I think it is a good idea to collect early notes on topics and trouble that most classification systems runs in so they can be addressed as soon as possible:

* material types: How is (if it does at all) OSC going to deal with maps, music recordings, movies, software etc.
* Material that skirts the fiction/non-fiction divide: this is a classic that has had people their heads for a while. There are not that many works here, but those that exists are a thorn in the side of cataloguers.

Message edited by its author, Jul 8, 2008, 4:05am.

How about books in other languages?

Do they nestle alongside the books. Or do they get a different top level classification?

Any decent classification system should be able to handle multiple-language materials within its own framework - if a particular library *wants* to separate out all the Spanish books, they can (and they can easily hack the shelfmarks to do it) but the classification shouldn't mandate it.

What always bugs me (a non-librarian) about classification is the difficulty of finding books in a bookshop when they could sit easily in multiple categories - I've found books on bridges, my pet subject, shelved under architecture, engineering, transport, history etc in the past. Is the idea to develop a classification system that allows books to be labelled with multiple facets such that I could find the same book under several different top-level headings?

Yep it bugs me too. But I have few answers. Faceted classification helps findability in an online system, but doesn't help too much with someone wandering around shelves.

Stuff to do with computer software is often spread out - for example Agile web development with rails is in TK5105 (according to the LoC) and not QA76 like most of the rest of my software development books. My books about computer graphics and graphics APIs are in T385. Calendrical Calculations (a book about algorithms) is in CE12 with the other stuff about calendars. My book about LATEX is in the Z235s with other books on DTP and typesetting systems.

When I bought these books the bookshop had them in their computers section and that is how I relate to them. Given the classification outline I could have worked out where to look for the book on Latex and Calendrical Calculations (given time). However who would look for a book on web development in TK5101-6720 Telecommunication - Including telegraphy, telephone, radio, radar, television? Yes it seems that TK5105.888 is all about the web but you cannot tell that from the outline. Why not somewhere in ZA3201-3250 Information superhighway?

Computing is quite a difficult case. Individual works may be spread out all over the classification system (presumably Beginning Perl for Bioinformatics is most likely to be somewhere in biology*, and Web Programming with Perl 5 will be in TK5105.888, and Programming in Perl would be in QA76. Graphics Programming with Perl is in T385. Who knows where Programming for Linguists: Perl for Language Researchers and Perl Programming for Medicine and Biology would be. From a computer programmer's POV all of those should be together - they are all primarily about computer programming (albeit in different domains). Someone from a different discipline might well disagree with me though.

* I checked and it is QA76 along with the other software development books. Bioinformatics, biocomputing and Perl : an introduction to bioinformatics computing skills and practice is in QH324. Perl for Exploring DNA is in QP625. All of those three books cover very similar material.

bduguid: Finding a book in the catalog and determining it's place in the shelf are two different things, for the reason you have suggested: some (most) books fit more than one subject heading, but each book must go into a specific place in the shelf. I don't think we need a rigid, coded system for tagging the subjects in books, thanks to the fact that we have computers. But libraries do need a way to locate books in sheves, and hopefully colocate them with other related books. This is what the Open Shelves Classification is all about (I really like the fact that it has "Sheves" in the name for that precise reason).

Message edited by its author, Jul 8, 2008, 10:23am.

circeus, you mentioned one of the potential stumbling blocks as being material types. That reminded me of something. Several people have talked about modeling the primary categories on the "bookstore" model. But at a bookstore Graphic Novels, Comic Books, and Manga have a separate section despite covering many different genres, simply because they are of a specific media. Even in my local library, all different types of comic books are shelved together, so that Peanuts (children's cartoons) is next to Y the Last Man (adult scifi) is next to Boys Over Flowers (teen romance)--in the non-fiction section.

Personally, I would prefer to have graphic novels handled just like regular novels in the fiction section of the OSC, but it is something we need to consider.

I can't help but devolve discussions, which usually leaves me with cavemen.

That being said, I was thinking about how the beauty of creating a classification system now lack of limits to each item's record. We don't have to worry what three things we're going to identify the item as - we just need to find a place to shove the book.

Since we're talking about where books will physically live, then really they could be in Library of Congress order - by size. We could add a color sub-scheme.

Now, I realize that's silly, since if we're talking about public libraries as our primary example, people will often be browsing. I would like to see an order based on how often items are checked out. I know this won't work, since that would create some major side-by-side randomness, but how glorious would it be to look on the first-level shelf to see if any of the bestseller novels or new-release DVDs are in. Or poke around the back end, looking for Death's Flotilla. What? Is that not feasible? :)

*Aside*
One lesson I learned from Dewey is that when the numbers change, catalogers tend to start cataloging new materials the new way, and leave old items the way they are. When I was a children's librarian, our nonfiction animals section (DDC 500s) were all over the place, because of this. My hope would be that this would get to some sort of full system before anyone tries to put sticky labels on the spines of books.

>Ninjapenguin

When treating Graphic novels (GNs) in DDC, I customized the system by classifying comic strips and GNs in a new "literary form" (805) which was appended with the genre numbers (-301 to -308 minus 30-) from table 3. 1 ("short stories") was used for comic strips that did now diractly qualify for another category.

Hence Garfield is 818.51, but Narbonic would be in 818.5762. Obviously my custom system sorts by genre over time period, which are in 01-09 (placing Agatha Christie in 823.872). I won't pretend it would be ideal for a library collection (but then I can't think of a library that uses its nonfiction classification for its GNs and strips volumes anyway, except for stuff like historical collections), but it certainly works well for my mostly-contemporary-fiction collection.

