IRC log for #koha, 2006-11-25

All times shown according to UTC.

Time S Nick Message
12:37 kados paul: congrats!
12:37 paul you'r welcome ;-)
12:38 kados hehe
12:38 kados now I will have less questions about the Z-client ;-)
12:42 paul for sure. and me too...
12:54 kados paul: got a second for a question?
12:55 kados paul: does SAN plan to import data into Koha?
12:55 kados or do they start from scratch?
12:56 paul heum... with 300 000 items in their 8 branches, you can imagine they don't plan to strart from scratch :-D
12:56 paul so they migrate, of course.
12:56 kados i suspected this :-)
12:56 kados so how will they migrate with a broken bulkmarcimport? ;-)
12:57 paul hpassini wrote a specific bulkmarcimport, based on a 2.2 setup.
12:57 paul then 2.2 => 3.0
12:57 paul so, they don't use rel_3_0 bulkmarcimport
12:58 kados why?
12:59 kados it seems to me that rel_3_0 has the potential to be _much_ faster
12:59 kados for import
12:59 paul kados : i suspect bulkmarcimport can be up to 10x faster in 3.0, of course
13:00 paul but everything is done, well tested and will have to be run only once on dec, 8th at SAN, so, they will use their actual tools
13:00 paul (between dec, 8th and jan, 3rd, no acquisitions or cataloguing, just circulation on previous ILS)
13:01 kados ok
13:01 paul not a prob in france, as all budgets mus have been spent before dec, 1th or something like that, in public libraries
13:01 kados right
13:02 paul (that's why, a stupid thing imho, the SAN spended 11k¤ in 1 day, because it was the last day they could order... so they acquired anything and probably silly things :-( )
13:02 kados maybe they should have purcahsed software development!
13:02 paul lol.
13:03 paul impossible because not the same budget
13:03 kados ahh
13:03 paul in France, funds that are affected to "investments" can't be affected to "employees" or "daily spends"
13:04 paul (same empl => invest, or daily => emplo...)
13:04 paul they even can't replace books by computers, because computers, that are more than 500¤, are considered as investments (the rule uin France)
13:05 paul so they could have spend 11k¤ buying pencils or paper, but not computer or software development
13:05 kados wow
13:29 paul kados still around ?
13:29 kados yep
13:29 paul in rel_3_0, Search.pm, getMARCsubjects
13:29 paul it seems you create a marc subject line for each subfield.
13:30 paul in rel2_2, a marc subject was created for each field (subfields being concatenated)
13:30 paul sample :
13:30 kados yes, I can explain
13:30 paul $achemistry$bpetrol
13:30 kados it's not completely finished
13:30 paul ok, i'm listning
13:31 kados and allow clicking on the first part of a subject and only use it for searching rather than the whole field
13:31 kados so #achemistry$bpetrol
13:31 kados if you click on 'chemistry' it searches for just chemistry
13:31 kados if you click on 'petrol' it searches for both
13:32 paul (that's now what is done atm)
13:32 paul (or it's buggy ;-) )
13:32 paul s/now/not/
13:32 kados right, it's buggy
13:32 paul but the idea sounds cool to me
13:32 kados there is another issue
13:32 paul (& unimarc compliant)
13:32 kados zebra did not support cross-field phrase searches
13:32 paul ???
13:33 kados so you could not have a phrase index for both $a and $b
13:33 kados you would have to construct it manually:
13:33 kados su=chemistry and su=petrol
13:33 kados not su="chemistry petrol"
13:33 kados but I believe ID has added this feature to zebra 2.0
13:34 paul su="chemistry petrol" is found only if $achemistry petrol, right ?
13:34 kados we will have to ask them
13:34 kados correct
13:34 kados there is a new preference if I'm not mistaken for Zebra 2.0, but it might be undocumented
13:34 kados does that explain my choices?
13:35 paul yep.
13:35 kados (I 100% agree this could be improved, it is minimal functionality for now because i was in a hurry ;-))
13:35 paul although it's still unclear how i'll modify it for unimarc...
13:36 kados ? does it need to be modified?
13:36 paul yes it needs, at least to hide $9
13:36 paul (that is authority number)
13:36 kados ahh, well in usmarc too
13:36 kados quite simple to do so
13:36 paul (as well as $3)
13:36 kados yep
13:36 kados it can/should be done
13:37 kados one way is by harcoding a check for this in the sub
13:37 paul another thing that can be trouble shooting users :
13:37 kados another would be to add a new framework setting
13:37 paul imagine 2 authorities :
13:37 kados or to exploit the 'hidden' field
13:37 paul auth A : $91234 $achemistry$bpetrol
13:38 paul auth B : $94321 $a petrol $b ressources
13:38 paul if the user sees a biblio with auth A and clics on "petrol"
13:38 paul he will have answers including biblios with auth B
13:38 paul which librarians will dislike.
13:38 kados right
13:38 paul that's why, in 2.2, the link was on $9
13:39 kados if authorities are enabled it should be on $9
13:39 kados IMO
13:39 paul so, if there is a $9 => we use it, otherwise => as it is actually.
13:39 paul sounds good to me.
13:40 kados yep
