| Bush, V. |
[Aug. 2nd, 2006|10:13 am] |
1) Library of Congress Subject Headings - possible subject headings -
Inventions - T15 - T35 (history); Technological innovations - T173.8 (technology); Science - Q - "most appropriate"
2) Dewey Decimal Catalog - 000 - Generalities, 004 - Data Processing, Computer Science
Library of Congress Classification - QA 75.5 - 76.9 - Computer Science, Mathematics, Calculating Machines
3) a) Keywords from text - invention, innovation, memex, calculating machines, associative indexing, advanced data analysismicophotography
b) LISA - computer science, information science, history, innovations, predictions, technology
In reading and working with this article, I got curious about the author, Vanneva Bush. I 'googled' him and learned that he designed 'the world's most powerful analog computer' and lived from 1890 - 1974. Knowing about his work helps put this article in a better context. No wonder he was able to predict so clearly the advances that were to come after WWII. This website gives further information, including pictures.
|
|
|
| Batty, D. |
[Aug. 1st, 2006|10:59 am] |
Batty, D. (1989). Thesaurus construction and maintenance: a survival kit. Database 12 (1), 13-20.
The article opens with a good summary of a thesaurus, then outlines what is needed in the basic construction of one. He provides very useful questions about how to discover who the users are and what they will need and comments that a mix of free text serch and controlled index vocabulary is currently considered a wise choice in thesaurus construction. He then gives details on what he sees as the six steps of this process. These begin with analyzing the range and depth of the subject area covered by the thesaurus. The are three basic areas, core, fringe and outside - rather like a bull's eye target, each one covering a larger area and requiring fewer thesaurus terms. The next step is finding the sources for the raw vocabulary. For a single source literature, samples of that literature are best. A more general source should have a variety of samples - immediate literature, general literature, user questions among them. Collecting the recording the raw vocabulary was explained in amazingly low-tech terms - each term on a card or slip of paper! This is so like the manner of quotation collection for the Oxford English Dictionary that I had to smile. Batty does say it is possible to use microcomputers for the same purpose, but it is clear - to me anyway - that he believes this is not the best way. About half way through this explanation, I realized he was describing a faceted approach to a thesaurus. He does not use the term 'facet', but he is describing a method of facet organization.
Organizing the raw vocabulary consists of examining the terms for sets with a basic semantic affinity and making piles of cards for various categories - Things, Problems, Places, etc. This is the most time consuming part of thesaurus construction, taking up to more than half the time of a project. It is important to remember that the clusters should be mutually exclusive, with hierarchies within them that are simple, not complex. Notation should be easy to say, write and remember - a quote from W.C.B. Sayers. I must say, LCC does not meet this goal by a long chalk! Notation should also be able to accommodate as many terms on one level as possible and hospitable to new terms. There is a discussion of the cluster order and use of compound terms and precoordinated terms, followed by four steps in putting the raw vocabulary into refined form. One of the last steps is to create the references. An interesting point was that a relationship such as Bees RT Honey cannot be derived automatically from the hierarchies and must be entered by the human lexicographer. The final section deals with thesaurus maintenance.
I appreciated Batty's comments about the need for a human intelligence at the heart of this process. Not so much because it insures work for people, although it does, but because there are some who believe artificial intelligence will replace the human in time. Anything is possible, of course, given enough time, but I believe this is very unlikely. As amazing as computers are and as much potential as they have, they are only as good as the people who create them. But I digress! This article is very useful, giving good details and practical questions for the thesaurus construction person to work with. It also makes clear why this type of work is paid well. |
|
|
| Berinstein, P. |
[Jul. 31st, 2006|07:44 pm] |
Berinstein, P. (1999). The big picture - Do you see what I see? Image indexing principles for the rest of us. Online 23(2), 85-86,88.