>sonyagreen

I've run into issues of this, primarily with the "random board game" (somewhere around the 790s IIRC) section, where Scrabble book got wildly variable placements. Also because this area also receives RPGs, I've seen Scrabble books shelved in the middle of gamebooks (another good example of classification stumper), which I did not puzzle out until years later when I first looked at that section of the DDC abridged vs. complete.

Message edited by its author, Jul 8, 2008, 4:00pm.

>1

The question of maps, DVDs and so forth is a serious one. LT doesn't have so many of these, so it's going to be much harder to test with them. Do you have any proposals? Can you survey how systems have fallen short here?

I'd like to suggest that the structure of the call number itself should provide for format/material type, since this is how actual libraries shelve their materials. With apologies to Ranganathan, I imagine something like material:class;properties/space,time.

>11

Not always, I think. Certainly there are some edge media here, that could go together or not.

If media were a basic top-level face, though, could the rest of the system be identical? I think not. Not unless "science fiction" becomes "techno" when applied to music :)

As a shot in the dark, I'd posit that we should follow the usual practice in many library by at least offering a segregated (or at least optional different placement) specific form of classifications for these. After all, it's not like libraries normally shelve DVDs (except some documentaries, I believe), CDs and maps alongside books, and they are not that keen to apply whatever general system they use to it, preferring custom author or genre-based systems.

Because of this, an option might be of specifying a method for shortening the call numbers of these specific cases in relation to the most closely related schedule format, so that e.g. if physical geography and geomatics is B34X.XX..., then maps when sorted separately are truncated to X.XX or BX.XX. here's a quick partial outline sketch for movies (Tv series being different, but similar):

-fiction by country (with an option where countries are subordinate to genre)
--General entertainment by genre
---children movies
---historical (including biographies and fictionalization of real events)
---Horror
---Speculative fiction (including supernatural, fantasy and steampunk)
---action/adventure
---comedy (including parodies and romantic comedies)
---Romance
---Animated (subdivided by above genre)
--Independent by genre
-Documentaries

Message edited by its author, Jul 8, 2008, 4:20pm.

When creating a system that kids will ultimately use, there needs to be a good method for cutting call numbers shorter. We (in schools) generally stop after three places when we can. The numbers get too long and the kids get confused.

One obvious shortcoming of Dewey that's been mentioned other places is the squishing of all religions other than Christianity into a small chunk of numbers.

But couldn't classes exist to define those cases? If one item is Book-fiction-scifi, can't another be CD-techno-British, and another Audiofile-techno-American? Even in DDC, the position of a number can tell you what it refers to.

Not to get too crazy with colon classification, but couldn't some form of punctuation help kids understand the internal organization of a call number? If it's clear that there are just four slots, for example, and their separation is indicated either with punctuation or by starting a new line, then kids can process the meaning of each slot separately.

Why can't physical location be just another facet in a faceted system. In computer terms this would be the physical address that results from resolving a pointer. This physical location could be DDC, LCC, SAB or whatever the specific library, bookshop etc. uses. The new system therefore would serve as a wrapper for the old system and legacy classifications could be easily incorporated into the new. I don't really see a future where any book depository of any size does not have a computerized catalogue and better yet an internet acessible catalogue. Look up the book by whatever attribute you want, obtain physical location, obtain book.

I think we're all gravitating towards a faceted system for the material type because it is important to the shelving in our libraries right now. Basically, libraries already have ad hoc facets when they put all the DVDs in one place or all the CDs in another. I was actually quite lost and confused for a while until I realized my local library was shelving audio books alongside the printed books from the same author. To some extent, that is why material type is not currently part of call numbers, because the systems are leaving it up to a judgment call by the librarians at each library.

This is why I support a faceted system similar to what circeus proposes, but without the facet at the beginning of the call number. I don't want the material type to predominate in the shelving order. It really ought to be more optional than that. In fact, for spine label purposes each facet should probably have its own line. Something like:

XXX-XX-XXXX
CD

Also, another useful facet would be language. Whichever facet a particular library chooses to use in shelving could even be moved up on the spine label to indicate shelving order, but that does start to get pretty confusing if the facets become less obvious than things like MAP and RUSSIAN.

>5

I think the biggest issue with these specific titles mentioned is not the classification system being used on them, but rather that they are likely cataloged by people who haven't read those particular titles or aren't intericately familiar with the subject of the material.

I'm in agreement with what Tim has said other places that tagging as well as dyvving up work to subject experts and people "who have actually read the book" will help counter this issue, as opposed to relying on a classification schema to solve this. I think that the original OCLC model (with different libraries with different subject fields and specialists contributing records) was intended to account for this, but due to other issues it just never worked out that way.

>18

We just use prefixes in our library: CD, DVD, Map, Video, etc. The problem I find is that these have to be entered manually, into the same database field as the rest of the call number. I wouldn't be surprised if this was an issue with our local ILS and other systems handle this better. But ideally, imo, a prefix would be automatically displayed based on the specific material designation. I don't really consider it part of the classification in this regard.

>18

I completely agree that each facet should have its own line, potentially with punctuation. In fact, if facets were enclosed in punctuation, those who were building a number would have the option to leave a facet out entirely if it didn't suit their purposes. Don't want to separate the material types? Just leave out the first facet! But this only works if there are indicators for the beginning and end of each facet.

I think facets should be at the top, not the bottom, and should be optional and location-specific. I picture the bottom being used for alphabetical ordering—usually author, but it could be different for different classes. Also edition, if that's important to note in the call number.

