IRC log for #koha, 2005-10-06

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

All times shown according to UTC.

Time Nick Message
11:02 thd paul: At the bottom of Encode::Arabic::Buckwalter POD is a comment "Perl is also designed to make the easy jobs not that easy ;)"
11:29 kados hi everyeon
11:29 everyone even ;-)
11:29 paul hi kados
11:29 kados paul: do your libraries have a report to show 'items on hold'?
11:30 paul: how do they know if a patron makes a reserve?
11:33 paul: you around?
11:36 paul kados, was on phone
11:37 reserves are usually done only on issued items. so when the book is returned, the librarian is warned.
11:37 however, I agree we should integrate into Koha a report like the php one you use in NPL
11:41 thd paul: This Arabic question looks very tricky.  I see the problem more fully now.  The problem as your English examples showed relates to prefix and suffix characters attached to logical words that have a separate meaning as semantic words.  Arabic is too inflected to be easily indexed with stop words.
11:42 paul: The prefix character for 'the' may have other meanings depending upon the base word to which it is attached.
11:43 paul: There seems to be no widely accepted word stemming solution for Arabic.
11:48 kados paul: thanks
11:48 paul: one more small question
11:48 paul: :-)
11:48 paul yes ?
11:48 kados paul: if you have a moment
11:49 paul: I have noticed that 'request.pl' does not use permissions correctly
11:49 paul: it requires 'paramaters' or 'superlibrarian' privs
11:49 paul: but it should work with only 'reserveforothers'
11:50 paul right.
11:50 kados paul: is this the problem line: flagsrequired => {parameters => 1},
11:50 paul fixed
11:50 kados paul: ahh ... good
11:50 paul: thanks
11:50 paul yes, just replace by reserveforothers
11:50 kados paul: I'll check CVS version then
11:50 k ... just making sure ;-)
11:50 paul (don't do it immediatly, i'm commiting now ;-)
11:54 kados (ok ... I manually change it ;-))
11:57 paul: another bug found for permissions you might want to know about
11:57 paul: member-flags.pl
11:59 paul: flagsrequired => {borrowers => 1}, should be
11:59          flagsrequired => {permissions => 1},
12:00 paul: I'm happy to commit the fix unless you already have
12:00 paul: (or if you plan to ;-))
12:01 paul i'll commit on 2.2
12:02 thd paul: Having to determine the meaning of a base word in order to know whether a prefix should be a stop word or not is a significant amount of work for an indexing program.  This does seem like a fun problem though now that I understand it.  Too bad I do not know any Arabic at all :)
12:02 kados: are you still there?
12:02 kados thd: yea
12:02 thd: I've spoken with indexdata about arabic language
12:03 thd kados: yes
12:03 kados thd: zebra is capable of doing all the stuff that would need to be done
12:03 thd: but there would need to be extensive configuration files written
12:04 thd kados: yes, I saw documentation about the configuration files.
12:04 kados: Do you know any Arabic?
12:04 kados thd: a few greetings ;-)
12:05 thd: and as a linguist I know a few odd characteristics of the language
12:05 thd: but no, i don't _know_ arabic
12:06 thd kados: What about the problem where the 'the' prefix character changes its meaning to another word depending upon the base word to which it is attached
12:06 ?
12:06 kados yea ... so that's no problem in zebra
12:06 seb explained it to me
12:07 let me see ...
12:07 I think what happens is that you can setup envoronments in which a prefix is to be ignored
12:08 thd kados: Does not that require looking up every base word in a dictionary to determine the applicable part of speech or meaning for the prefix context to use the as a stop word?
12:09 kados no ... you just specify the exact environments in the config gile
12:09 file even
12:10 thd kados: So, by your comment, Zebra cannot cope reliably with 'the' in Arabic?
12:13 kados: I would expect that very little software manages to identify 'the' in Arabic reliably.
12:13 paul kados : about members-flags.pl => i have already permissions=> 1 on my cvs copy
12:14 thd kados: Have libraries ever asked you about title sorting without using leading articles?
12:16 Sylvain hdl ? Quand on fait la liste des plus gros utilisateurs, pourquoi est-ce qu'on ne prend en compte que les prêts "retournés" ?
12:17 ou paul peut ête ?
12:17 r
12:17 paul non, hdl
12:17 thd kados: are you still there?
12:17 paul ;-)
12:17 Sylvain ok :)
12:17 paul (joshua is never away, but often not 100% here ;-) )
12:18 hdl Je ne connais pas assez la pratique des bibliothèques pour savoir ce qui est le plus pertinent.
