IRC log for #koha, 2006-07-23

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

All times shown according to UTC.

Time Nick Message
13:23 owen kados: I notice call numbers aren't appearing on the opac-detail page. Could that be what Gloria is talking about?
13:28 kados hmmm ... maybe
13:31 owen kados, when you get a chance could you look at permissions again
13:33 kados owen: yep, changed back now
13:33 I've eliminated SearchMarc.pm now
13:33 all search routines are now in Search.pm
13:34 and I'm gonna be working on refining support for the old API a bit today
13:34 owen: as far as call numbers, if you want them to show up in the detail page, change 'classification' to 'dewey'
13:34 owen Still no luck
13:34 kados owen: but note that next time I migrate they will need to be changed back
13:34 owen: how about now?
13:35 (too many damn synlinks :-))
13:37 owen Finally!
13:37 kados hehe
13:37 sorry
13:49 owen kados: are there additional options for what information can be displayed with the Amazon reviews? I'm thinking reviewer name, review date, and rating
13:49 kados rating should be there
13:49 but yea, we can add stuff
13:50 anything exposed in the API can be added
13:50 owen Average rating is there, but not individual reviewer ratings
13:56 kados ahh, right
13:56 last time I checked that wasn't available in the API
13:56 but it may be now
14:24 owen kados, are you still around?
14:24 kados owen: yep
14:24 owen: oapcnav gone?
14:24 musta lost it with my last cvs update
14:24 owen Take a look at this: http://zoomopac.liblime.com/cg[…]ail.pl?bib=171479
14:25 Under "Format:" we're getting one of the notes fields
14:25 kados huh ...
14:25 it's in 520 a
14:59 the field in the Format line is 'abstract'
15:00 and abstract is linked to 530 $a
15:00 520 $a even
15:00 owen: what should it be linked to?
15:01 maybe we just need to change the name of the label to 'Abstract'?
15:02 well ... 'Summary' maybe
15:02 owen I expect the data under "Format:" to be 245h
15:02 The abstract is already showing up with the other notes
15:02 kados I see
15:03 so NPL musta co-opted abstract by calling it format
15:03 I mean that they put format data into abstract
15:03 and by they I mean we :-)
15:04 owen It was working before: http://search.athenscounty.lib[…]ail.pl?bib=171479
15:04 kados yep ... but we've got a completely new framework now
15:05 so I just need to adjust that slightly
15:06 but for the rest of the world, we shouldn't assume that they are mapping format to abstract
15:06 can we just have the format show up as part of the title or something?
15:07 do we need to have it in it's own spot?
15:07 (if so i can just add a new column to the biblioitems table for it)
15:09 owen There's no reason why it couldn't show up with the title
15:10 kados ok, I've got an idea ... give me a sec
15:15 thd I am back
15:15 kados: are you there?
15:17 kados thd: yep
15:17 thd kados: were you asking about a fast method of adding '--' subfield subdivisions?
15:17 kados well ... I was ...
15:17 I looked at the getMARCsubjects code
15:18 and I think I can use that as a point of reference
15:18 thd kados: that code relied upon MARC in SQL but the same idea should work.
15:19 kados yep
15:20 thd kados: to join subfields in a particular order would be making the same design mistake which the original MARC Koha design had in proscribing the order instead of reading it from the record.
15:21 kados: getMARCsubjects corrected that design mistake for subjects.
15:23 kados: 245 $h is the general material designation and had been misapplied for abstract in the original MARC 21 framework.
15:26 kados: if you read the MARC data from the marc_subfields instead of from the original Koha SQL bibliographic subfields you should not have any actual abstracts in your 245 $h.
15:30 kados: If it was not clear from earlier, the search links should match the whole subject as previously.  When there is some time, we could add support for mix and matching the individual subject subdivisions independently in an intelligent manner and not progressively in a crude manner.
15:31 kados: I have a question.
15:32 kados: I have noticed during the past few days that some important Z39.50 targets have 000/09 which actually lies about the record encoding.
15:34 kados: I have even noticed national libraries lying about the actual encoding, although, I have not noticed that for LC.
15:34 kados: we need an encoding lie detector.
15:35 kados: do you have some code that can be partly used for detecting encoding lies?
16:01 kados: where did you go?
