IRC log for #koha, 2007-10-18

All times shown according to UTC.

Time S Nick Message
12:12 kados paul++ for seeing 'insecure == 'NO''
12:13 kados paul: I think that was the cause of many problems for me :-)
12:13 kados and I couldn't tell because in sysprefs it displays as 'no' :-)
12:13 kados (even though it's 'true' for perl) :-)
12:18 paul wow, chris is now a "famous kiwi"...
12:18 paul
12:24 kados w00t
12:24 kados congrats chris!
12:25 kados and Koha too, for being a Finalist
12:25 kados :)
12:29 paul kados : there are 12 blo/cri bugs, at least 5 of them waiting for an answer from you.
12:29 paul could you pls take care of them?
12:37 kados paul: yes, it's on my list for today
12:38 paul kados: just to let you know what it means to have a company in France. I just recieved a paper from my gov taxes service. There is a list of "dates to respect" (to declare or pay taxes). The list contains... 29 dates !!! we are not concerned by all of them, but the 1st game is to find which ones are for us (as it's not specified).
12:39 hdl Congrats to all
12:46 hdl kados I sent a first commit for ISBN-13 management. But it was buggy.
12:46 hdl I just sent a new patch which works now
14:14 paul kados : about amazon images. Do you know that the terms of use changed recently and are now limited to business partners ?
14:14 paul So it seems that this feature can't be used by our libraries.
14:15 paul It has been reported to me by jerome pouchol, so feel free to comment/complete this information.
14:17 atz that is bad news!
14:17 liz_ aw bummer
14:17 paul (I would be pleased if someone checked & confirm this information)
14:17 atz do you have a link or a communication from them about it?
14:18 paul nope, just some mails in french from SAN-OP and another guy that uses amazon images in it's opac (internally developped)
14:18 atz we'll have to check that out... I don't imagine it is very easy to get ahold of anybody there via phone  :[
14:20 liz_ well maybe if they hear from enough libraries they would consider a "library partner" designation
14:20 paul liz_ ++
14:21 atz if you are going to pay for content though, you might choose other sources
14:21 paul (even if I feel it will be hard, because amazon business don't grow through libraries. someones may even think it is lowered by libraries...)
14:21 atz like syndetics
14:21 twiggle i'm not sure that's true really (libraries lowering amazon business) especially if the pic links back to their interface
14:22 twiggle i've been known to be a "try before you buy" library user
14:22 paul twiggle: i never said it's true. I just said some ppl may think it's true ;-)
14:22 twiggle :)
14:22 twiggle i kno, i'm just being contrary ;)
14:23 paul (and if the guy who think it's true is the amazon president ...)
14:23 twiggle (yea, that's a problem)
14:23 twiggle ^.^
17:28 kados hdl: you around?
17:29 hdl passing by.
17:29 hdl yes
17:29 kados hdl: I'd like to ask you about your ISBN 13 note ... I don't understand it
17:29 hdl I sent 3 patches.
17:29 hdl The only one that will work is the third.
17:29 kados ahh, ok
17:30 kados gmcharlt did some work on ISBN 13 normalization I think
17:30 kados gmcharlt: you around?
17:30 gmcharlt yep
17:30 kados[…]351bc26e49bff7a9c
17:30 kados hdl: is that related to your patches?
17:30 hdl It is only a rewrite for DisplayISBN
17:30 kados (I don't think I've seen them yet, chris gets up in a few hours)
17:31 gmcharlt my patch probably isn't related to ISBN13 specifically, but should be compatibile with them
17:31 kados gmcharlt: your intention with that patch was to normalize the isbn query ... does it assume a specific format that isbns are being saved in?
17:31 kados and in index vs sql vs marc ?
17:32 gmcharlt it makes no particular assumptions -- it just strips out parenthetical stuff in the 020$a that catalogers add
17:32 gmcharlt the parentheses currently cause a crash in the search code during hte dedupe check
17:32 hdl kados : my patch display ISBN13 with correct dashes. It is all it does.
17:33 gmcharlt but also, the parenthetical information is mostly not needed for that sort of check
17:33 kados so i think we need to be explicit about how we're storing ISBNs internally
17:33 hdl (assuming they are not STORED in database with dashes.