Eg.,

AUDIOVISUAL 12-142-234 SMITH
AUDIOVISUAL 12-142-234 SMITH 2007

Can I get a name for the latter part—in cataloger-speak?

Author names are traditionally made via cutter numbers (so that all Smiths have their books guaranteed to be together rather than mixed), I don't know how DDC and LOC call post-number additions.

Message edited by its author, Jul 8, 2008, 10:27pm.

What's the reasoning for Cutter-izing in this way? It seems to me that some of the uses should be in the main part of the classification.

Why not, when necessary, move from one system to another. To take the example at http://www.itsmarc.com/crs/cutr0028.htm

Why not do

134-123-Leonardo-34

Where 134-123 gets to you Italian artists of the Renaissance, Leonardo names the artist, and 34 give you his letters, or whatever?

well, obviously DDC and LOC (and more often than not often libraries) use Cutter numbers in different ways. I worked in a library were fiction was simply e.g. CHR.agh for Hercule Poirot's Christmas!

This case, however, is a peculiarity of that part of LC's classification (works on an artist/writer). I don't think it is used outside those ranges.

In fact, this is an application of the frequent use of cutter numbers for ranges of topics (similar examples can be found for subtopics of Maths, and geographical areas).

Message edited by its author, Jul 8, 2008, 11:33pm.

What do we call the "patterning" of DDC—essentially a little facet, with an embarrassed look? I mean when the languages repeat all over the place.

Is this a useful feature? If so, to whom? I worry that it can lead you astray. You get your language list together and start applying it here and there, until you've got a section for Ancient Greek speakers in Nova Scotia.

>24

One reason to Cutter-ize is that it's a way to alphabetize without actually writing out a full name, which might not fit on a spine label (especially with the increasing prevalence of hyphenated last names - can you imagine putting a name like Schwartzman-Atkinson on a label that's only a 1/2 inch across?)

"S39" or "S399" (depending on whether you use the 2- or 3-figure table) fits much better on a small label, and when you put things with the same classification number on the shelf in cutter number order, the end result is the same as alphabetizing by the whole name.

Someone point me to the tables?

Gosh, I'm going to Nova Scotia next week - I'll have a look round for the ancient Greek speakers.

I'm still thinking through the language thing. What is the value in separating by language at a top level? I don't see one (yet). Language is a MARC code kind of thing to me, not part of the classification number. Maybe in literature (an area I've never catalogued in or studied formally) - could it be useful to separate the works of Shakespeare in English from the works of Shakespeare in Chinese? But science? I don't see it.

Getting back to Cutter numbers: In my library, we use Cutter numbers for a) filing within a subject by author and b) making a call number unique. Having one unique identifier (the call number) for a book had value at one time, because it was your guarentee that you were pulling the right book off the shelf. Now, the bar code provides the uniqueness and we don't worry so much about duplicating call numbers. Is there any value in uniqueness here?

>29

No, I agree. I'm thinking of grammars, surveys of literature, etc.

Right—good question about uniqueness.

26> Any repeated feature is going to lead you astray eventually, but it does make the system a little easier to negotiate. The 400s mirroring the 800s, however imperfectly, makes it easier to remember where certain sections were. What, fundamentally, is the problem for having a section for Ancient Greek speakers in Nova Scotia? If you're getting down to that level, I'd be more worried about the opposite; being able to shelve books about French speakers in Nova Scotia, but not Chinese speakers in Nova Scotia. When that happens, and it turns out that now Chinese is the second most spoken language in Nova Scotia and the University of Nova Scotia is publishing several books a year on Chinese speakers in Nova Scotia, then you have to revise the system and move books around, which is worse than having the blank space in the first place.

See http://www.librarything.com/talktopic.ph... , message 14. Does the bookstore model—aped by Maricopa County, famously—mean that Cutterization is unnecessary?

>31

Agreed. But if you're mechanically transposing, say, a classification of legal systems into a culture that doesn't use those concepts, you've gone too far.

>28

Here's the basic Cutter 2-figure table: http://www.itsmarc.com/crs/cutr0020.htm

There don't appear to be good, free online sources for any of the others that I can find in a quick search (anyone else know of a good source?), but there are several different versions available...

Cutter 2-figure
Cutter 3-figure
Cutter 4-figure
Cutter-Sanborn 3-figure
Cutter-Sanborn 4-figure

Can i propose that any Cutter-ization be postponed anyway? It's not necessary to have an answer while developing the first levels.

re:Cuttering

Most of the print tables for Cuttering with DDC have been eliminated in favor of OCLC's online version (suprisingly enough, it's free to download and use):
http://www.oclc.org/DEWEY/support/progra...

It's hard to get an idea of how the original manual print tables looked and worked from that, but it will create numbers. I do have an old print copy of the Three Figure Cutter-Sandborn table I could probably scan and post and excerpt from if anyone was really desperate.

>26

I adore the patterning of the DDC. it's one of the things that keep me using it. Again, it's not without its flaws, but I find the "standard subdivisions" (as they are called) to be very useful. They allow me to expand out, so I can have Fashion design (746.92), fashion of the 20th century (746.920904), fashion designers (746.92092), fashion design in/from Japan (746.920952), which are all necessary in the library where I work. (Most libraries would lump all under 746.92.) I like some of the other subdivisions even thought I don't use them as much: -022 for pictorial works, for example, would benefit a lot of our patrons if applied well, I think. Lately I've been lamenting some sort of subdivision that would cover the whole "green" treatment, as we have materials in every section (fashion, interiors, manufacturing) that deal with that, but it's not feasible for us to pull them out into their own section.

