Classification systems compared

ForumCutter, reloaded

Melde dich bei LibraryThing an, um Nachrichten zu schreiben.

Classification systems compared

Dieses Thema ruht momentan. Die letzte Nachricht liegt mehr als 90 Tage zurück. Du kannst es wieder aufgreifen, indem du eine neue Antwort schreibst.

1amandaellis
Dez. 14, 2006, 5:54 am

The original post for Cutter Reloaded compared Cutter to Dewey and LCC systems.

Whether a system is free or available seem to be key points, especially to Thingamabrians or personal librarians.

So what are other key points of difference?

2ciciha
Dez. 14, 2006, 8:58 am

Just to warn you at the get-go: one of my library school classmates works at a library which uses Cutter for classification. She says it is an absolute bear to work with! Before any decisions are made to add this feature to the LT site, maybe someone should do real-library research... Abby are you there?

3sabreader
Dez. 14, 2006, 10:05 am

I posted this over at the blog but will post it here too:

hmmm. just looking over the classification system as presented at Forbes Library, I think it would need some major tweaking if it is to be used as an international, not just US, cataloguing system. For example: categories B, C, and D (where non-Christian and -Jewish religions are lumped together with "occult", and "Church history" -- meaning Christianity -- gets its own entire category -- obviously an artefact of the time/place it was developed); and H and J, where for some reason political science gets its own entire category while "social sciences" is lumped together with business, investment, etc.

In principle, great idea, but this particular system needs some serious tweaking.

4timspalding
Dez. 14, 2006, 10:06 am

Which one? (There are only four!)

I certainly don't suggest that LT drop other classifications. This would be LT helping to develop a system which, while LT would use, but would not be taking over the site.

It would be interesting to hear from the Bliss guys, another small system, but one that has kept up.

The difference between the history of Bliss and Cutter is brought out in the Conrad Winke article (link).

"However, EC still might have been salvageable in the immediate years after Cutter’s passing had the librarians using the scheme at the time banded together and worked cooperatively at maintaining the schedules, as happened with Henry Evelyn Bliss’ Bibliographic Classification, now maintained by the Bliss Classification Association and still in use in a number libraries of libraries in Great Britain. Instead, librarians at EC libraries seemingly did not pursue working together, but worked on their own until, in all but four cases, this became impractical and they abandoned it."

The reasons for dropping EC relate to this:

"Librarians at eighteen institutions, either via telephone interviews or from Mowery’s correspondence, were able to provide their institution’s reasons for abandoning EC. These were broken down into three general categories. By far, the most frequently cited reason was the lack of coordinated revision of the schedules. None of the libraries surveyed, be they past or present users of EC, had ever worked in conjunction with another library on schedule revisions or sharing call numbers. Therefore, all updates to the scheme had to be carried out in house, a time-consuming and costly undertaking."

Unfortunately, the libraries' systems has kept changing, without being in contact--finches in the Galapagos, if you will:

"However, even if EC numbers did occur... (in OCLC) ... with some regularity, because the EC libraries did their own revisions in house and not in conjunction with each other, EC numbers in MARC records would not be able to be shared between libraries with the same facility that LCC and DDC numbers are."

5MrKris
Dez. 14, 2006, 10:17 am

Nachricht entfernt.

6timspalding
Bearbeitet: Dez. 14, 2006, 10:52 am

I think major tweaks would be problemmatic—it would be better to make a new system.*

As much as I use the similarly myopic Dewey religion categories as an example of the inflexibility of traditional systems, I think the case needs to be made as to why these things "matter." The primary purpose of the system is to establish an order of books on the shelf. So long as the order is regular, do the levels matter?

Take these examples:

A. Christianity
B. Judaism
C. Islam
D. Sikhism

A. Judeo-Christian
A1. Christianity
A2. Judaism
B. Islam and its spin-offs
B1. Islam
B2. Sihkism

A. Christianity
B. Other religions
B1. Judaism
C2. Really obscure religions
C2a. Islam
C2b. Religions they didn't teach me about at Milton
C2b1. Sikhism

Each system has a sort of understanding of the world behind it. "Fair minded but ordered from a Christian perspective, "attempting to systematize," and "a parody 19c world view." But does that matter if the ORDER is the same? Millions of people use Dewey every day, with its stingy allocation of space to Judaism. I've seen no Dewey-inspired pogroms.

Now, the upshot of this is call-number-length. Judaism is one thing, but Buddhism in DDC has some REALLY long call numbers.

Anything else?

