IRC log for #koha, 2007-03-20

All times shown according to UTC.

Time S Nick Message
17:56 kados hey owen
17:56 owen Hey kados
04:08 hdl hi world
04:10 chris hi hdl
04:11 hdl hi chris.
04:11 hdl paul is around.
04:11 chris hi paul :)
04:11 paul hello world.
04:11 paul i confirm i'm here ;-)
04:17 chris i see france won the 6 nations
04:18 paul yep. last minute try, that gave us enough points to win...
04:18 chris :)
04:21 chris hi toins
04:21 toins hello chris
09:29 kados morning all
09:29 toins hello kados
09:29 kados toins: !
09:30 kados toins: good to read you
09:30 toins kados, how are you ?
09:30 kados toins: will you be applying to Google Summer of Code?
09:31 toins kados, i'll be really happy & proud to participate to the ggl soc...
09:31 kados excellent
09:31 kados do you think open cataloguer or opencontent is a better project?
09:31 toins but i don't have enough info
09:31 toins opencontent ?
09:31 kados oops
09:31 kados openshare tags
09:32 kados paul's idea
09:32 kados that's what I meant :-)
09:32 toins ah ok
09:32 toins hehe
09:32 hdl morning kados.
09:32 kados morning hdl
09:32 toins i think opensharetags could be a good project, yes.
09:32 kados hdl: I have set up a git repo
09:32 kados hdl: fyi ... just testing it out
09:33 kados hdl: it seems nice so far
09:33 kados toins: what is the status of opencataloger?
09:33 toins kados, a first version will be released very soon
09:34 kados toins: are you the primary author?
09:34 toins kados, yes
09:35 toins the only one in fact..
09:35 kados toins: are you happy to work on opencataloger? or are you tired of it? :-)
09:35 toins kados, i'm never tired !!
09:36 kados hdl: I have been integrating the YUI libraries into Koha with the following result:
09:36 kados http://kados.org/stuff/yui_borrowers.png
09:36 toins and i'm happy to work on opencat
09:36 kados hdl: we need to decide hwo to approach templates in 3.0
09:36 kados toins: cool
09:36 hdl paul
09:37 kados is paul back?
09:37 toins hdl, paul not around yet
09:37 hdl kados : paul should be back soon
09:37 kados *nod*
09:37 kados hdl: look at the link above for a screenshot
09:37 hdl what you showed me was pretty.
09:37 hdl I saw.
09:37 kados cool
09:38 hdl But I am not keen on big +
09:38 kados hehe
09:39 kados toins: http://wiki.liblime.com/doku.p[…]summerofcodeideas
09:39 hdl Maybe a sign just as create new file on Kate would be better.
09:39 kados toins: look at the first paragraph
09:39 toins ok
09:39 kados hdl: Kate?
09:39 kados hdl: do you have a image in mind?
09:40 kados hdl: I can put any image of a similar size behind the buttons
09:42 kados for instance, the button is represented as: <span class="linkbuttons" id="linkbutton1"><span class="first-child"><a href="/cgi-bin/koha/members/membere​ntry.pl?op=add&amp;guarantorid=<!-- TMPL_VAR NAME="borrowernumber" -->&amp;category_type=C">Add child</a></span></span>
09:43 hdl Not so clean.
09:44 kados right
09:44 kados so maybe it should be a /liblime/ template and we should leave prog alone
09:44 kados but ... the problem is that prog is not so clean either :-)
09:45 hdl I agree.
09:46 hdl But I'd rather not have to many nested span declarations without any content. I donot find it handy on long term dev.
09:46 hdl Don't you think ?
09:47 kados I agree
09:50 toins kados, my main problem for ggl soc is that i'm not available during august....
09:50 toins and if we need to respect this deadline => http://code.google.com/support[…]60325&topic=10729 it's impossible...
09:51 toins :-/
09:51 kados bummer
09:55 toins kados, is it not possible to begin in june and to finish in july ?
09:55 kados toins: it's fine with me :-)
09:56 toins kados, so do you think it's possible ?
09:56 kados toins: absolutely
09:56 kados toins: but we should check how it will work
09:56 toins ok
10:02 kados toins: actually, since you can work before the deadline it will be fine
10:04 toins kados, good
10:04 kados toins: so feel free to apply :-)
10:05 toins yeah !
10:06 toins kados, what i need to do exactly ?
10:09 toins kados ?
10:09 dewey well, kados is helping us LibLime folk with something at the moment
10:09 kados toins: http://code.google.com/support[…]ic.py?topic=10442
10:09 kados toins: you have to submit a application
10:10 toins kados, so you prefer opencataloger than opensharetags ?