I admit though that I'm a cataloger and just because I like the way the system works and repeats and patterns doesn't mean it's the best thing ever for the patrons. But I have to say that patterning is probably my major reason for personally preferring DDC and I'd be dissapointed to see a move towards something more arbitrary and not as scaleable.

>37

Just another vote for patterns/mnemonic devices to be built into the system. It's why I prefer Dewey over LC. The problem with Dewey, of course, is that it can be hard to tell where your standard subdivision begins. Punctuation, as in colon classification, provides one more visual cue to complement the mnemonic repetition. I like that Japan = 952 whether it is before the decimal or not, but wouldn't it be even more clear if fashion design from Japan looked something like 746.92/952 ?

I think the first rule should be that the system output should always be able to be put on three or less rows, short enough to fit on a 1.5 inch label in size 12 font!

Or something like that. These are going to go on actual books, some of which are quite skinny. It's hard to find them when you can't see the details.

Okay, pivy, spell out how these patterns would work here. What cues them? Surely they are context-dependent. 43 can't always mean French. Sometimes it has to mean "bas relief."

Standard subdivisions are a special section of the tables (T1), you can append them to any number unless the schedules tell you otherwise.

From OCLC:
"Subdivisions found in Table 1 that represent frequently recurring physical forms (dictionaries, periodicals) or approaches (history, research) applicable to any subject or discipline. They may be used with any number in the schedules and tables for topics that approximate the whole of the number unless there are instructions to the contrary."

Subdivisions usually begin with 0, so you can see in the classification where the subject ends and the subdivision begins. There are ten major classes of subdivisions, just as there are ten major disciplines in DDC. I can't remember them all off the top of my head, and my copy of the schedules is at work, but I know -074 is museum exhibits, collections etc.; -09 is "historical, geographic, persons treatment"; -092 is specifically personal treatment (it conviently parallels that 92 of biography), -022 (IIRC) is pictorial works, etc. You'll see them indicated with the preceeding hyphen because that indicates that they are always tacked on to something else; they never function alone. These subdivisions are the same regardless of what number you append them to (unless indicated). So museum exhibits of textile arts= 746.074; museum exhibits of fashion design=746.92074; museum exhibits of painting = 759.074, etc.

What gets me, of course, are those "unless otherwise indicated" sections. If it's going to be consistent, it should be consistent for all sections. I like that works on japan are classed uner 952, and Japanese art is 709.52, and japanese fashion design is 746.92/0952 (the slash indicates where the subdivision starts--you will often see this in records, or sometimes an apostrophe is used as well--this shows libraries who might want to truncate the number an appropriate place to shorten). That "952" consistency is a strength, imo.
But, is japanese language 520? is it 452, or 409.52? The 400s are one place i know of where things don't follow the regular rules. Literature (800s) is another. they each have special tables that trump Table 1 subdivisions. I'd love it if there was a way to just have the one table, one single set of subdivisions that could apply to anything and everything, although I'm not quite sure if that's even possible. I think some of these current issues might stem from evolution, and accounting for addition of subdivisions to already established classifications that don't fit the consistency patterns, but that's just my hypothesis.

I'm not sure if that helped at all or if it was just more confusing.

BTW, France would be -09/44, not 43. :)

So, I think the message is that, if we're going to use standard subdivisions, we need some sort of marker for them, right?

To be radical, what if we use letters for the non-standard parts and numbers for the standard subdivisions?

>20

I'll expand on using prefixes for special media. Let's remember we are talking about shelving here. My library shelves entertainment DVDs in their own area, but non-fiction DVDs with books on the same subject (eg., exercise videos with exercise books). The call numbers reflect this. Entertainment DVDs begin with the code 'DVD'. Exercise DVDs just have 613.71. The fact that exercise DVDs are DVDs is entirely omitted from the call number because it does not affect shelf placement. In this respect, the 'DVD' code is part of the call number, but it designates a place in the building more than the material.

Going back the the cutter issues...
One of the main problems with cutterized names, is that they aren't inherently user-friendly. The average patron is not going to know by looking at the spine label that when they see K587, they've reached Stephen King vs Barbara Kingsolver. (At the point they get close to the right name, they usually stop looking at labels, and start looking at the full spine for title and/or author.

Cutter numbers definitely give you the unique place on the shelf, but if we are going for a browsing shelf system, I'd argue strongly for non-cutter names, even if they have to be truncated somehow for spine labels.

Message edited by its author, Jul 10, 2008, 1:31am.

Again, do we need to decide this? As something that's tacked on after the regular classification is over, does this need to be resolved?

Going farther, provided that you use semi-intelligent algorithms, do you even need to have the same answer between libraries? Can't you just store the base classification and the author name and then decide how to treat that as a display issue?

Message edited by its author, Jul 10, 2008, 1:32am.

Well, depending on how the different facets are organized, the author name could be playing a major role in how this classification scheme shakes out. If that's the case, not addressing how the names are represented could cause problems further down the road.
At least, that's my impression from the suggestions I've seen so far.

That being said, if the name plays a lesser role, then definitely it becomes a display issue to be decided by individual libraries, but some standards should still be applied.

What I want to do is take SOME issues off the table. If we can do that without totally screwing ourselves, we'll make some progress. So long as nothing at all is decided and everything is equally in play, we'll never get anywhere.

42

UDC (and Ranganathan) use unique markers for all facets (and mostly it's punctuation). So (forgive me here for using an unrealistic example I use to illustrate a point in class) - "Privately owned mobile libraries in Belgium in the 19th century - a bibliography in Swedish" is 021.65.003.52“18”(493)(01)=113.6, where 021.65 is the number for mobile libraries and the rest is from tables, signalled by punctuation. .00 is point of view, "" is for time, () with numbers beginning in other than zero is place, () with numbers beginning in 0 is form, and = is language. And there are tables for all of those things (and a few more).