Berinstein opens with the challenge to describe 'what is in figure 1', which looks to me like a paintball gun. It is clever way to get the reader's attention and make them immediately aware of the difficulties inherent in indexing for images. She even invites the reader to send her their description, giving an e-mail address. I do wonder if it is still open for messages. This challenge makes her point well, pictures do not describe themselves, unlike text, and are difficult for different reasons than text might be. Because pictures are both of something and about something, they really have two challenges to an indexer. The first has to do with the power a viewer has over what the picture contains, the second with the levels of meaning in the pictures. (I am using the word 'picture' here, but 'image' is probably the more exact term.) Berinstein describes the approach to image indexing outlined by two people, Panofsky and Shatford.Shatford's idea seems more exact, indexing the image first into the facets who, what, when and where, then each into of and about, with of being separated again into generic and specific. As useful as this may be, it is still far too open to interpretation. Berinstein closes with three ideas to help with indexing choices. These seem most useful - consider the audience, install help files and let the user browse.
This article makes clear why I am not always successful when I search for images on the Internet. The concept of an image having levels of meaning is so clear once you read it, but I had never thought of it in those words before. I have wondered at the captions of pictures in newspapers and magazines before, now I know why. Even a news photo cannot be objective and often is not meant to be. My idea of the photographer's meaning is likely to be very differnet from the caption maker, let alone the photographer. Text indexing may be difficult, image indexing is in a category all it's own!
|
|
|
| Aitchison, J. et.al., #6 |
[Jul. 31st, 2006|02:20 pm] |
Aitchison, J., Gilchrist, A., Bawden, D. (1997). Section B. Planning and design of thesauri (pp. 5-12). Section C. Standards for thesaurus construction and development (p. 13). In Thesaurus construction and use: a practical manual, 3rd ed. London: Aslib.
The most useful part of this section, for me, is Figure 1., the side by side comparison of natural and controlled language. I disagree with the idea that a weakness of natural language is the 'intellectual effort placed' on the searcher. If a system has been well chosen to fit the needs of its users, most people will have the ability to find what they need. There are always the few who have trouble, but you can only plan for the general user, it seems to me. There is no question that a term having too many synonyms can be a weakness, however. False drops can also be a problem here, although some readings expressed the opinion that this problem is about the same in either language. I prefer natural language searches, but I do see the usefulness of controlled language, especially with certain disciplines - computer technology, natural sciences, mathematics come to mind.
As for the rest of this section, it is again difficult to pick specifics because it is all so new for me. I believe this section would prove very useful to those making a decision on planning and designing a thesaurus. The detailed discussion of each aspect of a database would help in evaluating the database under consideration. Breaking down the individual characteristics of a thesaurus makes those elements easier to evaluate as well. I do agree that cost is a key factor in this decision. The cost per hour could be prohibitive if the thesaurus was too detailed. Picking certain characteristics and not using others would help in this area. |
|
|
| Aitchison, J. et.al., #3 |
[Jul. 31st, 2006|12:41 pm] |
Aitchison, J., Gilcrist, A., Bawden, D. (1997). Section F. Structure and relationships (pp. 47-52). In Thesaurus construction and use: a practical manual, 3rd. ed. London: Aslib.
This section introduces structure and relationships. Specifically, F1 deals with the classification techniques used to determine the relationship between the the structural relationship and the terms it contains. Because this is all new information to me and I find this work less approachable than Hunter, I find most of it informative, which I assume means it would be useful. One detail I did learn is what specific terms are used as entry vocabulary. Words that would come under the UF category are those when are the non preferred terms - synonyms, lexical variants and quasi-synonyms. The equivalence relationship also has upward posting , narrower to broader terms, which Cutter did not have - that is an interesting bit of information as well. |
|
|
| Hunter, #3 |
[Jul. 31st, 2006|11:44 am] |
Hunter, E.J. (2002). Classification made simple, 2nd ed. (pp. 89-136). Aldershot: Gower.
These chapters deal with the compilation of a thesaurus and classification as a search tool in a variety of areas, including the Internet. As always, I find Hunter very readable. I especially like the opportunities he gives the reader to practice a concept he has just described. There was one after he introduced the faceted classification scheme for use in a thesaurus. This is useful because it gives the reader a chance to try the scheme and discover which areas might need to be looked at again for understanding. I also find useful his discussions about faceted classification in general. For some reason, his approach works better for me than that of Aitchison. In the section on classification as a search tool, I found useful his explanation of a Boolean search. While I was told about this in 401, the explanation then was not very helpful, in part I think because the instructor felt rushed. Hunter's helped a great deal. I will be able to use this type of search more efficiently in the future.