16:01 kados thd: sorry ... super busy today
16:02 thd kados: I thought that you were busy fixing all the problems in Koha :)
16:02 kados: does that mean you have no time for my question?
16:11 kados I didn't notice it :-)
16:11 I don't have any code for that
16:11 but I have noticed the problem
16:11 thd kados: do you not have code for detecting encoding problems in records?
16:12 kados yes
16:12 but it's suspect code because it relies on the MARC::* stuff
16:12 thd kados: show me please, maybe I can adapt it in some way
16:12 kados which we're still not 100% sure works
16:13 thd kados: well If it does not work I will prod Ed summers to fix it.
16:14 kados: the MARC::* stuff has to be made to work if we are using it and Ed seems eager if there are good tests.
16:26 kados I agree
16:56 chris kados: u about?
16:57 kados chris: yep
16:57 chris just want to run something past you for sanity checking before i post to the list
16:57 kados sure
16:57 chris one of the public libraries here in nz using koha are quite keen to do more copy cataloguing
16:58 kados cool
16:59 chris the way they would like the system to work is, they order items in full acquisitions, then when it arrives they get the marc record for it, and load it into the reserviour .. then they would like a way the marc bit to choose to replace a record with one from the reserviour
16:59 kados yep, pretty common technique
17:00 chris so before i go to write something to do it, i wanted to check we cant already do it in koha?
17:00 kados SMFPL need the same thing
17:00 http://wiki.liblime.com/doku.php?id=koha226bugs
17:00 look under critical enhancement requests
17:00 chris sweet
17:00 kados we can't do it currently
17:01 but ...
17:01 chris well i have to do an estimate
17:01 kados with zebra it is really simple
17:01 you just do a specialUpdate
17:01 with the bib number
17:01 chris right
17:01 what i will try to do
17:01 kados in the old koha it might be more tricky
17:01 chris is build the interface first
17:01 kados yep
17:01 well I'd do it with the MARC editor
17:02 chris which can be used in head/dev_week/and 2.2
17:02 yep
17:02 kados what you want is to overlay stuff
17:02 and prompt the user to approve changes
17:02 chris yeah
17:02 kados I think
17:02 chris i was thinking side by side
17:02 kados yep
17:02 that's how a lot of marc editors work
17:02 but while you're at it ...
17:03 might be a good time to look at the whole MARC editor
17:03 cuz you're talking pretty major interface change
17:03 chris i was actually thinking a step before the editor
17:03 kados that's the kind of thing you'd want to do in xul or something
17:03 also ...
17:03 chris because i want it to be able to work with opencataloguer too
17:04 kados Evergreen's already done it
17:04 so I bet you could nab their editor
17:04 wrap it around a Koha API
17:04 yea ... ++ on the opencataloger stuff
17:04 this community has always had good ideas
17:04 we just don't have the resources to make them happen :-)
17:04 chris ie i was planning to do something not tied to the existing marc editor
17:04 yep
17:05 kados cool
17:05 one thing that's gonna be important
17:05 chris should get better now i hope
17:05 kados is using the frameworks
17:05 chris yep
17:05 kados in fact, what I"ve seen of opencataloger thusfar is pretty impressive
17:05 chris i think im gonna do a 2 setp
17:05 or 2 level process
17:05 kados best bet might just be talking to paul/toins about that
17:06 chris yeah was gonna mail the list, was just checking it didnt exisit already
17:06 kados right
17:06 we need a good spec to work from
17:06 chris my 2 levels would be
17:06 1/ simply replacing a holding record
17:07 where you just choose the new marc, choose the little placeholder and overwrite it totally, then add your barcode
17:07 2/ fixing up full records with other newer/better full records
17:07 thats where the side by side thing comes in
17:08 kados right
17:08 that makes sense
17:09 chris cool, ill think some more and do a post to the list, and maybe a page on the wiki
17:10 just didnt want to spend the day thinking about things,and then find out someone has done it already :)
17:10 i will look at evergreen, and i will pester toins and paul
17:10 kados still haven't head back from gary
17:11 chris bummer
17:11 i got the list of reserves showing in the opac
17:11 kados sweet
17:11 chris so its really just the issuing/returning and the postage labels to go
17:12 plus a little change to placing reserves if gary says thats nessecary
17:12 kados sweet
17:12 right
17:12 chris so less than a days work
17:13 thd chris: what is "the little placeholder"?