10:10 kados toins: opencataloger please :-)
10:10 toins ok
10:10 kados toins++
10:11 toins kados, may i work with other friends ?
10:11 kados yes!
10:11 kados your friends can apply too!
10:11 toins gus, here is one of them
10:11 kados you can all get paid :-)
10:11 kados hi gus
10:11 toins hehe ... good !
10:11 kados gus: you're a student too?
10:11 gus hi
10:11 dewey bonjour, gus
10:11 gus yes with toins
10:12 kados excellent
10:12 kados so toins and gus can apply separately do work on opencataloger
10:13 toins why separately ?
10:13 kados so that you can both get paid separately :-)
10:13 kados even though you will work on the same project
10:13 toins ahh...
10:14 kados yes
10:14 kados gus too :-)
10:15 gus yes !
10:16 kados toins: do you remember sara?
10:16 kados toins: she will apply too
10:16 toins kados, yes
10:16 toins ok !
10:16 kados toins: maybe she could help with opencataloger too
10:17 toins excellent
10:17 kados it will be her first programming project other than school projects
10:17 kados paul: hi :-)
10:17 kados toins: make sure to read carefully the GSoC specs and submit an application asap
10:17 toins kados, ok
10:17 kados toins: because the deadline may be cut short if many students apply
10:18 kados toins: this happened with the deadline for organizations
10:18 toins i read it tonight
10:18 kados toins: 2/3 were rejected
10:18 kados toins: so liblime is lucky to have been accepted!
10:18 toins yep !
10:19 toins why liblime has been selected rather than other ?
10:20 paul kados : "I have been integrating the YUI libraries into Koha with the following result'
10:20 paul what are "YUI libraries" ?
10:20 paul (hovewer, it's really neat, I agree.
10:20 rch paul: Yahoo User Interface
10:20 paul just one note : "+ add adult" buttons are text buttons of course ? (except for the +)
10:21 paul otherwise it's untranslatable, and we have a problem
10:21 rch they have released all their UI code under OS license
10:22 kados paul: yes, they are fully text buttons
10:22 kados paul: css is used to style them to look like buttons
10:22 kados paul: and the + is a background image
10:23 kados paul: so fully translatable
10:23 paul YUI : ++
10:23 paul why is prog not reliable for this screen
10:23 kados but the question is, should it be prog?
10:23 kados or shoudl it be 'liblime'?
10:23 paul what are the diffs for tmpl ?
10:23 paul between prog and this screenshot ?
10:25 kados there is some extra markup in that screenshot
10:25 kados for instance, the buttons are represented as:
10:25 kados <span class="linkbuttons" id="linkbutton1"><span class="first-child"><a href="/cgi-bin/koha/members/membere​ntry.pl?op=add&amp;guarantorid=<!-- TMPL_VAR NAME="borrowernumber" -->&amp;category_type=C">Add child</a></span></span>
10:25 kados however, note that prog templates are quite 'dirty' too
10:25 kados at least the members section that I have been looking at this weekend
10:26 kados paul: look at the moremember and memberentry* tmpls for examples of very 'dirty' templates
10:27 paul what is dirty according to you for moremember ?
10:28 paul it seems very "regular" to me.
10:28 kados <p><label></label></p>
10:28 paul why is it a problem for you ?
10:29 paul its not valid html ?
10:29 kados it's not semantic
10:29 paul how would you write it ?
10:29 kados <p> means 'paragraph'
10:30 kados http://www.topxml.com/xhtml/xhtml_tag_label.asp
10:30 kados label is attached to a 'control'
10:30 kados so they are not appropriate for this context
10:31 paul shame on me. I thought it was valid as I wrote it. There are zillions of <p><label>...</label></p> in the templates
10:31 paul And most of them comes from me...
10:31 paul so, how would you write something like :
10:31 kados also 'legend'
10:31 paul surname:      MYSURNAME
10:32 kados that's a list
10:32 paul name:     NAME
10:32 kados so it should be:
10:32 kados <ul>
10:32 kados <li>surname : MYSURNAME</li>
10:32 kados <li>..</li>
10:32 kados </ul>
10:32 paul nope, that's not a list, as I wanted to have MYSURNAME & NAME vertically aligned.
10:32 kados that can be done with css
10:32 paul otherwise it's too hard to read.
10:33 kados can you say again what you want?
10:33 paul how ? with a <span> for "surname:" ?