I did not think a notation search would be as useful as he seemed to. Or rather, I think it has a limited use. Professional library persons would find this useful, I can see that. But the general public would not, I believe. In the section on use of classification on the Internet, I found a quote I disagree with as well. Hunter quotes Neely as saying that a strength of Yahoo is its categorizing of content. I did not find this to be the case in my Yahoo search assignment. For that matter, I do not find any search engine's classification system wonderful. I think this will improve with time, but they do seem to be in a learning process at the moment. I was surprised to read that Google's search engine was considered recent in 2002. I was not aware they were that new.
The conclusion section is interesting, but rather repetitive - as I suppose they are meant to be. I was glad to see him say that "no one classification scheme will be suitable for all purposes" and that use determines type. |
|
|
| Foskett, A. C. |
[Jul. 31st, 2006|11:05 am] |
Foskett, A.C. (1996). Chapter 8: Alphabetical subject headings: Cutter to Austin (pp. 123-134 only); Chapter 23: Library of Congress Subject Headings (pp. 336-347). The subject approach to information, 5th ed. London: Library Association Publishing.
In the first part of this reading, Foskett outlines the major contributions of Cutter, Metcalfe, Kaisa and Coates to subject headings. (I'm not certain where Austin comes in.) Cutter, in 1876, was the first to attempt an established set of rules for alphabetical subject headings. He had a specific approach idea using natural language. When ever I see this term now, I think of the earlier article we read by Wellisch who pointed out that all language is controlled, whether by an organized, written system or just the circumstances the speaker is in at the moment. I think that idea is very important here, because Cutter is using this term, natural language, to mean what is 'generally accepted' by - I am assuming - the community of users for a specific discipline. Foskett goes into some detail in discussion Cutter's organizational idea. I found interesting the ideas that double entries were used to retain 'names in common usage', and terms were put in subject or place depending on what area or type of information was being dealt with. Again, according to whom? Links were to go downward - Broader to Narrower, not upward. Cutter's system may have had problems, but he did make a real contribution by introducing a direct entry system in natural language, backed up by syndetic network of cross-references.
Metcalfe, a supporter of Cutter, introduced the alphabetico-specific catalog with known names in known order. While this system is intended to be easier for the reader, this has not always been the case. His work was improved by the work of Coates, who approached the Cutter/Metcalfe systems by looking at the psychology behind Kaiser's selection of concrete rather than process as the entry point. He pointed out that humans are not able to visualize an action out of context. We have to think of the thing involved in the action. Working from this idea, Coates started with Thing-Action and organized his system from there. It is an interesting approach. I agree with his idea of how we visualize actions. I am not certain most people would necessarily come up with the same order as Coates does in the further expansion of Thing-Action.
The chapter on LCSH was interesting because it provided information on the history of this system. It is no wonder there is no rule to show how a topic is divided. I can imagine that people who work with this system in cataloging learn the different headings by using them over and over. I gave up on this long ago and just rely on a search to provide me with a place to go!! Historic context, which both these chapters provide, is useful for because it shows the growth of a system, scheme, process. Understanding that development makes it easier to deal with the flaws inherent in any approach to classification. I doubt the world will last long enough for any human or computer to discover the perfect system. We are surrounded by a world full of exceptions to any rule - a classification 'birds' that includes, at some level, the Platypus and ostrich is a good example. Perfect would probably make humans very nervous! |
|
|
| journal assignment - Yahoo! |
[Jul. 28th, 2006|04:00 pm] |
I looked at the Yahoo! Directory organization and discovered that a general search for the word 'classification' brought up several pages of website listings such as the Wickipedia definition of the word, several websites which deal with taxonomy and plant names and a number of classification software sites. A search in the Yahoo! directory brought up more specific listings such as the British Board of Film classification and an outline of the LC Classification Outline. At the top of this page were bullets for a number of more formal looking classification sites starting with NAICS. At the bottom is the word More... Clicking on this brings you to a page with these headings at the top, under the information "Top 10 Related Directory Categories. These include a Library listing, one for Bliss Bibliographic, DDC, UDC and LCC. It also has two for the U.K. There are 11 listings in all, the 11th one being rather amusing. It is Practical Jokes and Pranks > Gnome Liberation. Clicking on this brings you to a page of sites about Garden Gnomes and the Liberation movement that, I am sad to say, is world wide!
