to be quite sensitive to copyrights and law matters.
09:20 hdl Not to throw it to Trash, but to understand and keep in mind what is Legal and what is not.
09:21 thd[…]513-h/12513-h.htm
09:21 thd There is at least one later edition in the public domain.
09:24 thd hdl: Representing the whole current DDC is something that OCLC reserves to itself as a part of a cataloguing service.
09:27 thd hdl:  Actually, just to be clear, numbers past the decimal point in DDC , not that they can be offered in a public browsing service, sometimes involve tables.  Therefore their meaning is not as straightforward as at least the first three numbers are.
09:28 hdl Yes.
09:30 thd hdl: OCLC objects to public services representing the whole DDC you could perhaps have something for librarian's that used the numbers past the decimal point if it was not in the OPAC.  Maybe OCLC would object to that as well though.
09:32 hdl ok.
09:34 hdl Using DDC as proposed in your link would be using public domain DDC ?
09:34 thd hdl: The whole UDC can be licensed.  The LCC is in the public domain but funtions very differently and requires use of adding number tables throughout to a much greater degree than the DDC or UDC.
09:36 thd hdl: As long as the copy is from 1923 or earlier now under the current Berne convention you should be safe.
09:39 thd hdl: From the point of view of OCLC trustees, public use of the DDC is seen as a threat to wholesale copying as an alternative to some libraries obtaining a license at all.  They actually claim that this was a real problem before they adopted their rather draconian attitude.
09:42 thd hdl: There may be ways to get avoid hazards of copyright as in your suggested possibility of excluding the numbers (maybe with changing the official matching names also).
09:43 hdl Manythx.
09:44 thd hdl: However, OCLC is very litigious and not at all nice even if they do have nice people who work for them.
09:45 paul thd : in my idea, Koha just deals with the number. whatever the meaning.
09:45 paul it's the library that fills the "meaning" of each code.
09:45 paul thus, it works with any classification, and we don't have to deal with copyright
09:46 thd hdl: They actually brought a case to court against a hotel in New York for using a quasi DDC theme that was not even DDC.
09:46 paul (as I hope a classification based on a number of digits is not @ i hope !!!)
09:46 paul I know.
09:46 paul and iirc, they won !
09:47 thd paul: The hotel settled and paid something modest to OCLC then made some changes in their marketing.
09:49 hdl But I think that to avoid  any problem, we (Koha team) could get in touch with OCLC and ask for a license... Or right to use.
09:49 thd paul: I understand that your idea would merely enable a library to do something to which OCLC might object.
09:49 paul yes, but that's already the case with every ILS isn't it ?
09:49 paul (as every ils can store dewey number)
09:50 paul and in France 90% of libraries uses dewey
09:50 paul (on a 3 digits basis often)
09:50 paul (only the 3 first digits I mean)
09:50 thd paul: I am merely asserting that, given the litigiousness of OCLC, they might choose to act against Koha and Katipo at least as well.
09:53 thd paul: Well, other ILS systems do not have what could be done very nicely that I had hoped hdl was proposing to implement.
09:54 thd paul: BNF does have classification browsing in a good and very useful sense, although it could be much better.
09:55 thd paul: Many ILS systems have call number browsing without a hierarchy to which OCLC has no objection that I know.
09:56 paul could you describe it ?
09:56 thd paul: Koha has that somewhat already.
09:57 thd paul: Are you asking for a description of call number browsing?
09:58 paul yes. (and where it differs from what we plan)
09:59 thd paul: call number browsing starts from a search rather than a classification hierarchy.
10:00 paul can the user go up and down in the hierarchy ?
10:00 thd paul: Search for a complete or truncated call number or follow a call number link in the retrieved set of another search..
10:01 paul yes, but the user has to know what a number means !
10:02 thd paul: In classification browsing the user does not see the hierarchy but follows the endless sequence of all call numbers sorted by call number.
10:03 paul that's the "see nearest" feature we skope with joshua
10:03 paul s/skope/spoke/
10:04 thd paul: As soon as the meaning of the number is given you start to cross over into the area where OCLC asserts its copyright for DDC.
10:05 paul my opinion is that it's stupid to copyright an idea.
10:05 paul and dcc is nothing more than an idea !
10:06 thd paul: Of course it is stupid but my work has concentrated on the public domain or copyright held in trust by the US government for all the people with LCC.
10:06 paul and is there a classification under public domain
10:07 paul (except the 3digits dewey you pointed)
10:07 thd paul: Decimalised classification is merely an idea but a particular fully specified scheme can be copyrighted.
10:08 thd paul: LCC
10:08 thd paul: Library of Congress Classification.
10:09 paul can the complete LCC be downloaded somewhere ?
10:09 thd paul: However, the construction of the DDC and coding around it is simple by comparison.
10:11 thd paul: Unfortunately, LCC had historically been a paper only set of schedules.
10:12 thd paul: Since 2001 the full schedules have been available in the MARC 21 classification format.
10:15 thd paul: Because that is recent I have not been able to identify any library where the complete set of LCC MARC records are held for the possibility of copying them.
10:17 thd paul: Also no commercial service that redistributes LC MARC records at reasonable prices has had enough customer interest in MARC classification records to justify offering them.
10:20 thd paul: LC CDC service uses some cost recovery formula to come up with the rather high prices for databases of MARC records by FTP.
10:21 thd paul: There are free test records from LC including a special set of classification records that they prepared for me.
10:23 thd paul: But I had stopped working on that problem over a year ago when my foot swelled very painfully to double size.
10:24 thd
10:26 thd paul: $4,360 for the backfile and  $4,360 annually to maintain a subscription.
10:33 thd paul: Well LC CDS seems to have removed the username and password for the test files.
10:42 thd paul: I am always willing to hear any ideas about how I might share my own classification browsing code without having Amazon or TLC for example putting it on their server and taking most all of the potential commercial value without any return for what I had developed or even getting modifications back.
10:43 paul my opinion is that such a work should not be copyrighted, so I can't help you on this. But you also can put a GPL licence on it : use, modify freely, but keep it free.
10:44 paul (& I still don't understand why it can be possible to @ an idea. I think i'll never understand...)
10:45 paul earning money just because I was the 1st to have an idea is something un-accessible to me, sorry
10:48 thd paul: My basic model for exploiting it was to offer free web services.  If Amazon were offering those identical services without even acknowledging my contribution then I fear that I would not have enough of a market left for me.  This is possible with web services under GPL version 2.  My trumpet is very much smaller than Amazon's for anything that I did where they might directly use my code.
10:51 thd paul: The idea is free.  You have it already in general.  Only very precise implementations should be subject to copyright.
10:53 thd paul: The only protection for ideas should be trade secret laws not copyright.  Ideas should be everyone's as soon as they are disclosed.
10:55 thd paul: US courts have only recently allowed the patenting of what they had previously held to be mere ideas.
10:55 thd paul: However, patents and software should never mix.
10:59 thd paul: I would not be first.  Maybe I would have the first free to use service considering that Amazon for example would never do the necessary work on their own and the nonfree services collect subscription fees.

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