12:18 thd paul: like you as well, when you are not marked paul_away? :)
12:18 paul ;-)
12:19 hdl Pour moi, je pensais que les prêts en cours n'avaient pas à être dans les stats.
12:19 Sylvain.
12:19 thd paul: I would settle for 50% of you or kados at times if that were possible :)
12:19 hdl Mais ce n'est pas trop difficille de faire le plus.
12:20 large.
12:20 Sylvain oui pas compliqué hdl :)
12:20 c'est juste sur le plan fonctionnel ça me paraissait bizarre
12:21 et c'est aussi que l'école des mines me demande pourquoi il ne voit pas d'utilisateurs dans cette liste ;)
12:21 hdl j'essaierai de leur demander ce qu'ils en pensent et le changerais le cas échéant
12:22 de la même manière pour les prêts au niveau des dates, le fait qu'on soit sur date < et non <= pour la date de fin les a troublé
12:22 c'est vrai que c'est moins intuitif je pense
12:24 hdl Merci pour ces remarques. J'en tiendrai compte par la suite.
12:25 Sylvain je demandé confirmation aux bibliothécaires concernés qu'ils sont bien de cet avis avant de faire la modif.
13:32 thd kados: are you there?
14:05 kados thd: yea ... what's up?
14:07 thd kados: Did you receive the INEO and SAN contributed roadmap 3.0 features in their current English state?  Did you have them in French and pass them through the babelfish? :)
14:09 kados: Or did they add them to the developer's wiki themselves, as they are now?
14:09 kados that's a secret ;-)
14:10 thd kados: Why secret :) ?
14:10 kados I could tell you but then I'd have to kill you ;-)
14:11 thd kados: so tell now, kill very much later :)
14:11 kados hehe
14:11 I added them after very rough translation with babelfish
14:12 thd kados: They have no one who writes English?
14:12 kados they could use better translations
14:12 thd: right
14:12 thd: feel free to translate ;-)
14:13 thd kados: To attempt a translation I would need the original text.
14:13 kados yea ... and I can't give that to you unfortunately
14:13 because the anonymous folks who are planning to do that dev work have asked to remain anonymous
14:13 thd kados: Do you no longer have it?
14:13 kados until they have committed to doing it for sure
14:14 and are ready to make an announcement ;-)
14:15 thd kados: Is it wrong for me to send them the message I have almost finished drafting about interpreting these features?
14:15 kados please send it to me ... I'll pass it on
14:17 thd kados: perhaps you could supply me with the original forms of the features without names attached.  I could then ask the interpretation question much better if I still had one.
14:20 kados: yes, no, or maybe?  for my suggestion above
14:25 kados maybe ;-)
15:50 chris morning
15:55 owen Hey chris, how's it going?
15:55 chris doing good owen, busy ... but isnt everyone :) other than that good
15:56 owen Yeah, busy here too
15:56 You hear much in the news over there abotu avian flu?  It's a little closer to home there than here
15:58 chris there is a bit of hype
15:58 but very little in the way of facts
15:59 owen We're starting to see some 'not if but when' articles in the mainstream here, but none of the terrifying death tolls predicted in other more obscure sources
15:59 chris mostly it seems to be "if this mutates and becomes a killer virus, it will be a killer virus"
15:59 which isnt rocket science :)
16:00 crowded cities .. those will be the places to stay out of
16:02 i suspect where you guys are would be quite a safe place :)
16:03 got plans for xmas?
16:03 owen Not a trip to Bali like my wife wanted
16:04 chris good thinking
16:05 bali is almost to nz .. just come here :)
16:05 though, not this xmas, cos ill be pittsburgh so wont be here to show you round :)
16:05 thd chris: Then, I am in one of the most dangerous places on earth :)
16:06 chris shanghai? ... or more likely new york city?
16:06 owen thd, are you in one of the most crowded cities on earth?
16:06 chris ahh
16:07 i have a soft spot for NYC
16:08 bummer
16:09 thd chris: It is bad if on the subway everyone in the same car is coughing or otherwise manifesting some illness.
16:09 chris: Then it is unavoidable.
16:09 kados bradl: hey chris
16:09 oops ;-)
16:10 chris heya kados
16:10 kados how's laurel?
16:11 chris shes doing much better
16:11 had a good nights sleep last night
16:11 kados sweet
16:12 thd sleep is very important
00:00 thd-away chris: Is there any reason that branches.branchcode cannot be changed from varchar(4) to varchar(8) or something to support more flexible naming of virtual branches?
