IRC log for #koha, 2006-07-24

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

All times shown according to UTC.

Time Nick Message
12:04 thd kados: I still think that the top level node should be the full form with all the subdivisions on a line
12:04 kados: $a should start underneath the full form when expanded
12:05 kados I could do that quite easily
12:05 thd kados: then you would technically have a four level hierarchy
12:05 kados right
12:05 I'm not 100% convinced though
12:05 I think the three-level is a bit more what users expect
12:06 at least it's what I'd expect
12:07 thd kados: you would not want to waste people's time expanding nodes which only have $a so those should be automatically expanded or not show a sublevel
12:07 kados they are/do
12:07 only nodes that have more than $a have a +
12:08 and I can auto-expand them if that's the desired behavior
12:08 thd kados: yes but they might with my suggestion above where $a is a sublevel of the full form
12:08 kados ahh
12:09 thd kados: default expansion would be good if the result set was small
12:10 kados: the biggest problem is that you need to locate this where there is more screen space for an interesting display
12:30 kados thd: ok, I think the prototype is finished
12:30 thd: it only uses $a $x $v currently
12:30 thd: but it will create trees of $a->$x, $a->$x->$v and $a->$v
12:30 thd $z is very important
12:30 kados yes, I can add that once we confirm my logic is correct
12:31 (very simple copy/paste to add it)
12:31 thd not to forget $y but that is a little less common
12:33 kados: maybe everything should not be expanded initially but I think $x should always be expanded unless the user has closed it
12:35 kados: checkboxes for selecting $a Kings and rulers as well as $v Fiction while ignoring the $x subject subdivision
12:36 are missing
12:37 kados hehe
12:38 that part can't be done in this design
12:38 thd kados: what is the limitation?
12:38 kados my skill :-)
12:38 thd kados: I did it
12:39 kados well, I'm quite happy with this one at the moment
12:39 thd kados: that is a fundamental part of one of my secret experiments from two years ago
12:40 kados yea, but I think too many options confuses the user
12:40 thd kados: of course it is good
12:40 kados as it is, knowing how to expand the list is pretty hard for most users
12:40 thd kados: you have to give the user a chance to be confused first :)
12:40 kados hehe
12:40 well, I'll give it some thought
12:41 currently, everything is implemented as a link
12:41 to do it your way, I'd have to change that to use input tags
12:41 which would be a major overhaul of the javascript stuff
12:41 thd kados: yes so it would be a form
12:41 kados yep
12:41 this is good enough for my initial purposes
12:42 I think we should give it a nifty name
12:42 KohaBrowser or some such
12:42 (course, it's not as fancy as the AcquaBrowser )
12:43 thd kados: there is a usability problem for not being able to easily and obviously select the fully expanded form
12:43 kados: yes they built that themselves
12:44 kados: another non-intuitive interface
12:45 kados: did you notice that dancing term relation eye candy
12:47 kados: their progressive subdivision link search is what you have implemented but it belies the actual structure of the subdivisions
12:49 kados: as $v is not necessarily exclusively related to $x.
12:50 kados_: or better $v is not a subdivision of $z as implied by the Queens Library subject linking interface
12:52 kados_: I used the implied continuous progressive hierarchy at Queens Library as an example of a mistaken understanding of subdivision hierarchies.
12:54 kados_: more importantly it is not obvious to the user that searching for the fully subdivided subject requires clicking on the last term
12:55 kados: that is why I was recommending a separate link with the full subject as the top node
12:56 kados_: are you there?
12:59 kados_ thd: I'm here
13:00 kados server musta timed out
13:01 search on neal stephenson
13:01 thd kados_: did you read what I posted about the mistaken progression in the Queens Library interface?
13:01 kados they don't have what we have
13:01 you can't expand "Kings and rulers"
13:02 thd: please explain a bit more
13:02 thd kados: they have the ability to expand or restrict the search progressively as if it is one continuous hierarchy which is what you have
13:02 kados you mean the aquabrowser?
13:02 thd kados: yes
13:03 kados i see that as something completely different
13:03 I don't get 'Kings and rulers' in the acqua browser
13:04 the aquabroser looks to be a bit more intelligent
13:04 though perhaps not quite as useful :-)
13:04 and I have a design for it as well using an open source toolkit
13:05 the thing we just designed is an improved version of what they have on the right-hand side
13:05 the 'Refine your search' column
13:06 thd: http://aqua.queenslibrary.org
13:06 though I won't get any arguments that their's looks way better :-)
13:07 that came out wring
13:07 wrong even
13:08 s/a/as/
13:09 thd kados: I cannot see the eye candy because it must require Flash 8 which is not available for GNU/Linux.