10:33 paul surname:      MYSURNAME
10:33 paul name:        N AME
10:33 paul mmm.... M and N should be aligned vertically
10:34 kados ahh, ok
10:34 kados but not surname?
10:34 kados and name?
10:34 paul I just want 2 columns : one with the labels, one with the content.
10:34 paul the "label" being left or right aligned, I don't care.
10:35 kados to preserve semantics, the whole element should be a list element
10:35 kados but you could wrap the label in a <span> tag
10:35 kados which would allow a css designer to change it's position visually
10:36 paul http://o5.bureau.paulpoulain.c[…]ha/opac-search.pl
10:36 paul Any word / Title / author are column 1
10:36 paul and the forms entries are column 2
10:36 kados right
10:36 paul <label> is correct in this case, because that's a form.
10:36 kados exactly
10:36 paul I wanted the same behaviour for display
10:37 paul screens
10:37 kados paul: it's possible that <table> is also legitimate in this instance
10:38 paul I don't think so.
10:38 kados paul: though should be avoided if possible as it's hard for screen readers, etc.
10:38 paul (but I may be wrong...)
10:38 kados owen is the expert in this
10:38 kados we can ask him the best way to acomplish it
10:39 kados also there are <legend> where there is no form
10:39 kados in moremember, etc.
10:39 paul arent's they legitimate when there is a <fieldset> ?
10:39 kados legend is specifically a caption for form elements associated with a fieldset
10:39 paul maybe a fieldset is invlid when there is no form...
10:39 kados only if there is also a form :-)
10:39 paul ok, my fault too then...
10:39 kados fieldset can be used anywhere
10:40 kados though IMO it's a useless tag
10:40 kados <div> is better
10:40 kados if you need special style, <div class="specialclass">
10:40 paul if you let me know the best way to do this, I agree to parse all templates to fix the problems.
10:40 paul (/me + toins, of course ;-) )
10:40 kados I have already fixed member.tmpl
10:41 kados and half-way for moremember.tmpl
10:41 kados paul: ok, I will talk with owen and file a bug
10:41 paul anyway, I think we must do anything needed to have only 1 set of templates.
10:42 paul I won't change my mind on this.
10:42 kados ok
10:42 kados I think it's best too
10:42 paul and as you & owen proved you're from far better designer than me, i'll follow your instructions.
10:43 paul one note : today, toins will work on a new default value for MARC editor.
10:43 kados like a plugin?
10:43 paul nope.
10:44 paul every subfield of each framework will be able to have a default value
10:44 kados ahh
10:44 kados cool
10:44 paul will be very useful in many circumstances, I wanted to write this since months...
10:45 kados paul: another question
10:45 kados in members templates
10:45 kados we have three 600+ line templates for memberentry
10:45 kados wait, 4 in fact
10:45 kados 2,400 template lines just to input a member
10:46 kados IMO it's a nightmare :-)
10:46 kados very hard to maintain
10:46 kados do you have ideas how to make it less redundent?
10:46 paul I may agree, but that's SANOP stuff, it was their 1st stuff, so maybe we could clean it progressively
10:47 paul imho, the worst part is not that it is long templates (even if I agree it's a shame)
10:47 paul it's that you need 3 clics and form submit to validate a member.
10:47 kados yep
10:47 kados but if you read the template
10:47 kados it's quite repetative
10:47 paul and it's even worst if you just want to modify the password : you'll have to submit the 3 forms anyway !
10:48 kados paul: memberentry.tmpl in dev_week is 178 lines
10:48 kados paul: just to give some perspective
10:48 kados paul: I have an inclination to 'reject' the new members stuff based on this :-)
10:48 kados new memberentry is 20X longer :-)
10:49 paul I think you should list all things that could be better done, then we could set priorities...
10:49 paul Note that there are some/many new features that can explain some growth...
10:49 kados only 4 or so fields have been added
10:50 kados so we should have instead about 200 lines at most in memberentry IMO
10:50 kados especially in a 'prog' template
10:50 kados right
10:50 paul because if it works fine, it's less important that something that don't work.
10:51 paul and we know how many ressources we have at our disposal.
10:51 paul so, sometimes "better" is the biggest "good" ennemy
10:51 kados right
10:52 kados but working is relative
10:52 kados what I mean is
10:52 kados if I want to convert memberentry to YUI for instance
10:52 kados it will actually take less time if I re-write it first :-)
10:52 kados otherwise I will have to deal with 20X as much code
10:53 kados anyway
10:53 paul give the ball to SANOP who write it ;-)
10:53 kados paul: from our tests, acquisitions, authorities, and serials are broken in 3.0/head
10:54 kados paul: does san-op use these features?