Deciding whether to make a whole new system would be a good upshot of this conversation. In general, my idea here, however, is to build on some data. Take all the Cutters and put them together FOR THE FIRST TIME, and I think you'll have a body of data that can automatically suggest some good classifications. Start with a totally new system and you start from nothing all around.

Ultimately, I don't think you can EVER devise a really good classification system. Putting Christianity and Sikhism on the same level now seems very correct. Five hundred years after a comet hits the Punjab (or Christianity crumbles, etc.), it might not seem so.

If you designed from scratch, I'd argue the system shouldn't be excessively concerned with being logical. A good test of a system would be to take 1,000 books, get their "similars" on LibraryThing and Amazon and compute how far apart the books would be. Of course, that wouldn't be static over time either.

7timspalding
Bearbeitet: Dez. 14, 2006, 10:26 am

MrKris: The argument goes as folows. Dewey is owned. A library can't use it without paying. And LibraryThing can't tell you what your call numbers "mean" without getting OCLC's permission.

8PossMan
Dez. 14, 2006, 10:55 am

One disadvantage of Dewey is the distinction made between pure and applied science. I suspect the main reason Dewey is so popular is a bit like Microsoft Windows - everyone else (nearly) has it. Public lending libraries here are more or less locked into it and change is not really an option for them. If there are people out there with the time and the skills to start from scratch in a collaborative way that would be a very beneficial development (a bit like some projects in biological taxonomy).
And as an aside the novel The Grand Complication through its hero sings praises to the beauties and delights of Dewey Decimal System

9smellthecoffee
Bearbeitet: Dez. 14, 2006, 2:44 pm

I think Cutter is not a practical way to go because it is so out of date and has no communally recognised updating process. If a new system is to be considered i would vote for Bliss 2nd edition which is to my mind the best subject classification around today.

Bliss 2nd ed., as well as Ranganathan's Colon Classification, are the only fully-faceted general classification schemes, which means it can also serve as a thesaurus for alphabetical subject indexing (UK spelling but how hard would it be to come up with an algorithm to automatically choose which spelling you want).

In fact the terms and concepts of class H Medical and Health Science are mostly the MeSH Thesaurus (with UK spelling) A fully faceted classification optimally is designed to be used as both a classified and alphabetical subject retrieval system all rolled into one.

The main problem with Bliss is each of the volumes is expensive. The lack of a one-volume abridged version of the whole classification is also a serious defect. The reason this is needed is because say a library has certain subject areas in depth, but also has a more general collection not requiring in-depth indexing.

For example for my library i what i would want is the full edition of Class P Religion Occult Ethics and class W The Arts, and a brief 1-vol for everything else.

The volume of general auxiliaries (which is like Dewey common subdivisions) does give a brief outline of the whole classification. However it is too brief and incomplete to actually be used for actual classifying.

Their volumes are so expensive probably because its at the moment relatively few users are having to bear the costs among themselves, not because the Bliss Classification Association is trying to make a huge profit. Putting the classification on the web would be a boon.

There are also ways in which they could perhaps simplify the notation a bit (although it's no worse than LC). And personally i think having a one-volume version of the classification is vital.

I am not a member but i plan to join, especially now that i see from my fellow Library Thingites that the desire to improve subject classification is not as dead as I thought! I don't know what the members of the Association might think but perhaps this could be the direction to a successful growth for the classification.

I have not said much about the Colon Classification because i have seen it alot less. It has been used quite a bit in India, at least in the past, but I don't know the current status nor if it has a common system for maintenance. But Ranganathan is the creator of the concept of faceted classification, and for all I know at the present, it may be even better than Bliss. I certainly would like to hear from any LT folks who know about the Colon Classification and especially its current status.

One thing i think would have to be simplified in the Colon Classification is the notation which makes great use of punctuation marks. Shelvers must have a hellacious time shelving with the system. But notation can be changed without changing the conceptual structure of the system. I think that perhaps some steps may have been taken to make the notation more user friendly but i am not sure.

10MMcM
Dez. 15, 2006, 11:24 am

>6 timspalding: (a bit OT)

Perhaps it's not central to your point, but Sikhism is either a spin-off of both Islam and Hinduism or of neither. Guru Nanak was born a Hindu. (Rev. Lovejoy: "Miscellaneous." Apu: "There are 700 million of us.") The Sikh holy book says ਨਾ ਹਮ ਹਿੰਦੂ ਨ ਮੁਸਲਮਾਨ ॥ I am not a Hindu, nor am I a Muslim.

If the keys are arbitrary, then why not make them totally arbitrary? If they attempt to encode some meaning, then aren't there bound to be problem cases (beyond just length) like this?