00:03 thd chris: My idea is to have virtual branch codes better related to the branches to which they correspond.
00:05 chris: So, a virtual branch code for the branch MAIN might be MAINREPA or even MAINREPAIR.
00:06 chris: Would such a change in the size of branches.branchcode break anything?
00:08 rach Chris is out at the moment
00:09 thd rach: Do you know if my idea would break something?
00:11 rach: Do branch categories do anything meaningful?
00:11 s/meaningful/functional/
00:11 rach I think that's actually a question for paul, he put them in
00:11 I'm not sure what he uses them for
00:11 chris yes they do
00:11 im back
00:11 rach oh look he's back :-0"
00:11 chris and i think i may have put those in
00:12 rach shows what I know :-)
00:12 thd chris: What do branch categories do then?
00:13 chris well often ppl only want certain branches to show
00:13 for instance at hlt, only branches of type IS show at circulations
00:14 thd chris: how are the others hidden?
00:15 chris you just tell get branches to only get branches that belong to a certain category
00:15 ahh now i remember
00:15 i had category as a column
00:15 thd chris: With a custom modification of the code?
00:15 chris paul made it a seperate table, which allows a branch to belong to more than one category (which is handy)
00:16 soon no thd, when i put the fix back in
00:16 it used to work, but got busted when they got made a seperate table
00:16 ive fixed it for 2.2
00:17 well the next release after this anyway
00:17 thd chris: So there should be a brarches.hidebranch column?
00:17 chris no
00:17 you should just set a variable in the template
00:18 which is used to determine what category should be showing
00:18 thd chris: Where is the variable stored?
00:18 chris in the template
00:18 and passed as input
00:19 eg <input type=hidden name="branchcategory" value="IS">
00:19 by default it will just show you all branches
00:20 its not so much hidden branches, just some should only show on certain pages
00:20 it doesnt hurt to have the others there
00:20 thd chris: How can that work without storing the invisible branch value somewhere?
00:21 chris quite easily
00:21 its not an invisible branch
00:21 is is category
00:21 is a
00:22 thd chris: Invisible in OPAC category?
00:22 chris no
00:22 for HLT IS = Issues
00:22 you could have ZQ = Can circulate books
00:22 or anything
00:23 it doesnt matter what the actual code is
00:23 getbranches($type)
00:23 will find all branches belonging to $type
00:23 getbranches() will find all branches
00:24 thd chris: How does getbranches() know which not to display in the OPAC?
00:25 chris 2 ways, you can put a hidden input in the form that calls whatever page shows branches
00:25 to tell it
00:25 or you can just whack it in the code
00:27 opac-renew.pl is the only script that uses it
00:27 sorry, opac-reserve.pl
00:29 thd chris: Yet the status of hidden must be stored somewhere after the original branch category creation form has been submitted and expired?
00:29 chris why?
00:30 next time you submit the form you will be passing it to the script
00:30 thd chris: I hope the user only has to define branch categories once :)
00:31 chris what user?
00:31 thd chris: superlibrarian
00:31 chris they can define them once, none, 20 times
00:31 its entirely up to them
00:32 thd chris: I mean every time that getbranches() runs there is no need to resubmit a branch categories form marking a category hidden :)
00:33 chris no no no
00:33 you dont mark a category hidden
00:33 you tell getbranches what category you want to show
00:33 eg
00:34 http://opac.rangitikei.katipo.[…]koha/opac-main.pl
00:34 thd chris: You make a custom modification of the script?
00:34 chris no
00:34 you make a modification of the template of script which calls the script
00:34 thats showing all branches
00:35 thd chris: sorry template, yes.
00:38 chris: would it not be easier to define branchcategories.hidden, branchcategories.hiddenopactemplate, etc. ?
00:38 rach I think the short answer is "yes"  If you use the same branch categories say as we do, and our templates, then you'll get the same branches showing in different places as we do
00:39 you could if things were more stable - but there are a lot of different places you want different things to show
00:40 so you want different categories to show in isuues, returns, transfers, different places in the opac as well I think
00:40 thd rach: There would seem to be an advantage for default templates to support maximally useful default behaviour.
00:41 rach so this isnt' necessarily how it will always be, but we haven't necessarily achieved consesus on what you would show where
00:41 I suspect it is easier (for people who have been doing the development) not to try and abstract them out yet
00:42 at the moment it's very flexible if you're working with a bigger library
00:42 because realistically you're in
00:42 thd chris: I originally wanted to ask the following.  Is there any reason that branches.branchcode cannot be changed from varchar(4) to varchar(8) or something to support more flexible naming of virtual branches?