13:40 paul (& very flexible, because in UNIMARC, 610 are subject related, but "free subjects", so without authorities !)
13:40 kados exellent
13:40 kados so it's a new feature even ;-)
13:40 paul yep & i should commit it in a few minuts.
13:41 kados woot
13:41 paul (as it's simple to implement i think)
13:41 paul another question : reordering don't work on my setup.
13:41 kados hmmm
13:41 paul (I mean, selecting title, author, or any other order don't change anything)
13:41 paul do you have an idea why ?
13:42 paul + how do you specify the order in ccl syntax
13:42 kados do you have sort indexes defined?
13:42 paul how can I check ?
13:42 kados if your record.abs has :s for some fields like title, etc.
13:43 kados (and note that 'rank' is different than 'sort')
13:43 paul melm 200$aTitle:w,Title:p,Title:s
13:43 paul melm 200$cTitle:w,Title:p,Title:s
13:43 kados so it seems yes
13:43 paul so I suppose yes, I have title sorting
13:43 paul melm 200$fAuthor,Author:p,Author:s
13:43 paul melm 200$gAuthor,Author:p,Author:s
13:43 paul and authors as well
13:45 paul mmm... in fact it seems there is an ordering, but I can't find how it works :
13:45 paul http://i17.bureau.paulpoulain.[…]_by=1%3D1003+%3Ci
13:45 paul with Author:s on 200$f, 200$g, 700$a, 700$b ...
13:45 paul which field is selected ?
13:45 paul (for ordering I mean)
13:45 kados that could be confusing
13:46 kados I usually only define one sort index
13:46 kados for each 'type'
13:46 paul ok, sounds logic.
13:46 kados so for title, i always use Title-cover which is 245$a
13:46 paul i'll try to change my setup.
13:46 paul a last question : Relevance & popularity are based on what ?
13:47 kados relevance is a very long topi
13:47 kados I could spend a long time talking about it ;-)
13:47 kados popularity is simple
13:47 kados it's items.issues
13:47 kados which is auto-incremented on every circulation
13:47 paul yes, I know. I wanted to know how you declare it in ccl.properties & record.abs
13:47 paul (& you update it with update_items.pl)
13:48 kados first it must be in the framework
13:48 kados so that it appears in the MARC data
13:48 kados (remember, items table is authoritative for items, marc is just a copy)
13:49 kados in my ccl.properties I have:
13:49 kados popularity 1=9523
13:49 kados then in bib1.att I have:
13:49 kados att 9523            popularity
13:50 kados and finally in record.abs I have:
13:50 kados melm 952$s      popularity:n,popularity:s
13:50 kados (of course, this is dependent on 952$s being mapped to items.issues in your framework)
13:51 kados paul: does that make sense?
13:51 paul yep, 100%
13:51 paul and for relevance, there is something specific ?
13:51 kados that's the default
13:51 kados at least in dev_week
13:51 kados it means 'relevance ranked, field weighted, stemmed'
13:51 kados but it's a very advanced topic
13:52 kados and difficult to explain on IRC ;-)
13:52 paul ok, i'll try to investigate to undertand by myself...
13:52 paul thx anyway
13:53 kados the zebra guide may be of some help
13:53 kados http://indexdata.dk/zebra/doc/[…]ation-ranking.tkl
14:56 toins bye all
16:12 pecisk Hello there, when there will be next Koha conference? I know there was one in may this year. :)
16:14 kados pecisk: we don't have one officially planned yet
16:16 pecisk but is there any thougts, "unofficial" ideas? :) About date, for example? Is there planned any representation of Koha in other conferences or events? :)
16:17 kados Koha will be represented at ALA annual by LibLime
16:17 kados also, Katipo typically goes to Lianzza
16:17 kados other than that I'm not sure
16:18 pecisk I see
16:18 pecisk thanks for info :)
16:18 kados np
16:18 kados do you have an event in mind?
16:19 kados pecisk: I work at LibLime so I'd be interested in hearing if you have suggestions about good cons (maybe even in Latvia? ;-))
16:19 pecisk not yet, but I could have :)
16:20 pecisk We would like to get know Koha better, so we look if there is some event where we can meet people who uses and develops it
16:20 kados my email is jmf@liblime.com, so if you have ideas feel free to let me know, or else post to the koha list
16:20 pecisk sure :)
16:20 kados ahh, right
16:20 kados pecisk: have you seen the Koha ZOOM system? : http://search.athenscounty.lib.oh.us
16:21 pecisk yeah, I seen it, it is really great :)
16:21 kados (thanks)
16:21 pecisk sure :)
16:21 kados pecisk: Labdien!
16:21 pecisk kados: Labdien, labdien :)
16:21 pecisk it is Labvakar now, but anyway
16:21 pecisk :)
16:22 kados hehe
16:22 kados in US it's Labdien :-)
16:22 pecisk ohhh
16:22 pecisk how do you know that?
16:22 pecisk :)
16:22 pecisk I mean, language?
16:22 kados ahh, just picked it up
16:22 pecisk hehehehe
16:22 pecisk information junky like me
16:22 pecisk :)
16:22 pecisk ohh
16:22 kados exactly
16:22 pecisk I see
16:22 pecisk cool
16:22 pecisk :)
16:23 kados anyway, lunch
16:23 pecisk yeah, bone appetite :)
06:36 btoumi hi all
06:36 toins hi btoumi
07:20 _btoumi _btoumi
10:04 js hello #koha
10:59 hdl kados tumer around ?

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