13:09 kados ahh
13:09 well it looks very pretty in Firefox on OSX with Flash
13:10 thd kados: you can make it pretty and the eye candy is merely decorative rather than useful
13:10 kados well, it's useful for sales :-)
13:10 thd kados: so you see the animation in OSX?
13:11 kados yes
13:11 it's pretty cool
13:11 I can reproduce it though :-)
13:11 thd kados: and yes the eye candy would create sales but I do not know of any free or open source version of the dancing term nodes
13:12 kados using http://www.touchgraph.com/
13:12 that's open source and it works as well
13:12 uses java
13:12 (do you have java installed?)
13:12 thd kados: yes I do
13:13 kados: although Java is not free, well neither is Flash
13:13 kados true enough
13:15 thd kados: Java is also much slower than flash
13:16 kados hehe
13:16 thd: you need to trade work for a new computer that works well :-)
13:18 thd kados: ok, what can I do, I have a new computer that just needs an expensive hard drive
13:18 kados you do?
13:18 thd kados: It may also need an expensive DC power supply
13:19 kados: I have a couple of Intel rack servers from two years ago
13:19 kados ahh, cool
13:54 thd kados: I know that Flash would have been much faster than this.
13:54 kados: this eye candy stuff is good for selling systems but it does not scale into a useful tool.
13:55 kados yep, in general I agree
13:57 thd kados: the general problem is that given a large enough term set no one has enough RAM to store all the nodes and the tags start cluttering up the screen and blocking the view of each other fairly quickly long before you have RAM problems.
13:57 kados yep
13:57 thd kados: there is also not enough network bandwidth for a large term set.
13:58 kados: however, I show people good stuff and then I show them this and they think that this is the real thing.
13:59 kados this? meaning nearly what we just did this morning?
13:59 thd kados: the eye candy is probably a requirement for selling the systems so that the ignorant can feel comfortable with having taken a good choice for the wrong reasons
14:00 kados yep :-)
14:01 thd kados: 'this' meaning dancing visualisation tools not what we were working on this morning
14:01 kados: there is a problem with what we were working on this morning
14:02 kados ahh
14:02 thd kados: now that my system has recovered from loading the eye candy I can check the comparison with the aqua interface at Queens Library
14:02 kados well, it doesn't have the checkboxes you want :-)
14:03 thd kados: that has an important consequence
14:04 kados but my problem with that is I don't think it can be done in a way that users can understand
14:04 thd kados: you have a tree diagram that shows the relationships correctly but you cannot do anything with the nodes individually
14:04 kados true ...
14:04 but we'd have to come up with a new interface to handle that interaction
14:05 the current design can't support it in a user-friendly way IMO
14:06 thd kados: e-tailers with some sense are paying $100,000 plus to add this sort of thing for their customers
14:06 kados: library users are more clever than the average person trying to by a shirt online.
14:10 kados: I know what it looks like because I did it.
14:11 kados: the web is full of interfaces which show your search terms with a checkbox next to each term at the top of the result set.
14:11 kados yes that's true
14:11 I agree it's trivial to do
14:12 but maybe not in a single day :-)
14:12 thd kados: it is far from the dominant usability enhancement but is more likely to be ignored than to create confusion.
14:54 kados: your five minutes are up and I have a link for you with an example
14:54 kados: are you back yet?
15:31 kados thd: welcome back
15:31 thd kados: http://www.ovid.com/site/help/[…]_guide.html#Tools
15:34 kados: also in the same page http://www.ovid.com/site/help/[…]uide.html#Mapping
15:46 kados: I have a better link from McGill which is now broken
15:47 kados: It has much better screen shots and is partly working in the internet archive
15:48 kados: http://web.archive.org/web/200[…]utorial/broad.htm
15:49 kados: look at the table of contents
15:50 kados: the images are broken for subject searching
15:50 at the top level
15:51 kados: however the images are there for subparts of subject searching after mapping
15:52 kados: requires clicking on each separately
15:53 kados: the images are good and explain well but the more recent versions of the same manual on Ovid's own web site are not as clear or at least lack enough good screen shots
15:54 kados: the Ovid link which I supplied was for a more general manual
15:55 kados interesting
15:55 it's still hard to picture it without actually using it
15:56 (at least it is for me :-))
15:58 thd kados: the example in the internet archive at least gives you a step by step search example
16:04 kados: do you have a link for tree expansion using XML without page reloading?