17:13 chris just a little record you make when you placing an order in full acquitisions
17:13 usually just title, author, itemtype, isbn
17:14 then when your order arrives
17:15 you want to either catalogue the rest, or hopefully use an existing record
17:16 its a process libraries like to use ... i think mostly cos what you order isnt always what you get, so they dont like to spend a bunch of time cataloguing it before they have it in their actual hand
17:17 at least in nz thats the case
17:27 kados http://zoomopac.liblime.com/cg[…]ail.pl?bib=136919
17:27 in related links
17:27 first link works ... second one doesn't :(
17:36 chris right
02:35 osmoze hello
07:57 kados thd: you awake?
07:57 thd: http://zoomopac.liblime.com
07:57 thd: I believe I've done as you asked with the faceted searches ...
07:57 thd: need you to verify it's working as expected
07:58 faceted results I mean of course :-)
09:42 thd kados: that is very nice but not what I meant recently.
09:43 kados: that is close to what I had suggested as the least that might be done for subject displays.
09:45 kados: A basic issue with what you have done is that I do not see how the user has a browse path to any fully completed subject assignment with all subdivisions.
09:46 kados: a fully specified subject with all subdivisions should not disappear from a result set list.
09:47 kados: are you there?
09:48 kados thd: helly
09:48 hello even :-)
09:48 thd kados: do you see above?
09:48 kados thd: the browse path _is_ to a fully completed subject assignment with all subdivisions
09:48 it may be that NPl doesn't have any of those though :-)
09:49 for instance
09:49 do a search on 'neal stephenson'
09:49 thd kados: I mean you have made the full form which was actually found disappear
09:49 kados and expand the first faceted subject result
09:49 please expand a bit
09:51 thd yes
09:52 the result number has disappeared past the table width
09:52 kados ahh ... yes, of course, that is a display issue
09:53 but just a minor detail
09:53 what I'm asking is whether the actual results are allowing you to browse the subjects as you wanted
09:53 thd kados: but I see it would have had a '0' for the catalogue as a whole.
09:54 kados ?
09:54 ahh yes
09:54 that is another one of those search anomolies
09:54 I'm working with ID to resolve it
09:54 (sorry :-))
09:55 (bad example ) :-)
09:55 thd kados: I had thought that was the point you were trying to make until I reminded myself that you did not have the data to make that point :)
09:56 kados hehe
09:58 thd kados: I see in this case there are now subdivided subjects really for the result set or I have not found them yet
09:59 kados: I found one 650 $a Kings and rulers $v Fiction
10:00 oops
10:00 kados it still needs some work of course :-)
10:00 thd kados: I found one 650 $a Kings and rulers $x Succession $v Fiction
10:00 kados you did?
10:00 where?
10:00 (that should show up ... if it's not I have a coding bug somewhere)
10:01 thd kados: yes, the system of the world has a few
10:01 kados k, I'll look
10:02 system of the world doesn't
10:02 ahh
10:02 I see
10:02 hmmm ... no idea why we're losing that $x
10:03 thd kados: however, su = kings and rulers succession fiction fails
10:04 kados yes, that is a zebra problem
10:04 I've been working with ID to fix it
10:04 thd: if you expand 'Kings and rulers' it shows 'Kings and rules Succession Fiction'
10:05 thd: Isn't that correct? or do you want it to expand even further to allow:
10:05 kings and rulers
10:05  -> Succession
10:05       -> Fiction
10:05 ?
10:06 thd kados: one point is that it should have both
10:06 kados this works btw:
10:06 su="kings and rulers" and su="Succession" and su="fiction"
10:06 so with some clever trickery we can trick zebra into giving us what we want with those links
10:07 thd kados: have you had any response about why the other is not working?
10:07 kados not yet
10:07 thd kados: so let me explain what I think would be good here
10:08 kados: use the example of Kings and Rulers--Succession--Fiction [1]
10:09 kados: you should have one link or linking mechanism that specifies that fully
10:10 kados: another set of links should be ...