You know, I think author name is a red herring. IMO, the important first thing that the classification should is subject division.

Fiction aside, what the average user wants is all the gardening books together, all the cookbooks together, all the biographies together, all the production economics books together. I would even argue that format is irrelevant because not everyone chooses to separate their collection that way. Top level needs to be about subject.

49:

I agree with you entirely about the irrelevance of format. If separation by format matters to an individual library then the system should be flexible enough to accomodate them by simply adding a prefix to the start of the number.

>48

I have to say, that's fiddly! You'd need to be Victor Borge to pronounce the things.

http://www.youtube.com/watch?v=lF4qii8S3...

What about the age old problem of mixed formats. "Kits" if you will. Books with CD-Roms. Items with CD's, books and other stuff??? Do we mix them all together by topic or is the format the most important thing.And if so which format.

I think this has to be a library-by-library decision.

53: agreed, which is why I think this system should ignore format as much as possible, leaving room for individual policy decisions.

Yes media needs to be ignored, it changes to fast. just in my lifetime music has come on 45 & 33 1/3 RPM discs, 4 and 8 track cartridges, cassettes, and now CD's and MP3 files. However, do we want the system flexible enough to catalogue all of a libraries holdings? Music, films, photos, microfilm, and computer files?

It should ignore format, but not... well, what's the right term for the distinction between recorded music, recorded video and books? Assuming that a recorded book has the same call number as the paper book--with an option for the library to add "AUDIOBOOKS," should the sheet music have the same call number as the CD of the same? A DVD of Hamlet and the book?

What's the standard here?

We really should leave all media to each library. Even in my one little library, we handle media differently depending on the type of media and the collection involved. Fiction videos are separate from fiction books, but non-fiction videos are integrated with non-fiction books, but non-fiction audiobooks and filed at the start of the fiction audiobooks, except language instruction which is integrated with language books. And this is just the beginning. Even in one library, media is too messy and too tailored to local need to standardize.

We could recommend "best practice" nomenclature for media which are applied optionally and as the library sees fit. If every library used V instead of VHS, VCR, VCS, or BETA, how useful is that really?

I like the idea of "best practice" recommendations as a way to give libraries who don't want to come up with their own systems a means of indicating format (or language as we've discussed elsewhere), while leaving it an optional part of the classification for those libraries who have their own rules and ways of doing things that work for them.

Ideally, you can add "VHS" or whatever automatically, if the library checks some radio button, as we know the media.

Message edited by its author, Jul 13, 2008, 11:47pm.

I don't think there is a standard for format. In my library we have a "media" section where vhs tapes, dvd, cd-roms and audiotapes are filed together, but separate from the books. My local public library separates everything.

My preference is to ignore format (and language) and allow the library to choose how they file the material. So the DVD and the audiorecording and the book of Hamlet in English and the French translation of Hamlet would all have the same number, but libraries could add a location before (or after, depending on local practice) the number to tell users if the items are filed separately. The big advantage to this is online browsability - a user who finds one version of Hamlet can check that call number to see what other formats or translations are available.

If we're going to ignore media, we still should ignore that large pieces of the cataloging will be music or movies. I don't think we want all music to be shelved under Arts - Musical Arts - ETC. Music should be a top-level thing.

>61 Should it though tim? In 56, you brought up a good example: a CD containing a music recording and sheetmusic for that same piece. In my head these are just different manifestations of the same thing, and the whole "where does media go" thing aside, they should have the same call number, in my opinion. Now, with this point of view, maybe the section containing both of these things should be fairly toplevel, given how big it could be in a library with lots of audio recordings, but I really do think they are the same sort of thing. Similarly, dvds might be fiction or they might be a documentary or something, and should be numbered accordingly. If a library decides to shelve all fiction dvds (i.e., "movies") separately from their fiction books, thats fine, but I would think the call numbers would be in the same section, just the location prefix would be different. At least thats how I had envisioned people using the prefixes that have been thrown about around here in the past. E.g.: DVD JK-45-10-6 for Hamlet the movie on DVD, and JK-45-10-6 for a a book of the play (assuming no prefix means books, but whatever).

Just my four and a half cents.

No, I think you're basically right. But the idea of "movies" as Fiction DVDs is a strange one. Not a play DVD? (How about some Mammet play made into a movie?) "Movies" are their own art form, not a different medium of an existing art form.

Then again, if you're not going to separate out the movies, having Sci-Fi movies in with the Sci-Fi book does have a certain appeal.

You are right. I got carried away. Obviously Hamlet would probably be a play DVD, and belong with the plays.

I have the Royal Shakespeare Company's brilliant stage production of Nicholas Nickleby on DVD - it's a filmed play, not a movie. There are several actual movies of Nicholas Nickleby too (although I don't own those), not to mention the original book (which I do have). Filmed play should be classed with the novel and the other movies in Fiction.

BUT - I think in the case of a filmed play where the original work IS the play, it should be separate from fiction (presumably in a Plays category)

Now, what about a movie where the script was original (not based on a book or play)? Is that Fiction or Plays? Example: Star Wars. If you're cataloguing the script, it could go in plays, so the movie ought to go there too, but then where does the novelization go? I think movies and their associated print versions need to go in fiction. Except documentaries, of course...