16:05 kados no I don't ...
16:06 thd kados: you had described something to me but did that reload the page?
16:07 kados yes ... I've done it before but the code is somewhere on my terabyte hard drive
16:07 I really need to spend about a week organizing things
16:07 thd kados: you described something which did not use JavaScript and did not merely hide the content but fetched it when the node was expanded.
16:07 kados no, that usese javascript
16:08 using XMLHttpRequest()
16:08 a javascript method
16:08 you could do something similar with cleverly placed invisible frames
16:08 thd kados: is there no Perlish way to do that?
16:08 kados but i fear it would cause problems with browser versions
16:08 not without a page reload
16:09 thd kados: frames are almost always evil
16:09 kados yep
16:09 thd kados: I suspect even invisible frames are evil
16:09 kados :-)
16:10 thd kados: page reloads are not really a problem when the content is not large
16:11 kados: my secret test interface allows the user to have every node of an entire thesaurus open if the server imposes no limit so reloading could be expensive.
16:12 kados yea
16:12 evergreen has some nice interfaces in javascript
16:12 as does amazon.com
16:12 and google for that matter
16:12 thd kados: by every node I do not mean every node in one branch but multiple branches simultaneously
16:13 kados: some of the Google AJAX interfaces bring my system to a crawl
16:15 kados: is Evergreen any easier to install now?
16:29 kados no :-)
16:29 it's harder in fact :-)
16:57 thd kados: more features mean more difficult to install?
16:57 kados for?
16:58 evergreen?
16:58 thd evergreen
16:58 kados yes, in general ...
16:58 mainly it's the underlying framework it's built on that makes it so complex
16:58 it's even more complex than zope
16:58 thd :)
16:58 kados they basically invented their own Server Request Framework (transport layer for communicating between servers)
17:16 there is data loss if there are parallel elements within a given subject
17:16 (meaning it discards all but the last subdivision)
17:16 gotta fix that...
17:21 thd: you still there?
17:21 thd: do you have any example records that have repeated $x, $y $z (any or all)?
17:22 thd: I can import them into this dataset for testing purposes
17:23 thd kados: some of your other clients do have them
17:23 kados: and  many of them while NPL has few
17:24 kados: I basically use only one record and test over and over again with that record because it exhibits so many difficulties
17:24 kados if there is a repeated $x and repeated $v/y/z, how should they properly nest?
17:24 which record?
17:24 thd kados: I searched for a good example to test for problems
17:24 kados: I did not find it accidentally
17:25 kados should I pay you for it? :-)
17:25 thd kados: search for Forgotten Chicago at LC
17:25 kados: It is actually missing some problems
17:26 kados: so there is another record that I have used but not for what you have been asking
17:28 kados: Oh wait, a repeated $x well that is rare and always a cataloguing mistake when I have seen it
17:29 kados: Forgotten Chicago has a repeated $z
17:29 kados: and some significant multiple divisions
17:31 kados: every time I have seen a repeated $x except maybe once it should have been a $z instead
17:33 kados: the only other time which might have been legitimate there was no subdivision for the repeated $x
17:35 kados: there is a possible good repeated $x in Afognak if you wanted to search there.  I fixed the bad ones which should have been $z and came from obscure libraries in Canada.
17:35 kados thd: http://zoomopac.liblime.com/cg[…]Forgotten+Chicago
17:35 thd: it's imported
17:36 thd kados: it does not have a repeated $x if that is what you wanted
17:36 kados thd: tell me if the subject tree is acting like you expect with that record
17:36 thd: ahh ... :(
17:36 it does have repeated $z
17:36 thd kados: read above to see that is very rare
17:37 kados I don't have $z displaying currently
17:37 thd kados but yes it has repeated $z which is common
17:37 kados now in one of the subjects, the order is:
17:37 $a $z $z $x $v
17:37 does that mean that the $zs are branching from the $a?
17:38 or are they parallel to the $a?
17:40 thd kados: they are subsidiary to the $a necessarily but this would suggest that everything is equally subsidiary to the $a
17:40 kados man ... that's a paiin
17:40 thd kados: it may be a mistake
17:40 kados I'm stumped over how to represent all the possibilities
17:40 thd kados: or an outdated form
17:42 kados: I have been reading about the history of LC subject heading usage and almost no one had been keeping the authorised forms up to date.
17:47 kados hey mason
17:51 hi johnb
17:51 thd kados: work to correct that problem is only a few years old after research done through OCLC.