00:42 rach "touching
00:42 templates anyway - and it makes it quick and easy to get it right
00:44 would naming them differently imply then that they were something they aren't?
00:44 like if you give them a "plain english" name, would that fool you into thinking that they were going to display?
00:46 thd rach: Never mind my suggestion.  Templates may well be the right place to implement this.  However, template configuration features need their own documentation that does not depend upon the prospective adopter hunting for comments in every template :)
00:46 rach yeah, I'm not arguing that if all you want to change is this one thing it's a royal pain to work it out
00:47 the branchcode thing may be a hangover from our original development
00:47 i'm pretty sure the codes from HLT's old system were all 4
00:47 thd rach: I have to know that there is something there designated for configuration unless I reinvent a new method every time.
00:47 chris thd: its not hidden its exactly the opposite from hidden
00:48 thd chris: Sorry, what is not hidden?
00:49 chris branches of a certain category
00:49 you can tell it to only show branches of a certain category
00:49 of course a branch can belong to mulitple categories
00:50 thd chris: I see, that is opposite, 'non-hidden' :)
00:50 chris: visible
00:50 chris yep, by default koha just displays all, with a bit of tweaking its easy to make it display different ones in different places
00:50 depending on what you want
00:51 rach but that is the sort of thing that isn't very well documented for the DIYer I suspect
00:52 if you're thinking that you can get a setup like NPL or HLT without "touching" the templates I guess we should disabuse people of that notion
00:53 thd rach: 4 character standard library names are universal I was just trying to find a way for using a mnemonic in a branch code for a virtual branch with the same mnemonic for associating the virtual branch with the branch.
00:55 chris in what way rach?
00:55 rach um that sounds like cheating :-)
00:55 sorry 2 thoughts going
00:55 thd rach: Many sophisticated proprietary systems may have extensive configuration options for the user who is not allowed to touch the closed code.
00:55 chris then again they may not
00:55 rach um I think koha differes from commercial systems in that you are expected to alter the templates directly
00:56 chris or are in fact allowed to
00:56 rach rather than do everything through extensive configuration options
00:57 thd rach: However, proprietary systems can leave out user options for special expensive custom configuration that the user has no ability to configure himself because of no access to the code.
00:57 rach and we tend to move variables to configuration options only when we've got a clear understanding of what htey should do
00:57 so in the begining htere where almost no configuration options, each release there are more
00:57 chris thd: exactly thats what i was trying to say :)
00:58 thd rach: yes, of course, you should have a clear idea from experience about what options people want rather than turning the system administration into an airline instrument panel :)
01:01 rach: system behaviour has to conform to something useful practise model learnt from experience otherwise the user will crash Koha at first outing and look elsewhere for an ILS :)
01:02 rach: Crashing is a little too easy with MARC 21 Koha.  MARC 21 Koha has to catch up to UNIMARC Koha with basic safety defaults.
01:04 chris: Would any change in the size of branches.branchcode break anything?
01:06 chris nope that should be safe enough
01:06 ok, ive gotta go sort out some dinner
01:07 thd chris: There might be a simpler means of implementing my idea with branch categories but I cannot think of that well enough yet.
01:07 chris: happy nourishment
04:23 Malin hi
04:24 paul who is Malin ?
04:24 Malin guess koha 2.2.4 isn't released yet?
04:24 paul should be in 2 days
04:24 Malin cool
04:24 paul (maybe monday)
04:24 Malin paul, FYI: i'm a librarian from Poland :-)
04:25 paul a polish librarian using Koha ?
04:25 Malin no, not using koha
04:25 i'll be just writing about it
04:26 although there are some libraries in Poland using koha afaik
04:30 ok, gotta go
04:30 thanks for the info
04:30 see ya :-)
10:03 paul hdl_away :  bonne nouvelle dans ta boite mail ;-)
10:35 Sylvain hi all !
10:37 paul j'ai une petite question (comme à chaque fois, c'est ce qui me fait penser à lancer irc :)). A l'ENSMP ils m'ont fait une remarque pas idiote c'est que pour les prêts, ça serait bien que le titre de la page soit différent pour les sorties et les retours. T'en penses quoi ? Pour, Contre, Sans opinion ? (ça oblige à rajouter une un xx-top dans les includes)
10:37 paul plutôt une bonne idée.
10:38 mais comme on reprend toute l'ergonomie dans la 3.0, ca va peut être etre inutile
10:38 comme remarque
10:38 Sylvain ok.
10:38 je vais peut être me contenter de leur faire la modif chez eux alors

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

koha1