I do not believe Yahoo! makes finding this information easy. It takes three clicks to find the page with the top 10, something that seems to me should come up on the first click. I do not think browsing is easy on Yahoo! and I find their home page confusing. I know there is a logic at work somewhere, I just don't see it. Just for interest, I did a similar search for 'classification' in the Google Directory. I find their organization more logical and easier to do in general terms. But, finding the same information was not as easy. Nowhere I looked were they together as in Yahoo! and they were widely scattered on the pages that came up. The same applied for MSN Search. |
|
|
| Dykstra, M. |
[Jul. 26th, 2006|11:42 am] |
Dykstra, M. (1988). Can subject headings be saved? Library Journal, 113,55-58.
This article opens with a statement showing a clear bias, and by the time I finished reading it, I was thinking it might have been better as an op-ed piece because the bias is so strong. Her topic being the attempt by the Library of Congress to make a thesaurus out of the LCSH, she quotes from their "See also References" in which they say 'headings' are 'equivalent' to 'terms', a statement Dykstra calls "quite simply false." She continues with phrases such as "serious professional problem" and "some sort of disaster" as she discusses how libraries across North America will try to follow the LCSH usage. According to her, the LC Subject Cataloging Division has been selective in what they chose to use or not use (her word is 'ignore') from the principles of thesaurus construction. The premise of thesauri everywhere is that of terms, definitions of which are the 'building blocks' of a thesaurus. Through discussion of this problem with her students, she believes a solution has been discovered. Dykstra calls it the Corbett solution after the student who first presented the basic idea. The rest of the article is given to outlining the five steps needed to solve LC's problem of having a list of subject headings and wanting them to be a thesaurus.
The solution seems feasible, and the advantages are presented in a clear manner. She sees them as, first, a proper, rule-based thesaurus - to use her earlier metaphor, a game of tennis improved by the rules of tennis, second, the relevance and relative ease of use for all on-line searchers. She also says it would help facilitate the advancement of the automated analysis of text. Her closing paragraphs seems to make clear this is her real bias. Her frustration seems to be that as LCSH stands it will hinder this technology and a change to proper thesaurus structure will improve this advancement. She might have done better to start out with this positive note, instead of sounding angry and bitter at the beginning. Whatever her personal feelings, the tone of the opening paragraphs is a bit off putting, which rather spoils that fact that she may be correct . |
|
|
| Aitchison, J. et.al., #5 |
[Jul. 25th, 2006|03:36 pm] |
Aitchinson, J., Gilchrist, A., Bawden, D. (1997). Section E. Specificity and compound terms (pp. 35-46). In Thesaurus construction and use: a practical manual, 3rd ed. London: Aslib.
This discussion of vocabulary specificity explained the uses of different types of controlled vocabulary. How specific the retrieval language is depends on how well the indexing terms can express the subject in depth and detail. Specificity is important to the performance of the language because, among other things, it determines the accuracy of the concept terms. A good entry vocabulary is not affected by the specificty of language, precision is. The disadvantage of a highly specific vocabulary, for compound terms as well as single terms, is the number of terms increases, which means more expense to compile, maintain and operate. It also means the compilers need to have a greater knowledge of the terms and the problems of selection, organization, etc. are greater. Changes will be needed more often as well.
Compound terms are difficult to construct, in part because they need to be precise in order to be useful. They are made of adjectival or prepositional phrases and contain a focus ( the noun component) and a difference (the modifier). It is best to use the natural form or direct form in stating the compound term. The advantages of compound terms include ensuring good precision and preventing false drops. The disadvantages are listed above. Factored terms have the advantage of giving better recall than compound terms, but they also give poorer precision performance. The final section is given to a discussion of the rules for when compound or factored terms are retained and the rules or recommendations for creating factored terms. |
|
|
| Hunter, E. J. #1 |
[Jul. 24th, 2006|10:43 am] |
Hunter, E.J. (2002). Classification made simple, 2nd ed. (pp. 4-39, 59-69, 82-85). Aldershot: Ashgate.