17:52 kados: the 1980 LC subject cataloguing manual is no help
17:58 johnb Hi Kados
17:59 thd kados: you need a subscription to Cataloguer's Desktop.  They have all of the manuals and they are up to date.
17:59 mason hi
18:00 its sunny and freezing here in wgtn
18:00 im thinbking about having a morning shower
18:00 johnb We have had 2.5 inches of rain today.  Temp is in the 70's
18:03 thd kados: I just checked this excellent book and the answer seemed to be4 unknown
18:04 kados: http://www.amazon.com/gp/produ[…]v=glance&n=283155
18:04 kados bummer
18:04 johnb: have you guys managed to get zoomkoha going yet?
18:08 johnb Kados:  Cindy has been on vacation this week.  She will work on it next week.
18:11 thd kados: I may have found the problem
18:12 kados thd: yea? do tell
18:12 johnb: sounds good ... she deserves it :-)
18:12 thd kados: well I was mistaken
18:14 kados: Chan, in her revision of imroth's Guide to LC Subject Headings gives an example
18:15 oops Imroth may have had nothing to with the subject headings guide
18:16 kados johnb: mason you guys seen the new browse feature in the ZOOMOpac?
18:16 thd and I whipped it up this morning and have been spending the afternoon refining it
18:16 thd kados: I had imagined that the seeming anomaly fit the pattern but it does not
18:16 kados http://zoomopac.liblime.com
18:17 thd: (BTW: got $v and $z to play nicely)
18:17 thd: http://zoomopac.liblime.com/cg[…]Forgotten+Chicago
18:17 thd kados: Chan gives the following patterns
18:18 actually it does fit
18:19 so in the example for Place-Topic-Time-Form we have an example that starts with a topic before place
18:20 Art Criticism--France-Paris--History--18th century--Bibliography
18:22 kados ok
18:22 thd kados: the other common forms just omit some parts
18:22 kados thd: should we split up subjects into geographic, topic, etc.?
18:23 thd: also, 'Fiction' seems to be the most common subdivision
18:23 thd kados: so the only way to make sense of this in a rational scheme fitting with the theory for subject assignment is with a two level hierarchy
18:23 kados thd: should it really be nested or should it be it's own root?
18:23 thd kados: NPL has nothing but fiction :)
18:23 kados hehe
18:24 thd kados: fiction is a form subdivision
18:24 kados thd: http://zoomopac.liblime.com/cg[…]y=neal+stephenson
18:24 thd: look under 'scientists'
18:26 johnb Kados: Just looked over the zoomopac.  Looks like you got everything in there, but the kitchen sink!
18:26 thd kados: I think we need a two level hierarchy and I was mistaken previously
18:26 kados johnb: :-)
18:27 johnb: still don't have a aquabrowser equivilent
18:27 johnb: but I've got ideas for that as well :-)
18:27 thd: do tell
18:27 thd kados: Place-Topic-Time-Form should be all parallel under $a
18:28 kados thd: those being which subfields?
18:29 johnb: the nice thing is you can enable/disable things you don't want
18:29 thd kados: $z Place or Geographic -- $x Topic -- $y Time or Period -- $v Form
18:30 kados thd: k
18:30 johnb Kados:  We like options
18:30 kados thd: that simplifies the hierarchy
18:31 johnb: any comments on the new nested faceted results browse thingy?
18:31 Burgundavia: you too ... http://zoomopac.liblime.com ... do a search and watch the left-hand side
18:32 thd kados: there would be other elements for personal and corporate names where you have works etc. as subdivisions in addition to the less common subdivisions for non-name-like subjects
18:32 kados thd: currently I have only put in 100s and 700s ... but I could easily add 110, 710, etc.
18:33 thd: but I assume those should have their own category, right?
18:33 Burgundavia kados: interesting, but that section needs a title
18:33 kados Burgundavia: any suggestions?
18:33 Burgundavia such as "related subjects, titles and authors"
18:33 thd kados: I was also referring to 600 and 610
18:33 kados thd: ahh
18:33 Burgundavia kados: you also lose the navigation bar, which is kind of annoying
18:34 johnb Kados:  I like the nesty thigamajig (love all of this high tech jargon)
18:34 kados hehe
18:36 thd: should 600 and 610 be listed under subjects?
18:37 thd: or should they have their own category?
18:37 thd kados; yes definitely
18:37 kados thd: maybe 'People' for personal names?
18:37 thd: I think that's how queens does it
18:37 thd kados: definitely subjects but you should identify them
18:38 kados: every different type of subject should be represented separately with the option for the user to much them together
18:38 johnb Kados:  I have just looked up  some of our LC subject heading stuff
18:39 kados johnb: yea?