10:11 Kings and rulers [all with at least this]
10:12 -> Succession [all with this and above]
10:12 -> [choose another subdivision]
10:13 -> Fiction [all with this and above]
10:14 -> [choose another form subdivision]
10:16 kados: to which you could add options for choosing other free floating subdivisions, geographic, and chronological, etc.
10:16 kados hehe
10:16 you're nuts :-)
10:16 I'm not sure I can add the '[choose another subdivision]' bits
10:16 but I can probably do the rest
10:17 thd: one thing I don't understand is
10:17 thd: how are the subject fields supposed to be ordered?
10:17 thd: or should they always follow the order in which they appear?
10:17 thd kados: why could you not do the choose another parts you only need a table of which subdivisions go together
10:17 kados thd: (restricting ourselves to the 650 for now)
10:18 I don't have that table
10:18 thd kados: there is a usual preferred order which may have changed over time
10:19 kados: you may not have that table but you could create one from bibliographic records alone without authorities
10:19 kados even if it works in 99% of the cases it would greatly simplify my task
10:19 first I'd like to get this working without authorities
10:21 thd kados: you could create the previous table without authorities and you would primarily miss the ability to search without knowing what the correct authorised form is.
10:23 kados: so the usual order at present for 650 seems to be $a, $b, $c, $d, $e, $x $z, $y, $v
10:24 kados thanks
10:25 should they nest that way too?
10:25 $a
10:25 |-$b
10:25   |-$c
10:25       |-$d
10:25 etc?
10:25 thd kados: if you look at my ISBD system preference, the usual order is specified for many fields when that was not overridden by an ISBD specified order
10:26 kados: no nesting is a little of a misnomer
10:26 kados hmmm ... well the faceted results nest :(
10:27 thd kados: the nesting is not continuous
10:28 kados so there are only two levels?
10:28 subject
10:28  subdivision
10:28 ?
10:28 thd kados: and I have found a mistake in my ordering for one of those rare ones
10:28 kados how do I represent the subjects properly in a tree?
10:28 Kings and rulers
10:29 |-Succession
10:29 |-Fiction
10:29 ?
10:29 (wehre the 'Succession entry links to both 'Kings and rulers' and 'Succession')
10:29 (and the 'Fiction' entry links to all three)
10:30 ?
10:30 thd kados: I spent about ten hours one day on the subject subdivisions and never checked anything except for some annoying problems about the backwards design and the fact that it was a big improvement over the previous suggestion for a value which had ignored the ISBD rules
10:31 s/subject subdivisions/my ISBD preference/
10:32 kados: the hierarchical order with the correction should be something like ,,,
10:32 650 $a
10:35    ->$b # I may have been wrong about my mistake and had it right the first time
10:35 [that should be only one level of indentation above]
10:36    $c
10:36    ->$d
10:38 kados: let me start again I would have typed this already using vim
10:38 kados ok :-)
10:41 thd kados: ok now we have with the possible uncertainty of the place for $b-$e which are too uncommon for me to know well.
10:41 kados let's stick with the common ones
10:41 thd 650 $a
10:41    $b
10:41    $c
10:41    $d
10:41    $e
10:41    $x
10:41        $z
10:41        $y
10:41        $v
10:42 kados hmmm ...
10:42 I don't understand how $x can be parallel with $a
10:42 should not $a be the root of the tree
10:42 $a
10:42  $b
10:42 thd yes it is
10:42 kados  $c
10:42  $d
10:42  $e
10:43  $x
10:43 thd $a is not indented
10:43 kados    $z
10:43   $y
10:43 oops
10:43 ok, I think I get it
10:43 so our example would be:
10:43 Kings and rulers
10:43  ->Succession
10:43  ->Fuction
10:44 Ficting even :-)
10:44 thd kados: you have it I started with the 650 and it loaded as if everything were on the same level
10:44 kados ok
10:44 so there are three levels only
10:44 root, branch1, branch2
10:44 thd kados: yes only three levels
10:44 kados and branch2 should or should not be automatically expanded?
10:44 and should it be a branch directly from $a or from one of the others?
10:45 ie, how are $z, $y, $v related to $a?
10:46 thd kados: you could have as the root node a fully subdivided subject which when the link is clicked will search the whole thing
10:46 kados (are they branches off of $x?
10:46 I could do it several ways ...