17:33 hdl )
17:33 kados because i think hdl is assuming they don't have dashes in the db
17:34 kados gmcharlt: what's your experience with how that's generally handled?
17:34 kados gmcharlt: is it good practice to normalize the ISBN down to a digit with no dashes for internal representation?
17:34 kados gmcharlt: and then perhaps re-construct the dashes for display?
17:35 gmcharlt that sounds OK (keeping in mind that the last character of an ISBN can be an 'X')
17:35 kados will catalogers be upset if we remove their parentheticals?
17:35 kados :-)
17:35 gmcharlt for search purposes, they won't care
17:36 gmcharlt but the 020$a should be left alone, and in the full bib display, taken from the 020$a directly if possible
17:36 kados that presents a challenge
17:36 kados since we're indexing 020$a directly in zebra
17:37 gmcharlt are there hooks to do some normalization of the bib data before it is presented to zebra?
17:38 gmcharlt also one thing regarding hyphens -- per MARC 21, the hyphens are not supposed to be stored in the 020$a
17:38 gmcharlt so if the data is entered (or made to be entered that way) the ISBN will index as a single token
17:38 gmcharlt the problem is that some "inexperienced" catalogers may enter the 020$a with the hyphens
17:38 kados ok, good to know
17:38 kados right
17:39 kados ok, so we always remove hyphens
17:39 kados rule #1 :-)
17:39 kados we could add a hook to normalize bib data
17:39 kados before indexing it in zebra
17:39 kados that's actually a great idea
17:39 kados there will be a bit of a performance hit
17:39 gmcharlt and even if we don't need it for ISBNs, we'll undoubted need it for other things
17:40 kados but worth the normalization it'll give us
17:40 kados gmcharlt: definively
17:40 kados hehe
17:40 kados definitely
17:40 gmcharlt and we keep indexing as a background proces
18:12 thd kados: hyphens in ISBNs usually break queries when reused with hyphens on other systems.  Hyphens are nothing more than an aid to human readability and help in retyping but are very problematic when used in copy and paste queries.
18:13 kados thd: I'm here
18:13 kados thd: yes, I understand that now
18:13 kados thd: our default is to remove all hyphens from the data before the data is saved
18:14 thd kados: ther should perhaps be a user option to remove ISBN/ISSN/ISMN etc. hyphens even from display
18:16 thd ... merely because removing hyphens from copy and paste queries for other systems which cannot cope is a nuisance.
18:16 thd s/ther/there/
18:16 kados thd: good point
18:17 gmcharlt hdl: you around?
20:16 owen Yeah, I noticed some bug reports coming from Galen Charlton (gmcharlt, I assume)
20:17 owen Ooops, wrong tab.
20:17 gmcharlt gmcharlt = Galen Charlton, yes
20:17 kados gmcharlt: I was just telling Owen how impressed I was :-)
20:17 kados in a pmsg
20:18 kados and it seems to have spilled over into the main channel :-)
20:18 owen I was trying to remember your unpronounceable nick, gmcharlt  ;)
20:18 gmcharlt you can blame the CS department at my university for that :)
20:19 kados gmcharlt: meet Owen, he's responsible for the new koha 3.0 templates, and the beautiful
20:19 kados and not to mention the official koha template interface designer :-)
20:19 gmcharlt Nice to meet you-, owen
20:20 owen Likewise. Where are you located, gmcharlt ?
20:20 gmcharlt Chicago
20:21 kados owen: galen will be in athens week of nov 5th, you can meet him in person
20:21 owen I went to school in Chicago, '89-'92
20:22 kados hehe
20:22 gmcharlt Been living here for a while, ever since doing the Chicago -> Anchorage -> Tallahassee -> Chicago round trip
20:25 owen See y'all in 16 hours :)
20:49 hdl gmcharlt: are you there ?
20:49 gmcharlt yep
20:49 gmcharlt had a question -- hold on one moment
20:49 hdl you wanted to ask me a question ?