18:39 johnb Used a book and everything
18:39 kados hehe
18:39 thd kados: if you grouped personal and corporate names most users would not know the difference
18:39 johnb I can maybe answer some general concepts about lc subject headings
18:40 kados k ... but my questions may or may not be general in nature :-)
18:40 johnb Kados:  Most people would not know the difference between copr. and personal, librarians would though
18:41 thd johnb: do you have an authoritative source for the preferred sequence of every subject subdivision?
18:41 kados The overarching question is how to properly represent subjects from a MARC record in a hierarchy
18:41 johnb Kados:  Yes
18:41 Kados:  But you won't like it
18:41 kados hehe
18:41 bring it on :-)
18:42 thd johnb: what is not to like?
18:42 johnb Kados:  There is a book called LC subject headings based on tapes from LC  in verified status
18:43 Kados:  you can  either read the books (4 volumes) or but the tapes
18:43 kados hehe
18:43 johnb Kados:  I mean buy the tapes
18:43 thd kados: does the introduction give a preferred order for the subdivisions
18:44 johnb kados:  Yes
18:44 kados but order is only one aspect
18:44 johnb Kados:  you prob. wan the order, right?
18:44 thd johnb: the records are sent via FTP now
18:45 kados unless the tree is not more than one level deep
18:45 johnb <thd>  Yeah you can get them ftp but they still cost $$$$$
18:45 thd johnb: do not get them direct from LC
18:45 johnb: I assume that you do not have a subscription
18:46 johnb: if you know someone who wants to share they are legally in the public domain but a records supplier may have supplied a contractual restriction with the records
18:47 johnb <thd> no, we are a poor rural library
18:48 Kados:  I don't know anyone off the top of my head at the moment, but let me make a few phone calls on Monday
18:48 thd johnb: you paid for the creation of the records when you paid your taxes
18:48 kados thd: actually, that's not entirely true
18:48 thd: LC doesn't use tax money to create subject authorities
18:48 thd: they use money they receive in subscriptions
18:48 thd: or so I'm told
18:49 thd kados: really, but that would be contrary to law I think
18:49 kados dunno ... not an expert in the laws governing LC
18:50 thd kados: the subscriptions are only supposed to pay for the cost of maintaining the dissemination service.
18:51 kados hehe
18:51 johnb Kados here is an interesting quote "In 1974 the principle of free-floating subdivisions controlled by patern headings was officially incorporated"
18:51 kados hmmm ...
18:51 what are pattern headings?
18:51 thd kados: I have no good conception of how much money they receive from subscriptions but my suspicion is that it is not nearly enough for the cost of creation
18:52 kados: a pattern is one like what I cited from Chan
18:52 kados thd: $z Place or Geographic -- $x Topic -- $y Time or Period -- $v Form?
18:53 thd kados: you are actually supposed to be able to make an English sentence out of these now
18:53 yes
18:53 kados so that implies that there are a limited set of patterns?
18:53 johnb Kados:  We need to get our hands on "Subject Cataloging Manual: Subject headings"
18:54 thd kados: well my citation is only for the most common pattens
18:54 johnb Kados:  general rule is that you start with a subject heading and then use the subdivisions
18:55 thd johnb: is there a preferred order given for the subdivisions
18:55 johnb kados:  four types of subs:Topical, Chronological, Geographic, Form
18:57 thd john what is the preference shown in some reference for the order of those
18:57 johnb Kados:  I am reading give me a sec
18:57 kados k
19:00 johnb Kados:  You'll love this answer
19:00 Kados:  It depends on what is allowed in that particular subject heading
19:01 kados hehe
19:01 what the heck does that even mean? :-)
19:02 johnb Kados: Well if you look up the subject heading "architecture, gothic" in italics it says that (May Subd Geog)
19:03 thd kados: that was the non-answer given in Using Subject Headings for Online Retrieval
19:04 johnb Kados:  I am going to go down stairs and see if I can find the subject manual
19:04 kados johnb: May Subd Geog means something?
19:04 thd johnb: do you hide the manual? :0
19:05 kados but so far I have'nt seen evidence of that :-)
19:06 johnb Kados:  It is an instruction it means in english "may subdivie by geography
19:07 kados johnb: you've got to be kidding me
19:07 johnb Kados:  subject headings our constructed for each book
19:07 kados johnb: that's not machine readable!