65> It seems to me that what you are arguing here is that plays shouldn't be separate from fiction. Just as the media doesn't matter then it also shouldn't matter that a work of fiction is written as "blah, blah, blah" Horatio said and a play is written Horatio: blah, blah, blah.

>65

Wait, but the PLAY of Nicholas Nickleby would go in a *plays* section, wouldn't it? Then the filmed version of that play would go in that section—absent a media split-out section, as discussed. So

Nicholas Nickleby, novel -- Fiction
Nicholas Nickleby, play -- Plays
Nicholas Nickleby, filmed play -- Plays
Nicholas Nickleby, movie adaptation -- Movies
Scholarly study of Nicholas Nickleby -- lit crit or whatever
Nicholas Nickleby radio play -- Plays
Nicholas Nickleby audiobook -- Fiction

Right?

I want to know how LCC and DDC handle these issues—and Bliss, UDC, etc.

I can only speak towards Dewey in which film and music fall into the arts in the 700's and plays fall into the 800's with literature.

Just to confuse things in my library (this system predates me) we break feature films into their own category, including adaptations of plays and musicals, with the exception of Shakespeare. Then Scores and audio recordings go into their own sections with their own classification systems to be truly confusing.

And yes I would love to change how we do this.

>69

To what? What's the problem with DDC in your mind.

Incidentally, I think that should be our major question as we build this out—not what abstract representation of the world of books is most appealing, but how can OSC improve upon other systems.

I'm sorry for the confusion, my problem is not with Dewey but with the exception we make to it and for our use of two other systems for music (ANSCR for music recordings and I believe Dickinson for our scores).

Great. Never even heard of ANSCR. But then I've never thought about music classification—not even as much as image classification.

>67 - Well, my thought is that the original form of the work dictates where subsequent adaptations go.

Therefore, since the original of Nicholas Nickleby is a book (well, newspaper serial, but print, anyway), all subsequent adaptations (plays, movies, Classics illustrated comics, etc.) go in Fiction.

With Hamlet, the original work is the play so all subsequent adaptations (filmed plays, movies, pop-up books, etc.) go in Plays.

The Odyssey, by Homer is poetry, so subsequent adaptations (play, play script, movie, 8-part tv series on DVD, etc.) go in Poetry.

It makes lovely sense to me but might be too complicated. The whole thing of having to know where the original work came from may be problematic when most kids seem surprised to hear that any movie was based on a book.

Also, as mentioned, what if the original form was like the Hitchhiker's Guide to the Galaxy - a radio program? Or Yes Minister, which was a TV series first? I guess TV and radio programs counts as plays, but then all the subsequent books end up in Plays, which doesn't make sense...

Hmm - it seems >66 (jjwilson61) is right. I think I've talked myself into saying plays (including stage, film, tv & radio) ought to be classed with fiction. Won't break my heart if nobody agrees with me, but at the moment it actually makes sense to me.

I much agree that it's weird and complicated—like the lead character in High Fidelity's "biographic" record organization.

With all these FRBR related issues (manifestations vs. formats, etc.) being touched on, perhaps we need to make it a priority in the process to figure out how important a role FRBR and FRAD (authority control) should play in OSC.

It seems that a lot of people are keen on adding some kind of notation to the call number for media/format, but I think that if the idea here is to build a system that could legitimately be used by libraries as a replacement for systems they use now, there are going to have to be decisions that are left up to individual libraries, and there is going to have to be some flexibility built into the system. I'd say leave it up to the individual institutions to separate or not separate media, and focus on classifying the content in an intuitive way so that you have an intuitive browsing experience within in each section of media, or as a whole. Also, how often do you go to the library and say, "I want Hamlet, and I don't care if it's a movie or the original play?" No, you usually have something in mind. You can see both together in the online catalog, but they don't need to be together on the shelf.
Also, one major downfall of current systems is that new media sometimes have to squeeze into some weird "material designations." You don't want to get caught in the trap of saying, "there are plays, there are novels, there are movies, there is manga, etc. and nothing else," because there WILL be something else, and you'll wish you had room for it.

What I read was a proposal to have an optional designation for media/format possibly prepended to the call number that was free-form but possibly having some standard notations for media in use today, but libraries would be free to use those or invent their own. Sounds pretty flexible to me.

>77

Yes, this is what I get too. We are arguing over an optional recommendation. But this is good. We are distilling the essential from the optional.

>75

Interesting thought. I don't know much about FRAD, but for FRBR:

WORK (the artist's original concept)
->EXPRESSION (thing that is copyrighted)
-->MANIFESTATION (thing assigned a product number)
--->ITEM (thing you borrow)

The work is what gets the subject classification.

FRBR emphasizes that media has two layers of meaning. A music score or musical performance would is an expression, while a publication of the score, a recording on tape, a recording on CD, and a recording in a downloadable file, are manifestations. I suppose FRBRized media designations would look something like:

Music Score (music score in any form)
Music Score-Print (printed music score)
Music Score-MIDI (digitized music score in MIDI format)
Music Recording (music recording in any form)
Music Recording-Tape (music recording on tape)
Music Recording-CD (music recording on CD)
etc.

If nothing else, this spells out that media should not be classified as part of the subject. Perhaps we should have two facets, subject and media. Subject and media are each quasi hierarchical, like Dewey. But either can be put before the other, or omitted, as local need dictates.

To head off the most common question, the 1939 movie "The Wizard of Oz" is considered an expression of a derivative work, not a direct expression of L. Frank Baum's book The Wonderful Wizard of Oz.

#77

Yes, I agree. Language should work similarly as some libraries may well separate out fiction by language, or at least have separate local and foreign language fiction sections.