20:50 gmcharlt yes -- first, let me introduce myself : my name is Galen Charlton, and I was recent hired by LibLime to work on Koha
20:50 gmcharlt I was thinking of writing a tool to compare a current Koha DB against kohastructure.sql or another DB
20:50 gmcharlt among other things, to verify that any given upgradedatabase is working
20:50 gmcharlt kados said that if anything like this existed already, you might have done it
00:44 thd gmcharlt: I have some code for part of your quest
00:44 gmcharlt great
00:45 gmcharlt if email is convenient, mine is
00:46 thd gmcharlt: I also have some anecdotal evidence from my own experience that may not have always worked.
00:47 gmcharlt thd: by not including all of the schema changes, I assume?
00:48 thd gmcharlt: I discovered about three weeks ago that missed some columns for my installation which had been added between 2.2.5 and 2.2.6
00:49 thd gmcharlt: the code itself still does not use those columns so it was easy to miss them
00:50 thd ... without noticing any consequence
00:50 gmcharlt thd: right
00:50 thd I can give you a URL to some code
00:55 thd I thought that I could give you a link but kados has not committed my latest patch
00:58 gmcharlt thd: I can wait, or can you e-mail me the patch?
00:59 thd gmcharlt: only a small bit of that patch is relevant so I will extract it and email it to you.
00:59 gmcharlt thd: thanks
00:59 ryan thd: updatedatabase for 3.0 starts from scratch
01:00 thd ryan: I presume that it does but that does not mean that 3.0.2 will not miss something
01:01 gmcharlt thd: exactly why I want to write something to validate a DB against kohastructure.sql
01:01 ryan yes, true... in fact , it does :)
01:02 ryan there's some fk constraint that has to be manually removed.
01:03 gmcharlt where does the 3.0 updatedatabase do that check?
01:33 thd gmcharlt: I just sent the email
01:33 thd gmcharlt: are you Foxnorth?
01:34 gmcharlt no, I'm not
01:34 gmcharlt got the email, thanks
03:06 atz anybody hear anything stateside about the Amazon images ?
03:06 atz Frenchies say Amazon is restricting the image service to "partners".
03:38 kados I haven't heard anything
03:38 kados but I'll check right now
03:41 kados[…]me=A36L942TSJ2AJA
03:42 kados doesn't seem to have changed at all
03:42 kados since the last time I checked a few months ago
06:30 atz that's good.  maybe we're just hearing FUD.
06:40 atz I can't tell whether it was ever compliant with that page though.  
06:40 atz I'm sure we do more than 1 hit per second.
06:41 chris thats per ip isnt it?
06:42 atz I'm not sure how the API works... but one hits page is going to have 10 images on it
06:43 chris true
06:43 chris or even 20, depending on how many results per page its returning
06:43 chris joshua has corresponded with amazon quite a bit
06:43 chris theres even a blog post on it
06:44 atz and if it actually checks first w/ the service (it might, but I'm guessing it doesn't), then the servers would be connecting more often than 1/s under load
06:45 atz especially, say, in a hosted environment where many servers live on one box
06:46 chris its not the servers that do it
06:46 chris its the clients
06:46 chris ie the content is fetched from amazon to the browser
06:46 atz i realize the images are just statically mapped
06:46 atz right
06:46 chris if you are doing it realtime
06:47 chris if you decide to cache them locally, you can do that too
06:47 chris but you have to update them every 24 hours
06:47 atz other OPACs I've seen will query out first to verify added content exists
06:48 atz to know whether to include a "review" or "sample" section
06:48 chris right the amazon module as it stands just fetches
06:48 chris so you get empty review sections
06:48 atz right... that looks lame
06:49 chris yep
06:49 chris but it gives you more chance of staying with the t&c
06:49 atz but I suppose you can try some js to hide it after the fact
06:49 chris yeah hidden divs
06:50 atz so how do we know how to reference the correct image?  formulaic mapping based on ISBN?
06:51 chris not sure kados wrote the module, but i think so
06:53 atz actually, the C4::Amazon module does HTTP::Request or LWP connections to Amazon
06:53 chris oh really? i didnt think it did
06:53 chris i stand corrected
06:54 atz yeah, this doesn't seem to be related to opac code, necessarily
06:58 atz lol, he has LWP::UserAgent identify itself as "Mozilla/4.76 [en] (Win98; U)"
06:58 chris heh
07:00 atz well, sleeptime for me
07:00 atz ttyl
07:00 chris sleep well
07:04 masonj cya joe

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