19:08 thd kados: there is a chapter in Using Subject Headings for Online Retrieval about the discrepancies between the machine readable authorities and the bibliographic file contents
19:08 johnb Kados: oh, you want digital, librarians and or catalogers don't think like that
19:08 thd kados: that was obvious was it not
19:09 kados: the example johnb gave was from the printed volumes and was not meant to be understandable by machines
19:11 johnb Kados:  Not every word is a subject heading the manual will list a noun and then give the approrriate subject heading.
19:13 thd kados: machine readable is here: http://authorities.loc.gov/cgi[…]2201154&PID=15442
19:13 johnb Kados:  then once you have selected the main heading you could select a sub heading.
19:13 kados: the subheading are specific to each subject heading
19:14 Kados: then you can use the subdivisions I have just mentioned
19:15 kados: on the example there are no sub headings, look up corn
19:16 Kados: under corn there are sub headings and su sub headings in a few cases
19:16 thd kados: 008/06 is i - Subdivided geographically--indirect
19:17 kados johnb: corn where?
19:18 johnb Kados:  the subject heading "corn"
19:18 kados johnb: at LOC?
19:18 johnb Kado: in the printed manual there are three pages of sub headings for corn.
19:19 kados johnb: the data I am working with at the moment doesn't have more than $a and $x in the 650 entries for corn
19:19 ahhhh!
19:19 but ...
19:19 so ...
19:19 johnb Kados: some of those sub headings allows a further sub heading
19:19 kados are those authorities for subject headings?
19:20 johnb: what I have in my data is:
19:20 650
19:20 $a Corn
19:20 $x Fiction
19:20 thd kados: actually repeated subdivisions are always subsidiary
19:20 johnb Kados:  Corn is the subject heading
19:20 kados thd: subsidiary to which element?
19:20 thd kados: we have at least three levels again
19:21 kados is this FIFO?
19:21 johnb Fiction is the subdivision
19:21 thd kados: to the previous subdivision which has been repeated
19:22 kados thd: so it's like:
19:22 johnb Kados you can have multiple subdivisions, that is allowed
19:22 kados $a
19:22  $x
19:22    $x
19:22 ?
19:22 thd kados: yes
19:22 kados thd: that could nest indefinitely
19:23 thd kados: not in the real world but theoretically
19:23 kados thd: what made you come to that conclusion?
19:23 johnb Kados:  what stuff developed in the 19th century is outmoded?
19:23 thd kados: I always knew that but had forgotten in the past few minutes
19:23 kados: it is certainly true for every example that I know
19:24 johnb Kados:  Have we answered your question?
19:24 thd kados: consider $z Illinois $z Chicago
19:24 kados thd: so ... $a Architecture -- $z Illinois -- $z Chicago $x History $v Pictorial works.  
19:24 johnb kados: that is correct the general rule is you go from general to specific
19:25 kados so for the above subject
19:25 thd johnb: there is nothing specific about putting $x before or after $z
19:25 kados is this the proper nesting?:
19:25 $a Architecture
19:25  $z Illinois
19:25    $z Chicago
19:25  $x History
19:25  $v Pictorial works.
19:26 johnb Kados: I was talking about the subdivision words not the marc stuff
19:26 kados or this:
19:26 $a Architecture
19:26  $z Illinois
19:26    $z Chicago
19:26  $x History
19:26    $v Pictorial works.
19:26 and if $v came after $z would it be nested as well?
19:26 johnb Kados: You can do it in different ways depending on what the cataloger drank for lunch
19:26 kados hehe
19:27 I don't get how I'm supposed to know which elements 'belong' to other elements and which are in parallel
19:27 thd johnb: the sequence seems almost wholly arbitrary apart form the repeated subdivisions mostly and certainly not a principled English sentence.
19:28 johnb Kados: from what I can see your first solution is the way I would do it
19:29 thd kados: I think that cataloguing principles are supposed to treat Place--Topic--Time--Form as parallels
19:29 johnb Kados:  If you ask 10 catalogers to create subject headings for the same book you would get 10 different results
19:30 thd johnb: what you state is true of actual practise.  Yet, best practise is another question.
19:31 kados johnb: what I'm mainly interested in here is how the designers of the standards wanted it to be done
19:31 johnb: I realize in practice almost noone follows their design
19:31 thd johnb: I understand that best practise is not one choice only which is why machine readability is such a problem
19:31 kados johnb: but it's at least useful to start with an understanding of that
19:31 johnb Kados;  The order is subject heading, sub heading, (there can be multiple sub headings) then the subdivisions
19:32 Kados:  To answer the question we need the Manual
19:33 thd johnb: what is the subheading as you just described, distinct from the subdivisions?