10:54 paul authorities & serials yep
10:54 kados are you sure you've committed all code?
10:54 paul (although authorities only a little)
10:54 paul I ? in this matter, we should ask the question to SANOP.
10:54 paul but I think they did.
10:55 kados we want to check before we start filing bugs
10:55 paul what kind of error do you get ?
10:55 kados so there is no missunderstanding
10:55 kados it's not just 'an error' ... it's a suite of errors :-)
10:57 rch paul / hdl: have you seen this error with updating authorities... ?
10:57 rch [warn] Couldn't open gils.att [Too many open files]
10:57 rch 13:00:02-18/03 zebrasrv(71) [warn] Couldn't load attribute set gils.att [Too many open files]
10:57 rch 13:00:02-18/03 zebrasrv(71) [warn] bib1.att:5: Include of attset gils.att failed
10:58 rch 13:00:02-18/03 zebrasrv(71) [warn] Record didn't contain match fields in (bib1,Local-Number)
10:58 rch 13:00:02-18/03 zebrasrv(71) [log] error grs.xml <no file> 0
10:58 rch ?
10:58 paul the 2 first, no
10:58 paul The 3rd one, yes
10:58 paul Record didn't contain match fields in (bib1,Local-Number)
10:59 hdl Couldn't open gils.att [Too many open files] <- First time I come across this.
10:59 paul where is stored your authid in the authority ?
10:59 rch 001
10:59 paul and it's in the xml field for sure ?
10:59 paul and it's in the xml record for sure ?
11:00 hdl Is your authtypecode stored in 152$b ?
11:00 rch yes.
11:01 hdl which zebra version do you use ?
11:02 hdl rch : is there no other warnings before Couldn't open gils.att [Too many open files] ?
11:02 rch zebra 2.0.12
11:03 rch 10:00:02-19/03 zebrasrv(248) [request] Init OK - ID:81/81 Name:ZOOM-C/YAZ Version:1.116/2.1.54
11:03 btoumi hi all
11:03 rch 10:00:02-19/03 zebrasrv(248) [log] function: 1
11:03 rch 10:00:02-19/03 zebrasrv(248) [log] Waitaction: 2
11:03 rch 10:00:02-19/03 zebrasrv(248) [log] Received DB Update
11:03 rch 10:00:02-19/03 zebrasrv(248) [log] action
11:03 rch 10:00:02-19/03 zebrasrv(248) [log] specialUpdate
11:03 rch 10:00:02-19/03 zebrasrv(248) [log] database: authorities
11:03 rch 10:00:02-19/03 zebrasrv(248) [log] enabling shadow spec=/zebra/rel_3_0/authorities/shadow:1G
11:03 rch 10:00:02-19/03 zebrasrv(248) [log] cache_fname = /zebra/rel_3_0/authorities/shadow/cache
11:03 rch 10:00:02-19/03 zebrasrv(248) [log] record 0 type XML
11:03 rch 10:00:02-19/03 zebrasrv(248) [log] 314 bytes:
11:03 rch <?xml version="1.0" encoding="UTF-8"?>
11:03 rch <record xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.loc.gov/MARC21/slim http://www.l ...
11:04 rch 10:00:02-19/03 zebrasrv(248) [warn] Couldn't open gils.att [Too many open files]
11:04 rch hi btoumi
11:04 btoumi hi rch
11:05 btoumi kados u have some questions?
11:05 kados btoumi: hi :-)
11:05 kados btoumi: 1. memberentry in dev_week is 178 lines
11:05 kados btoumi: in head/3.0 it's 2,700 lines :-)
11:06 kados paul: http://www.w3schools.com/tags/tag_dl.asp
11:06 kados paul: to answer our question about semantics before
11:06 kados paul: defines a definition list
11:07 btoumi kados: about tmpl?
11:07 kados btoumi: yep
11:08 kados paul: so we can use <dl> <dt>Name:</dt> <dd>Joshua Ferraro</dd></dt> ... </dl>
11:08 btoumi kados can u tell me the file name please?
11:08 kados btoumi: memberentry*.tmpl
11:09 kados btoumi: they are 600 lines a piece
11:09 kados btoumi: in dev_week/rel_2_2 less than 200 lines and only one file
11:11 btoumi in according with paul we made some change in template
11:13 btoumi before the template was hardcoded
11:13 btoumi koados what is the problem
11:14 btoumi kados what is the problem?