Classification techniques are used in information and retrieval systems in order to provide access to that system. All information systems need to deal with simple and complex subject headings, and this is done in either a hierarchical - top-down - or faceted - bottom-up - manner. Because a faceted classification scheme is more easily understood by humans and computers, it is becoming a common classification scheme. Hunter then gives a detailed discussion of the process involved in this classification scheme. I found his step by step analysis very useful. I was finally able to understand the different terms clearly - facet, concept, notation, etc. This reading was much more useful than the Aitchison one in explaining faceted classification . He follows his analysis with an example from a brewery - a clever choice because this would peak the interest of many readers. He then gives the reader an opportunity to try their hand at the process with an example from an estate agent's listings. I did try this and did pretty well at the basic divisions. As he showed further variations, it became clear a faceted scheme does allow for growth and adaptation. Hunter made clear why the user of a scheme is an important part of the process, something Aitchison did not do. He concluded this section with a list of the eight steps involved in a faceted classification scheme.
Hunter then goes on to show how variations of this system can be incorporated into existing enumerative schemes. This is useful because it can help reduce the length of the schedules and improve efficiency. The two main ways to do this are by providing auxillary tables of terms to use throughout the schedules or include instructions within schedules for dividing one classification number in the same manner as another. There followed examples from Bliss, Dewey and UDC. UDC has grafted synthesis onto its enumerative scheme for some time. Section 9 listed the advantages and disadvantages of each scheme - faceted and enumerative. I can see why the faceted scheme is useful in computer systems and why the DDC is still being used in public libraries. Each one fits their user well - the faceted scheme is excellent for computer systems because it does not list complex subjects, which means the schedules are shorter although complete; the enumerative scheme has been in use far longer, and it is easier to display the notational structure. They seem to balance out, finally, which helps explain why the enumerative scheme is still used, even though the faceted scheme is very popular. |
|
|
| Aitchison, J. et.al., #4 |
[Jul. 23rd, 2006|08:40 pm] |
Aitchison, J., Gilchrist, A., Bawden, D. (1997). Section F2. Classification and facet analysis (pp. 66-75). In Thesaurus construction and use: a practical manual, 3rd ed. London: Aslib.
My reading of the paper on the music faceted access project helped with this section of Aitchison, and I believe I understand the basic concept of facet classification. It is difficult to summarize these sections from the thesaurus book, because the book's purpose is to help in 'building' a thesaurus. This section opens with a discussion of classification, an important tool in finding structure and relationship when constructing a thesaurus. The most common form of systematic classification in this area is faceted classification. This system is 'synthetic' because if provides for combinations of terms to represent detailed topics. It is also analytical because it is structured so classmarks represent simple concepts organized in specific categories. Faceted classification is useful in thesaurus construction for three main reasons. It 1) provides a tool for analysis of subject fields and determining relationships between concepts; 2) may be used as a systematic display in a published thesaurus; and 3) facets may be added to terms in an existing vocabulary. One distinctive feature of fundamental facets is division of terms, using one division characteristic at a time, into categories (facets) of homogeneous, exclusive groups. Another is the recognition of a limited number of fundamental facets.
The discussion of subject field and discipline structured, using the 'clothing technology' schedule, was quite understandable. I got lost in the section on 'fundamental facets as main division' thesaurus types. I can see how faceted classification is useful, perhaps better than enumerative classification. When the relationships between terms are explained with a classification system that uses the language of the subject area or discipline, not only is the indexer able to use the terms, the user is as well. It is logical that an index is only as good as it's usefulness to those who need it. |
|
|
| Music Library Association |
[Jul. 23rd, 2006|12:25 pm] |
Music Library Association. Working Group on Faceted Access to Music. (1994). Discussion paper: Faceted access to music" possibilities and ramifications. Avaliable at: http://www.music.indiana.edu/tech_s/mla/wgfam.dis
This discussion paper details the problems with subject access to music, explains why a thesaurus is needed and presents their faceted indexing proposal. The explaination of a faceted access system was very useful, clear and understandable to me, a novice in this area. A faceted access system is works with concepts in a specific discipline which are identified and organized into unique categories, called facets. The concept is then denoted by term - a work or expression with a precise mening or peculiar to a science, art, profession or subject. These term are listed in a thesaurus which is organized in a hierarchical manner, shoing the relationships between terms. The conceptual context created is a great advantabe of a faceted system because is allows searchers and indexers to understand the use of the term in a specific discipline. Since subject access to music requires both non-topical access points and multi-element access points (keywords search), on-line databases are often not useful for music-specific searches because the facets of music information are often not in the subject headings. A thesaurus is need which will provide the list of terms used, with classification by ideas represented. The thesaurus, wihch has created concept indexing constructed with facet analysis, will identify concepts, give them terms and determine the relationship amoung and between the terms. This thesaurus will make those relationships clear to the users. This, in turn, will provide success in information retrieval which depends on a strong collocation, enabling the user to see the work as a member of a class of related works.