19:33 kados: all the manuals are in Cataloguer's Desktop
19:35 kados thd but I think Place -- Topic -- Time -- Form are actually 651 -- 650 -- 655 (not sure about time)
19:35 thd kados: just to add to the fun from the concise authorities manual for 008/06 there is a suggestion about the subdivision
19:35    * d - Subdivided geographically--direct
19:35      The heading may be followed immediately by the name of the specific place to which it is limited without the interposition of a subdivision for the larger geographic entity.
19:35    * i - Subdivided geographically--indirect
19:35      The name of the larger geographic entity is interposed between the heading and the subdivision for the specific place to which the heading is limited.
19:36 kados: time is 648
19:36 kados there ya go :-)
19:36 johnb Kados: some headings allow further specific subheadings corn - anatomy  
19:37 kados johnb: wouldn't 'anatomy' in that case be a subdivision?
19:37 johnb Kados:  I am looking it up on my system to see how it looks\
19:38 thd kados: but all those are repeated as subdivisions within each subject so we can have very complete subjects no matter what whether type of subject was specified for $a
19:39 kados i guess I dont' understand why we have the 6XX fields ... and then within each we seem to reproduce what they already provide
19:42 thd kados: because we want to give the heading or most important part first in $a
19:42 kados: $a is not equal to the others
19:44 kados thd: do each of the LC's subjects correspond with a classification code?
19:44 thd kados: yet no single type of subject is more important except if it is given first for a particular record or the cataloguer has bothered to use the importance indicator
19:45 kados: now you are asking the good questions
19:45 johnb Kados:  Ok i just look up cats on my system
19:45 thd kados: about one third of them do
19:45 johnb Kados:  cats has a sub heading in the lc print manual under health
19:46 thd kados: keep thinking about that and you will soon know many secrets
19:46 johnb Kados: my system handled as _x health
19:47 thd: To answer your question, yes and no
19:48 thd johnb: were you answering this question?  what is the subheading as you just described, distinct from the subdivisions?
19:49 johnb Some headings have allowable subheadings that are specified in the LC subject headings.  
19:50 kados johnb: I'm less interested in what the allowable subjeadings are and more interested in where they are and how they should be represented in a hierarchy
19:51 thd johnb: do the subheadings go in $a with some punctuation separating them from the heading?
19:51 johnb Kados:  The usual rule is start with the subj heading then the sub headings if any then the subdivisions
19:52 Kados: in my system it looks like the subheadings are listed as _x
19:52 thd johnb: do you have examples of subheadings?
19:52 kados johnb: what are subdivisions if not $x?
19:52 johnb Kados: I have got a full book of them
19:52 Kados: give me something from a to c and I can cross reference
19:53 kados johnb: in my reference, $x is listed as 'General subdivision'
19:53 $v is Form subdivision
19:53 $z is Geographic subdivision
19:53 so they are all subdivisions
19:53 I think
19:54 thd johnb: we have an examples from LC showing $x following $z
19:54 johnb Kados:  according to LC they are called subject subdivisions
19:54 Kados:  According to LC terminlology the order is
19:56 kados: Headings, subject subdivisions (specific to the headings), then the four categorical subdivisions (non-specific to the heading)
19:57 kados johnb: so that would be:
19:57 $a
19:57  $x
19:57 thd johnb: how does one determine what is specific and non-specific to the heading
19:57 kados    $y
19:57 thd ?
19:58 johnb kados: The only way is to compare it with the authority file
19:59 kados bummer, since we don't have access to authorities in many cases
19:59 thd johnb: I think I have an answer for my own question
19:59 johnb Kados:  Sorry I guess I am not explaining it well
19:59 First there is a heading
20:00 Then, in some case, actually not very offten LC creates subdivisions under that heading
20:00 Kados:  these subheadings are specific only to that heading
20:01 thd johnb: a subdivision specific to the heading would be $x specific to 650, $z specific to 651, etc.
20:02 johnb Kados: each line is it own subject heading there is no sharing between the subject heading lines
20:02 thd johnb: yet we find that in the case of $z geographic subdivisions specified in 008/06 in the authority record for there may be a difference for that particular heading.
20:03 johnb Kados: yes because the catloger is authorized to create whatever entry he/she thinks is approriate
20:04 thd johnb: from what manual does you specific and non-specific subdivisions reference originate.
20:05 johnb thd: Library of congress; subject headings
20:05 thd johnb: I want the system to create everything for the cataloguer
20:05 johnb thd: the catalogers will just change it
20:05 thd johnb: do you mean the introduction to the printed manual of LCSH?