11bkalish
Dez. 15, 2007, 7:34 pm

Hey folks,

I'm delighted to have found this group. I work at a library that still uses the Expansive Classification as our primary classification. I love the system and think many of its devices and methods are fantastic. Integrity of symbols! Brevity of marks! The local list! Form and aspect subdivisions. It's genius.

Unfortunately there are some problems with implementing the system. What it comes down to is that it isn't so much a shelf classification scheme as a collection of suggestions and guidelines. Yes, of course there are published schedules, but those schedules are full of options and alternatives and varying, often contradictory practices. The idea, and a lovely one it is too, was that each library would select their own preferred methods and usages so as to classify their books so as to best meet the needs of their patrons. However, because each library implemented the system differently there was never any consensus about class numbers, or even the order in which the marks were to be arranged. It's not just that different libraries maintained the schedules differently; they implemented them differently from the very beginning.

It would certainly be possible for a group of LibraryThing users to agree on a single set of practices for implementing the system, but it would be very, very difficult. There is no instruction manual for implementing the system and simply becoming familiar with the options would be a formidable task. Deciding on a uniform set of practices and updating the schedules would nearly amount to the creation of a new classification system.

All that said, if folks still want to go ahead with this and it looks like such a project might prove of real benefit to some users, than I am all for it and would be happy to contribute my time and knowledge to the project.

Benjamin Kalish

12timspalding
Bearbeitet: Dez. 15, 2007, 8:48 pm

Hey, what library—Forbes?

13bkalish
Dez. 16, 2007, 10:38 am

Given that there are only four such libraries I don't suppose there is much of a chance of me keeping it a secret. So yes, I work at Forbes.

14timspalding
Dez. 16, 2007, 1:47 pm

Well, I'd love to chat some time. As you can perhaps tell, I spent a lot of time on this issue. I found the Forbes not particularly interested in talking to us about it, or doing anything. I'd love to bring Forbes' classification into LibraryThing. I'm sure members would help type up the top levels of the schedules which, I gather, Forbes has not digitized. There's no reason LibraryThing couldn't make suggestions for CEC's when you don't have the book yet, or have it cataloged.

15bkalish
Dez. 16, 2007, 2:49 pm

There is, or was, a digital version that combined the seventh classification with the cumulative annotations and modifications of Forbes catalogers over the years. I don't know whether a copy remains in digital form---I have only worked with the printouts.

Working from the Forbes annotations would have some obvious advantages---the library has already done much of the work of keeping the classification current---however there would be disadvantages as well. First off, Forbes would own the rights to any recent modifications, so we would require their permission. Also, the library has adopted many usages that may be less than ideal but are too entrenched to be abandoned.

The most important questions we must ask is who do we expect to use this classification and how much effort should implementing it require. Would it be used by individuals? By small public libraries? By only English speaking libraries? By libraries of any size? By libraries world wide? (CEC would be a bad choice for, say, Chinese libraries.) And will people want a unique mark assigned to each book? A small selection of marks at different levels of expansion? A range of marks to be selected according to local needs?

16timspalding
Dez. 16, 2007, 3:01 pm

The idea that a public library that is the only library in the world still actively using a 19th century classification would raise the issue of license rights over modifications... well, I give up. Cutter was a good idea, but not sharing is what killed it. If, with one library left, the issue is even raised, it doesn't deserve to live.

This kind of thing drives me up the wall. Red taep and not sharing is why there are hundreds of thousands of Amazon developers, and all the book data on the web is Amazon data, and LibraryThing is the only serious site using library data.

Apologies for the sour tone.

17bkalish
Dez. 16, 2007, 3:54 pm

As far as I know the library never has raised the issue. *I* raised it, and I honestly have no idea how the library (which in this case probably really means the board of trustees) would respond if asked. As an employee of the library I have an obligation to ask the question. It seems unlikely that it would be a problem, but I would hate to put my job in jeopardy by neglecting to ask for permission.

Also, I really don't believe that not sharing was what killed the Expansive Classification. Cutter was a huge proponent of cooperative cataloging and encouraged sharing between libraries. The problem was that his system, as published, required too much work to be implemented and so relatively few libraries signed on. If the system had been more attractive to catalogers and had been better supported in its early years it might not be so obscure today. What was lacking was not the willingness to share but the labor and funds to make it happen.

18bkalish
Dez. 16, 2007, 4:04 pm

By the way, I don't mean to dismiss what Winke's has said about cooperation among libraries. I just find it frustrating when it is all blamed on "Cutter's failure to provide for the continuing revision, expansion, and publication of his work" without asking *why* Cutter failed at these tasks.