The rest of the paper explains the work that has been done on the thesaurus to date, with specific examples taken from work on a database of LC music subject headings. It is clear there are specific advantages to facet-access, especially in very specific subject area searches. Classification systems such as Library of Congress, which use the alphabetical approach, do not use the language and termonology of a subject discipline. The headings they create often will not be of use to the type of search required by an architect, musician or scientist. Facet-access seems a useful alternative now that on-line searches are the norm in many fields. |
|
|
| Aitchison, J. et.al., #2 |
[Jul. 20th, 2006|03:31 pm] |
Aitchinson, J., Gilchrist, A., Bawden, D. (1997). Section F1.2 The hierarchical relationships (pp. 52-58), Section F3. Notation (pp. 75-79). In Thesaurus construction and use: a practical manual, 3rd ed. London: Aslib.
Again, this section of the thesaurus is interesting and informative, but difficult to summarize. Mostly it is the setting out of terms, their explanations and examples of what they are. In the hierarchaical relationship there are levels of superordinate - 'a class or whole' and the subordinate - 'its members or parts.' Built on a progressive sequence, it is a basic feature of a thesaurus and very important to proper recall and performance. There follows a discussion of the different aspects of these levels, with examples of how to place items in the correct level - what is a class and what is its members.
Notations are sets of symbols that are added to the classification system to represent various concepts. They are usually Arabic numerals and the Roman alphabet, but there are other symbols used such as hyphens and colons. There are also expressive notations, so called because they are used to explain subclasses and use words instead of symbols. The discussion pointed out that notations can become overwhelming, somewhat destroying their usefulness. I did not find as much in this section to apply to personal experience. |
|
|
| Hunter, E. J. #2 |
[Jul. 20th, 2006|03:07 pm] |
Hunter, E. J. (2002). Classification made simple, 2nd ed. (pp. 40-58, 70-81, 86- 88). Aldershot: Ashgate.
This group of selections begins with an explaination of 'hierarchy' which is very useful. The graphs help make the differences between types clear. Hunter goes on to explain the different terms in a hierarchial scheme, such as 'order in array', 'citation order' and 'notation.' Using examples from DDC and LCC, among others, he then details some of the practical examples of this type of scheme. Looking at the two with which I am most familiar, DDC and LCC, I was able to see more clearly just what their flaws and strong points might be. DDC does allow for expansion in the addition of additional numbers, but the classifications created are often limiting. The example Hunter gives about types of cooking makes that clear. The LCC can also grow, perhaps even more easily, because it makes use of numbers as well as letters. But I have noticed that the amount of space needed on a book spine to place the LCC number is getting longer and longer. There is only so much room!
The section on notation I found most interesting because Hunter used a number of different classification schemes by way of illustration. The history side of me was intreged by the NATO Codification System. It was both interesting and chilling to see terms such as 'weapons', 'guided missiles', 'chemicals and chemical products' and 'live animals' listed together. I do wonder what the live animals have to do with NATO! These readings, again, are interesting and very useful to a understanding of the process. I admire library catalogers even more now, as I do those programmers who do indexing for information retrieval systems. Card catalogs seem such an archice system now! |
|
|
| Olson, H. |
[Jul. 19th, 2006|07:58 pm] |
Olson, H. (1994). Universal models: a history of the organization of knowledge. In H. Albrechtsen and S. Oernager (Eds.), Knowledge organization and quality management: Advances in knowledge organization, vol. 4 (p. 72-80). Frankfurt/Main: Indeks Verlag.