So we seem to have a couple of optional items sitting at the front of the true call number. These optional prefixes may or may not be used for shelving as the library wishes.

I would hope that there isn't more (or at least not too many more) of these as it could quite easily get complicated.

75 & 78: I'm worried that heading into FRBR territory is getting a little off track (with the exception of the material/language designations). I think we should stick with subject classifications for now at least.

FRBR will be in there insofar as we're going to at least start by classifying *works*, not editions.

But then if this code is to only apply to works then media and language should have no part in them since, at least by the definitions used on LT, an audio book is the same work as the book it is derived from, and a French book is the same work as the English book it is translated from (I don't know if you want to carry over your dead language exception to this new system though).

I think the comment about:

Not to get too crazy with colon classification, but couldn't some form of punctuation help kids understand the internal organization of a call number?

And another earlier comment about parts of call numbers meaning something is telling. I think the point is, and correct me if I am wrong, to create a classification system that was intuitive and doesn't have to be a secret kernel of knowledge passed down by the fraternal order of librarians to be understood. The point is, patrons don't know and don't care to know why their book is labeled a certain why. They just want the book and to find books like it close by.

It shouldn't be about education. It reminds me of an antedote. If a website is designed properly you should have to train anyone how to use it. It should be intuitive. I was at a staff meeting once where someone asked if we could get a staff training (core skills) on using our website!

83> I find it very frustrating the way that people always assume that if something is designed "properly", there would be no need to train anyone on how to use it. Idiotproof systems annoy the intelligent and tend to fail to reach the idiots. To do anything moderately difficult is going to require some sort of education so the user understands how the system works.

>82 I think that is what the consensus here is reaching, yes.

The thought is that each specific library, if they wanted different types of materials, different languages, etc., in different sections, would actually end up essentially having multiple collections in the library, completely separate from one another. Each collection is marked by the magic meaningless undefined prefix people keep talking about, like MEDIA for the media collection, or FRENCH for the french language books, or GENERAL for the general collection of books. Each of theses collections would be placed into its own portion of the library, but internally would have the normal OSC classification of materials inside of that collection.

At least thats how I've been thinking about it, maybe others see things differently.

85: I see two options actually. The first is what you stated, that various item types are broken out into special collections.

But it is also possible if a library wishes to interfile these items and still mark them out as a separate format. At my library we treat our genres in this way. All of our fiction is shelved together alphabetically by name, but genres are marked with stickers. In the catalog the genre label is placed into a non-indexed subfield so that a call # search of the author's last name will still function correctly.

Just so we can make sure that everyone is on the same page, and to sum up some of the last 80 comments, it seems that what we're generally agreeing on is that an OSC call number will include these parts:

1. an optional notation for format and/or language to be used by an individual library as needed (possibly following a system we develop later, or their own system)

2. a subject notation to bring together works on the same topic/sub-topic - this is the bit that we need to develop as part of this project

3. a filing notation to designate the shelf order of books with the same subject notation (the format of which we're going to discuss at a later time - per Tim's request)

Is this correct?

>87
I think you've mostly got two and three a bit mixed up, or at least in our case they are both the same system. The OSC will be a filing system based on subject, like most such systems are. I don't think we're planning on using our subject classification for anything other than the primary subject that we file by, though.

I do think that format and/or language notation should be at the lowest possible priority. Current libraries do a fine job of deciding when it's appropriate to group by format or language with very little guidance from the classification systems they employ.

-----

Actually the most interesting part of the last 20 or 30 comments has been the discussion of fiction/poetry/theater/music/etc. I think there is enough overlap and commonality between those subjects that they all belong in a major subject heading of something like "creative works." That way operas and symphonies and plays and poetry and literature all reside in the same basic section. That would also leave room for additions of new works like video games and performance art and anything else creative that we haven't yet imagined. We may want to quickly start dividing the section out into theater and poetry and music and whatever else, but I think creative works makes sense as a top level category and I think I would put criticism and analysis of creative works in a separate category.

I think that would work for users and it also makes sense to me intellectually. Are there strong objections to that?

Message edited by its author, Jul 17, 2008, 10:32pm.

>88

Sorry, I should've been more clear about that - in #3 by filing notation I meant something to indicate how two things with the exact same subject number will be put in order on the shelf (author's last name, series, cutter number, whatever we come up with). Books would be filed first by subject number, then by this other number/notation within those subject sections.

Message edited by its author, Jul 17, 2008, 11:11pm.

88: In principal I like the idea of grouping creative works together but I do have a few reservations. Primarily I would want to run it by some patrons first, as I'm not sure if they would like the idea of browsing fiction and music in the same basic area (although I imagine that any system we create would have a firm separation point in this case so a library could break them apart into different physical areas if they so wished).

And as always there are a few outlying item types to address. Would creative non-fiction works such as essays and memoirs fall into this category?

I agree that the call number should be modular with modules being of two types, say, "core" and "local". The core modules would be universal and unalterable. Local modules would include format, language, etc. The use of the local modules would be the decision of the local library and could be rearranged or dropped as needed. The official OSC call number would include both, perhaps the local modules contained in square brackets (I'll use { } since square ones won't display here).

The various formats of a work should have the same core call number.