11:14 kados I will explain
11:14 kados 1. there are three steps to create a borrower
11:14 kados and in the template, the content for the borrower is repeated over and over
11:15 kados if you need to represent three steps, it could be done in css with only a single form
11:15 paul (and some javascript)
11:15 kados most libraries won't want three steps anyway
11:15 kados paul: yep
11:15 kados 2. there are three separate templates with three steps each
11:16 kados sorry ... s/three separate tempaltes/four separate templates/
11:16 kados it means that if I need to update the memberentry templates I have to update 4 files
11:16 kados since all the fields are identical for all four
11:16 kados we only need one template IMO
11:17 kados ultimately, 2,700 lines of HTML are unmaintainiable
11:17 kados especially since in dev_week and rel_2_2 memberentry.tmpl is only 180 lines
11:18 btoumi_bis so what do we do ?
11:19 owen kados: you around?
11:20 kados owen: yup
11:20 kados owen: good timing, we're discussing templates
11:21 hpassini kados so what do we do ?
11:21 kados hpassini: do you mean 'how to fix'?
11:21 hpassini yes
11:22 kados ok:
11:22 kados 1. merge memberentry*.tmpl into one file
11:22 kados 2. only have one form in memberentry.tmpl for submitting
11:22 paul (I disagree with #1, I can argue why)
11:22 kados paul: ok, sure
11:22 kados paul: why? :-)
11:22 paul (finish, I will say later)
11:23 paul (I wanted to wait until you had finished to type <enter> but my keyboard is too sensitive it seems...)
11:23 kados hehe
11:24 kados #2 is critical
11:24 kados so instead of :
11:24 kados <step1>
11:24 kados <form><input></input>...</form
11:24 kados <step2>
11:24 kados <form><input>...
11:24 kados <step3>
11:25 kados <form>...
11:25 kados you can just have:
11:25 kados <form><input></input></form> :-)
11:26 hpassini I agree but actually we don't have enought time
11:26 kados hpassini: you don't have enough time to fix it you mean?
11:27 hpassini yes
11:29 owen Sorry, I'm coming in late to the conversation: Why are there three steps to add a borrower?
11:31 paul owen : yes, adding a borrower needs 3 steps :
11:31 paul - basic informations
11:31 hpassini it's to have all information without use scrollbar
11:31 paul - administrative informations
11:31 paul - library use informations
11:31 paul (roughly)
11:32 owen Have you gotten feedback from libraries that a single page entry form is not working for them?
11:47 hdl owen : ppl like to see all information (of one type) on one page to check it all at once with a glance.
11:47 kados yea, scrolling--
11:47 hdl and not moving back and up with mouse.
11:47 kados I agree
11:50 hdl But Passing a parameter to the template could be a solution rather than having it on three pages. and having <TMPL_IF Name="step1"></TMPL_IF>
11:50 hdl Or we could have divs lake tabs in MARCeditors.
11:50 paul hdl : in fact, I think we should have 1 step only from software pov, but 3 steps from user pov
11:50 kados paul++
11:51 paul thus, javascript and something like "tabs" in MARC editor
11:51 hdl pov : persistence of vision ?
11:51 paul point of view
11:52 hdl (hdl was wondering why pov-ray came into conversation :P )
11:52 kados hehe
11:54 paul owen do you have a minut to speak of html
11:54 paul ?
11:54 owen Yes
11:55 paul in my prog templates, i've added many (zillions in fact) of <p><label>...</label>...</p>
11:55 paul I thought it was valid
11:55 paul kados: just told me it isn't as <label> are for form fields only
11:55 owen Yes, that was my understanding as well
11:56 paul how to replace them in detail.pl scripts to have a 2 columns display :
11:56 owen <label> is designed for use with <input>, since you can match up the id attribute
11:56 paul the left one with the description of the value
11:56 paul and the right one with the content.
11:56 paul like
11:56 paul surname : poulain
11:56 paul name : paul
11:57 paul both p being vertically aligned
11:57 paul kados pointed : http://www.w3schools.com/tags/tag_dl.asp
11:57 paul is that what i'm looking for ?
11:58 owen If you're not willing to use a table, I think that's the best option. I would argue that a table is not inappropriate, since you're presenting information which is structured as a grid
11:58 owen I don't see any accessibility issues with a table in this case
11:59 paul kados pointed the accessibility problem.
11:59 owen A screen reader used by a blind user would be able to handle the linearity of the table just fine (as I understand it)
11:59 paul anyway, I think we should choose 1 method and stick with it everywhere

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