Olson looks at theories on organizing knowledge by various authors starting with Cutter and Dewey. Her analysis determines that organizing knowledge is power and these authors show an assumption that universal models are necessary. This assumption grows from a mistrust of the differences in lanuage and a need for models to overcome those differences. She points out how Cutter and Dewey presented arguments in favour of universality, followed by other theories which lead her to see univerality as a dominant concept, so entrenched in the literature it has entered the efforts of feminists to create subject access to women-centered materials. Olson then shows the limits of universality by citing examples from DDC and library catalogs on the Internet. She argues for opening up the universal model, using the metaphor of a house which can have rooms added on as 'new family members' join.
Olson's discussion of the DDC classification of 'women' over time from 1913 to 1989 is very interesting. There is no question that the earlier placement shows a bias. I am not certain I agree with her conclusion that the most recent placement, of 'men and women' as sending a negative message to classifiers, if not the public. It may do so, I just don't believe her argument proves that to be the case. In an alphabetical structure, which Dewey used, the word 'men' comes before the word 'women.' I'm not certain it is useful to take that as proof of anything but the arbitrary nature of the alphabet. If it could be proved that 'm' was placed before 'w' in the original structuring of this system, the alphabet, so the word 'men' could come before 'women' the case would be much stronger.
Deconstruction theories always concern me as having a tendency to be argumentative just for the sake of argument. Is it always valid to take something apart from a different time and show flaws based on current social norms? I do not believe so. Her argument that the universal model needs to be more open to new additions is a valid one, to me. Both DDC and LCC have difficulties in this area and new ways of organizing knowledge should be studied with as little bias as possible. However, if organizing knowledge is power, and Olson wants to organize it in a new way, isn't she seeking power as well? When an author's bias shows as clearly as hers, maybe she needs to re-learn her idea of an open view of knowledge.
|
|
|
| Pietris, M. K. |
[Jul. 19th, 2006|03:59 pm] |
Pietris, M. K. (1990). Library of Congress classification. In B.G. Bengtson and J. S. Hill (Eds.), Classification of library materials: current and future potentional for providing asscess (p. 60-80). New York: Neal-Schuman Publishers.
The Library of Congree classification system has its beginnings in 1897. The first schedules were published before World War I and reflect the world view at the time. This as proved to be a continuing problem, since that still exists to some extent. This basic idea, that the time period when a schedule was created was reflected in it's classifications and can still be seen. A listing of the schedules and dates of publication show this. Another problem is ennumeration. While Dewey places an -05 to indicate a periodical, LCC has to create a unique number for periodicals. The LCC schedules take up a large amount of space, therefore, and are difficult to read or understand. Also, unlike Dewey, there is not a heirarchy of subject matter reflected in the notation. The area of shelflisting presents another set of problems, as does the renaming of countries. The article works its way through each problem, with examples to support the discussion. Recent changes are also discussed.
I found this article more confusing than the Dewey article. I see that the LCC is more precise, is updated on a regular basis, keeps adding new schedules. I am not convinced it is superior to the Dewey, just different. It does seem better suited to the academic and research library, because of the way new fields of study are added. It also seems it is more difficult for the library to work with, with the constant updates and new numbers being added all the time. |
|
|
| Dewey, M. |
[Jul. 19th, 2006|02:51 pm] |
Dewey, M. (1972/1876). Catalogs and cataloging. In A. F. Painter (Ed.), Reader in classification and descriptive cataloging (p. 7 - 14). NCR Microcard Editions.
First used in the Amherst College Library in 1873, what is now known as the Dewey Decimal System is explained in some detail. This system was first designed to use in cataloging and indexing. The classification system is divided into nine categories, and nine divisions under each category or class. Dewey makes a special point of the practical usefulness that was a major goal of the system. He also states more than once that he believes this system can grow as needed in future. Cross-references in the card catalog was one way this could happen. Books were to be shelved with others in their category and division, thus creating specialty libraries for each discipline. This was seen as an advantage as each specialist would have their own library. Other advantages were seen as lower costs, less space for shelving, ease of use and instruction, flexibility, and practical rather than theoretical.