10:46 I could attempt to represent the hierarchy as it exists
10:47 thd kados: expanding the fully subdivided subject should bring up the hierarchy starting with $a
10:47 kados and have each successive link automatically include the previous link when clicking
10:47 but I still need to know where $z,$y,$v branch from ... is it $a or $x?
10:49 thd kados:  $z $y and $v branch from $a if $x is absent otherwise they branch from $x
10:49 kados ok, great!
10:49 now I'll get coding :-)
10:51 thd kados: I am uncertain whether $b-$e branch from $a or elsewhere depending on how they are encoded but would say that the branch from whichever more important subfield preceeds them
10:52 kados for now I will only deal with a, x, and zyv
10:52 we can add the others anytime once we prove it's working properly
10:52 with those
10:53 thd kados: don't go yet: automatically including the previous elements is the week way that many implementations incorrectly imagine the hierarchy
10:53 kados ?
10:53 please explain what you mean
10:54 thd kados: much better would be to have a checkbox next to each element which is checked by default
10:56 kados: the user can remove a particular subdivision by unchecking the adjacent checkbox
10:58 kados: sometimes removing the term would be good and sometimes removing the whole subdivision containing two repeated $z for example would be good
10:59 kados: do you understand why automatically including everything above one point is weak?
11:01 kados thd: thd I'm still here
11:01 thd kados: do you understand why automatically including everything above one point is weak?
11:01 kados i do understand
11:02 and that part is purely interface
11:02 the tricky part for me is how to code the hierarchy so it comes out correctly
11:02 thd kados: most systems naively assume that everything is one hierarchy at successively lower levels
11:04 kados: and those are the ones which actually do something interesting as opposed to the ones that give you the whole search or the individual subdivisons without connecting them to the rest as you had done recently
11:06 kados: what is difficult about a three level hierarchy even if the levels do change in the absence of $x.
11:06 kados thd: you code it :-)
11:06 thd: I'm not a programmer :-)
11:06 thd kados: you need some loops and a test for $x
11:07 kados yes, i know how now that I know how it's supposed to be
11:07 should be done in about an hour
11:07 thd kados: what are you, if you are not an x?
11:09 kados: do not say that you are a non-x even if that would be implied logically we know that two valued logic will not work well for answering what you are.
11:12 kados thd: ok ... $x is correctly nested now I think
11:12 and searches should work too
11:15 thd kados: the searches search for the subdivision without including other parts of the hierarchy
11:15 kados not the ones I've tried
11:15 use Kings and rulers as an example
11:15 Succession is the subdivision
11:15 and it searches both
11:15 when you click on subdivision
11:16 thd kados: I chose one that is impossible Kings and rulers--Operating systems
11:16 kados ?
11:16 I don't think Operating systems is a branch of Kings and rulers
11:17 thd kados: where did Operating systems come from as a subdivision of Kings and rulers?
11:17 kados Operating systems is a $a
11:17 it's not
11:17 how did you get that impression?
11:17 only the items inside the folder are subdivisions of a given subject
11:18 I agree the root subjects should probably all have the same icon
11:18 do you understand now?
11:19 thd kados: most 650 $a can also be $x although they need a logical combination relating to the material catalogued
11:20 kados ok, now I'll do $v
11:20 as our next test
11:22 thd kados: where do the subjects outside the folder originate?
11:23 kados they are $a
11:23 parallel to the folder
11:23 (root elements with no subdivisions)
11:23 thd kados: in the same result set?
11:23 kados yes
11:24 all of the faceted results are pulled from the first page or results only (for now)
11:24 s/or/of/
11:25 thd kados: so I was confused because they did not have a top level icon which I imagined would always be a folder
11:26 kados yes, I see that now
11:26 the icons must be changed to reflect more libraryish things anyway :-)
11:28 thd kados: my own experiments did not have Icons so I did not have any icons to confuse me :)
11:30 kados :-)
11:57 thd: I've removed the icons for now
11:57 thd: hopefull it's less confusing
11:58 (though the font will need to be changed I think ...)
11:58 thd kados: that looks like my secret experiment except for no checkboxes
11:58 kados hehe
11:58 (almost)
11:58 still need to add the other sub subdivisions
11:59 well ... starting with $v

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

koha1