For example, "The Remains of the Day" by Kazuo Ishiguro:

Book: {Book} {F} {Fic} {Eng} Ishiguro 1989-00 {XX-123-456}
DVD: {DVD} {F} {Fic} {Eng} Ishiguro 1989-00 {XX-123-456}

In my example:
1. The format is local: Book, Cass, CD, DVD, etc.
2. Fiction {F} (and Nonfiction {N}) is a local module.
3. Local types Fiction {Fic}, Mystery {Mys}, Science Fiction {SF}, etc.
4. The language English {Eng} uses the ISO 639-2 Code (http://www.loc.gov/standards/iso639-2).
5. The core here is the author's last name, year, and code: Ishiguro 1989-00.
6. The last local module {XX-123-456} is a classification based on the subject matter of the work.

Some correct local library applications of this OSC call number (for the book) would be:

Book F Fic Eng Ishiguro 1989-00 XX-123-456
F Ishiguro 1989-00
F Fic Ishiguro 1989-00 XX-123-456
F Book Ishiguro 1989-00
F Ishiguro 1989-00 Eng
Fic Ishiguro 1989-00

Consider using standards wherever possible. If included in the classification, US states should use the two-letter USPS codes. Countries could use the IANA codes for top-level domains (TLD) (http://www.iana.org/domains/root/db).

Message edited by its author, Jul 19, 2008, 3:49pm.

>91

I just have one question. Please clarify what the code is in the core designation. Name and year make sense to me. What kind of code do you envision?

I'm not sure about the core part other than author, I just stuck in an example someone listed above. My post was meant more to illustrate the call number structure. The classification parts, "1989-00" and "XX-123-456", are just sample placeholders to be replaced by whatever actual classification is decided on.

>91

"The last local module {XX-123-456} is a classification based on the subject matter of the work."

Shouldn't the subject classification be the universal core, not a local module? I think that what we're trying to develop here is a subject classification system for shelving. In which case you want all things with the same subject to have the same number across all libraries. I agree though that format and language should definitely be locally designated.

I'd also count some of what you designate as local (fiction vs. non-fiction, mystery, sci-fi, etc.) as part of the fixed subject classification, because a particular work is a mystery no matter what format or language it's in.

IANA country codes are not useful for geographic areas, only for political entities. The LC has a standard list for Geographic Area Codes (GACs) which are hierarchically structured (continent - country or area - subentity) and thus also lend themselves well for a classification:
http://www.loc.gov/marc/geoareas/

>94

For a nonfiction work, the subject classification (given as a made-up {XX-123-456} in my example) WOULD be a universal core module.

For fictional works, the core is based mainly on author-chronology ("Ishiguro 1989-00" in my example), with the local subject classification used to supplement the core and direct a reader to more books on the subject(s) contained in the work of fiction.

I should have given a nonfiction example as well:

Book: {Book} {N} {History} {Eng} XX-123-456 {2008-12} ...

where the OSC call number structure is given as:

{Format} {Fiction/Nonfiction} {Broad subject/BISAC?} {Language} Subject Classification {Chronology} ...

What I call a "local" module isn't alterable by a local library, rather the library chooses whether to INCLUDE that module in the call number. The "core" module is the smallest part which classifies and (uniquely?) identifies the work.

#95

The LC list you linked to has problems of its own. Guernsey, Jersey and Isle of Man have the same code. The only thing in common is that they are all Crown Dependencies in every other regard they are different.

An alternative would be ISO 3166-2. Here is a list - http://philmcrew.com/countrysubentity.tx... - which also goes down to subentity.

UN/LOCODE goes further down to towns but I feel that may be too far. Also as there are tens of thousands of them they are difficult to memorise.

Talking about ways of indicating standard patterns with a marker (Pivy's message #40 or #41), the Universal Decimal Classification, which is an adaptation of DDC does just that. Stabdard area numbers, when added to another to another subject number, are in parentheses. e.g., as in DDC, history of Japan is 952. But birds of Japan is 582(52) and -- i am not exactly sure of the exact number but Japanese comic books is something like 741.5(52).

I used to work in a special Earth Sciences library that used UDC.

Actually one can usually tell when a standard area subdivision is being used in DDC because it is usually prefaced with 09, e.g., birds of Japan 598.0952. (but of course it's stuff like that that makes the numbers longer -- there's always a trade-off)

The Bliss Classification, a fully faceted classification, sort of does Tim's idea. Subjects are generally all letters and the Bliss equivalent of standard subdivisions begin with numbers. General multi-subject encyclopedias are 3A, mammals in zoology is GP and encyclopedias of zoology is GP3A.

Message edited by its author, Aug 2, 2008, 12:12am.

PS - Re my comments on how UDC is derived from DDC, UDC started out with the DDC subjects as is they were at the time back in the early 20th cent., just making further sub-categories, and adding devices like the parentheses, but over time UDC has veered somewhat from that -- a noticeable example being the whole language class, class 4, was merged in with literature, class 8. That right there is another difference -- calling it class 8, not 800s --- but not to stray from the topic here, one can check out Wikipedia's articles on the UDC and on Bliss, and also the links from there.

Message edited by its author, Aug 2, 2008, 5:47am.

Surely this is one of those areas in which a Colon classification system could be quite useful especially when linked with some kind of Machine readable Metadata?

The structure of the Colon Classif. seemed very interesting and flexible from what i remember of studying it a couple of decades ago.

The notation, on the other hand, was the most godawful mix i have ever encountered -- numbers, upper and lower case letters, punctuation, and even a Greek letter or 2 thrown in! I pity the poor shelver who worked in a librry with that system.

That said, since notation serves the classification and not vice versa, there'd be nothing stopping giving it a much more user-friendly notation.

kiwimac, I've been trying to get a copy of the Colon Classification later than the one from the 1960's that I had. I know there was an update with changes (I think they got rid of the Greek letters among other things.) As a student of classification i'd be interested in perusing it.