20:07 johnb thd: well I am using the first volume including the Introduction, and the subject heading lists from a to c
20:07 thd johnb: the cataloguers should be having fun reading the books
20:08 johnb thd: most cataloging in public libraries is just copied from MARC record provided by a vendor
20:08 thd johnb: I used to have two copies of that manual but my room is too small so I do not have ready access to either copy.
20:09 johnb thd: well when I went to Library School there was only two vol. now there are four!
20:10 Kados:  Is there anything else I can do for you tonight!
20:10 thd johnb: yes, copy cataloguing is the way to go unless the only records to coy from are ones which were not created correctly or none.
20:10 johnb Kados: maybe redesign subject headings so they 1 make sennse 2 are machine readable
20:11 thd: most librarians just set aside a book and wait for somebody else to catalog it!
20:11 thd johnb: they are much more machine readable than when the printed manual was only two volumes
20:11 kados http://kados.org/hierarchy.txt
20:11 so which is it in that list?
20:12 thd johnb: that is what I did at my bookshop
20:13 kados: I vote for number two
20:14 kados johnb: ?
20:14 johnb Kados: I beleive number 2 is the best
20:17 Kados:  beleive it or not not one of the sub headings is listed under architecture history should be Historiography
20:17 kados johnb: I got this record from LOC :-)
20:18 johnb Kados: Welcome to the wonderful world of cataloging
20:18 kados hehe
20:19 johnb Kados:  You also have to remember that there is alot of aold cataloging out there that follows older rules and standards
20:19 Kados: i still like the second solution, seems easy to read and makes sense
20:20 thd johnb: OCLC has a huge project to fix that in their records
20:20 johnb thd:  they have been changing the rules since Thomas jefferson sold his collection.
20:21 kados hehe
20:21 johnb thd: library catalogs are living entites that are never never completly standardized to the rules.
20:22 thd johnb: one day the machines will revolt against the living and put everything in good order
20:23 johnb hey guys its been fun but they are starting to shoot motars my way (fireworks) got a get out of here before the building catches on fire.
20:24 kados johnb_away: thx for the help
20:24 thd johnb_away: library catalogues are supposed to enable people to find things.  Between rules changes, changes in headings, and cataloguer differences: finding things is much less reliable than it ought to be
20:28 kados: did you understand the difference between heading specific and non-specific subdivisions and the one exception for subfield sequencing?
20:28 kados no
20:28 please expound on it :-)
20:31 thd kados: so if you have topical subject 650 $a will always be followed by topical subdivision $x, if present; unless the 008/06 from the authority record for that heading allows $z to come between them
20:34 kados: similarly, chronological subject 648 $a would have chronological subdivision $y just after, if present; unless 008/06 allowed $z to come between
20:35 kados: the issue is moot for geographic subjects 651 where is does not matter if $z comes between $z or not
20:38 kados: I am not certain how this would apply to genre/form subjects 655 but they tend not to have subdivisions in practise
20:39 kados: is that clear?
21:11 kados: I am back so was that clear?
21:11 kados thd: not entirely
21:11 thd: we need a complete specification before I rewrite anything :-)
21:12 thd: because I don't want to code this twice :-)
21:14 thd kados: what this means is that the sequence of subject subdivisions is changeable and cannot be known correctly except by reading the record and hoping it was correctly sequenced in the absence of authority records.
21:16 kados: I can try to check the Chan book and some other manuals tomorrow if the guard will honour my pass at the local library school
21:17 kados: otherwise this is another case where you need a subscription to Cataloguers Desktop
21:18 kados :-)
21:19 well the short answer is I think I can figure out how to get a 'good enough' system together
21:19 thd kados: otherwise I can ask on the autocat list, but I would always heck the best references that i can find before asking there.
21:19 kados but like all things, we should keep it in the back of our minds what the ideal system would be capable of
21:19 asking on autocat might not be a bad idea
21:19 I've thought about that myself
21:19 though I'm not sure I'd be comfortable posting a link to Koha just yet
21:20 it needs to be prettier first
21:20 and I'm not sure how to acomplish that
21:20 thd kados: I have already asked questions with the explanation there that Koha wanted to know
21:22 kados: was Katipo not going to make some very pretty templates for 3.0?
21:23 kados so I hear
21:23 but I mean really pretty :-)
21:23 I think to do that we're going to go outside our current community
21:24 thd kados: well really pretty needs image maps and other things which really slows down page loads
21:26 kados: I am still using the ugly websafe colour palate

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

koha1