I grew up this this system, learning the nine classifications in grade school and using it through my first graduate school experience. I lived in the 800's and 900's. Reading this description of the system and seeing a list of the divisions under some of the classes, I can see flaws. The names of some of the classes, "Useful Arts" for example, are simplistic and so broad as to beg the question of just what isn't a useful art? Murder is useful, for instance, and can be seen as being done in an artistic manner. Would a book on that subject go here or under some other class? And, of course, there are the problems with where certain divisions come, Women being the prime example.
I can see why public libraries still use this system, though. It is easier to understand than the Library of Congress system, and it has been in use so long, many people know it better. Given time someone will create a better system, I don't doubt, but it too will have it's flaws. |
|
|
| Aitchison, J. et.al., #1 |
[Jul. 18th, 2006|02:40 pm] |
Aitchinson, J., Gilchrist, A., Bawden, D. (1997). Introduction ( pp. xiii-xvi), Figure 1 (p.6), Section D. Vocabulary Control (pp. 15-34). In Thesaurus construction and use: a practicl manual 3rd ed. London: Aslib.
As a manual for the construction of a thesaurus, this reading is very specific and detailed, but rather difficult to summarize without just quoting the section headings. The Intro. explains the purpose of the thesaurus, creating a vocabulary of controlled language for retrieval of information and how this edition is different from the last. I found the table on page 6 very interesting, as it shows the differences between natural language and controlled vocabulary. Section D. explains in detail just how that vocabulary is set up, controlled. The rules are first set out. The four major ways are to control the form of the term, the choice of a sysnonym, the choice of which slang or loan words to use as well as which trade names and proper names. Also, the meaning of the term is restricted to the most effective one. There follows specific examples showing how to create these controls. Under 'form of the term' are the preferred or descriptor term and the general categories - concrete, abstract and proper nouns.
I found the examples on Homonyms or Homographs interesting because this is an area I have had trouble with in searching for articles. This also applies to the non-preferred term and the word USE that appears at times. Scope notes, SN, also was an interesting section. I now see even more clearly why this programming is so costly. It would take even a skilled person, well acquainted with the thesaurus, a great deal of time to do this indexing. No wonder systems often go with this instead of the natural language system. It also helps make cleas the words that appear on articles in some databases under the keyword list at the beginning of the article. I got out some articles from other class readings and looked at the lists. I was able to see why specific words were chosen. In a keyword such as "Electronic publishing" or "licensing", is an example of the use of noun phrases and verbal nouns such as gerunds. Sometimes these keywords are rather ackward, but the rules help make that understandable, even if, as a person who values language, I don't like the 'new' word. |
|
|
| Lancaster, F. W. |
[Jul. 18th, 2006|02:05 pm] |
Lancaster, F. W. (1998). Natural language versus controlled vocabulary: some general considerations. In Indexing and abstracting in theory and practice, 2nd ed. (p. 227-232). Champaign, IL; Graduate School of Library and Information Science, University of Illinois.
I was unable to find the terms "indexing by extraction" or "indexing by assignment" in any of the readings for the 18th. So, I am working on the assumption that the former refers to natural language/exhaustive indexing and the latter refers to controlled vocabulary/selective indexing. After reading all the articles that deal with this subject, I have decided that no system is perfect, and programmers make their choices based as much on budget restrictions as any other consideration. As Lancaster says, natural language is closer to the user's language, and will be more helpful to this type of person. Controlled language favours the informtion specialist. It follows that each such user, then, would have more difficulty with the other type of indexing. Indexing based on an extensive abstract could provide more results, but they might be more imprecise than an index with controlled vocabulary. However, if the controlled vocabulary used did not have the most recent terms, it would be of less use. This is an area where human intellegence would provide greater effecency.
Human intellegence would be more expensive as programmers have a high salary and it would take more hours to set up this type of index. Controlled vocabulary can also be expensive to install and mantain, so the costs might even out. If the goal of the index is to provide a high output of results, exhaustive indexing would be more productive. The arguments remind me of a tennis match. Pros and cons seem evenly balanced. |
|
|
| navigation |
| [ |
viewing |
| |
most recent entries |
] |
| [ |
go |
| |
earlier |
] |
| |
|
|
![[info]](http://l-stat.livejournal.com/img/userinfo.gif){kind=small}
eloiseariel's journal