025.431: The Dewey blog

Fiction Finder and Dewey, Part 2: Place, Genre, Language

(See previous post Fiction Finder and Dewey, Part 1, for background information about Fiction Finder. Search results reflect searches done June 18, 2014, with the underlying database current through April; results may change as the database is updated.)

In Fiction Finder, "Place" is usually where the story is set, and many authors write about places not their home.  For example, an Advanced Search under "Place" for Italy retrieves 8956 records, but none of the first ten records is for a work by an Italian author; seven are by Shakespeare (e.g., The Tragedy of Othello, the Moor of Venice), and the others are by Henry James, Laurence Sterne, and E.M. Forster.  A user who seeks fiction set in Italy by Italian authors will have to go to record 19 in the set to find it: The Name of the Rose, by Umberto Eco.  Here is the start of the full record.

Clicking the Dewey number in that record (853.914 Italian fiction--1945-1999) retrieves 4839 records for works originally written in Italian—many of them set in Italy, e.g., The Leopard, by Giuseppe Tomasi di Lampedusa.  Here is the start of the full record.

To broaden the range of Italian literature to include all literary periods and to focus on works set in Italy, one can truncate the Dewey number and search under "Dewey" for 85* and under "Place" for Italy; 1180 records are retrieved.  One can then narrow the search by selecting from the "Genre" list at the left.  For example, one user might select "Short stories," retrieve a set of 11 records, and choose Rome Tales: Stories. Here is the start of the full record.

Another user, faced with 1180 records, might choose "Detective and mystery stories" and retrieve 47 records.  Nine of the first ten records are for works by Andrea Camilleri, e.g., Rounding the Mark.  Here is the start of the full record.

Let’s try another place: Ireland.  An Advanced Search under "Place" for Ireland retrieves 7391 records.  Two of the first ten records are for works by American authors, e.g., The Death of an Irish Sinner: A Peter McGarr Mystery by Bartholomew Gill, classed in 813.54 American fiction--1945-1999), and eight are for works by Irish authors writing in English, e.g., Tara Road, by Maeve Binchy.  Here is the start of the full record.

The standard Dewey number for works by an individual author writing in English (e.g., 823.914 English fiction--1945-1999) does not indicate whether the author is Irish.  In this context, what is a quick way for a user to find out?  In the full record, the user can click the cover art to go to the corresponding record in WorldCat.org, scroll down in that record and click to "find more information about" the author—in other words, go to the WorldCat Identities page for the author, e.g., for Maeve Binchy—and if appropriate, scroll down to find the "Useful Links" to the Library of Congress Authority File record and the Wikipedia article for the author.

How can a user of Fiction Finder find works written originally in Irish Gaelic, if the user doesn’t have in mind a specific author or title, and doesn’t know the Dewey number? (Note: Dewey numbers are searchable in Fiction Finder, but not their captions; the captions are displayed only to explain the meaning of the Dewey numbers.)  An Advanced Search under "Language" for Irish and under "Place" for Ireland retrieves 364 records—many of them not originally written in Irish, but having at least one edition translated into Irish.  (The Editions section of a full record has information about specific editions and their languages.)  The tenth record in the set of 364 is in Irish, An béal boċt; : nó, An milleánach. Droċ-sgéal ar an droċṡaoġal curta i n-eagar le Myles na gCopaleen.  Here is the start of the full record.

Clicking on 891.6234 Gaelic fiction—Irish—1875-1999 retrieves 118 records.  The first record is for Exile, by Pádraic Ó Conaire.  Here is the start of the full record.

The listed editions are all in English, but the edition details say "translated from the Irish by Gearailt Mac Eoin."  

(Note: Information about Fiction Finder was supplied by Diane Vizine-Goetz, Senior Research Scientist, OCLC Research. She has pointed out that OCLC Research is currently working to improve FRBR clustering with respect to translations, and the results will be reflected in Fiction Finder and other works-based applications. See "OCLC Research Launches Multilingual Bibliographic Structure Activity" and "Multilingual Bibliographic Structure.")

Fiction Finder and Dewey, Part 1: Folk literature and Films

"Fiction Finder is a works-based based application that provides access to millions of works of fiction described in WorldCat records for books, eBooks, audio books, movies and television. You can search by person, place, topic, genre, character, Dewey and more." Fiction Finder was developed by OCLC Research.

As a hidden part of Fiction Finder, DDC has been used in processing the database to pull in a broad range of stories of imagined events, including folk tales, graphic novels, films, and television, in addition to novels, short stories, drama, and narrative poetry.  For information about DDC and broadening Fiction Finder beyond textual novels and short stories, see the previous blog post "Improved Access to Fiction" and the conference paper "Finding Fiction: Facilitating Access to Works of the Imagination Scattered by Form and Format" (2006).¹

As with Cookbook Finder, DDC numbers have recently been made visible, with assignable numbers from DDC 23 (including built numbers) appearing with captions in detailed-record displays.  To search by Dewey number, click the class number in the detailed record. Or, if you know the DDC number you want, you can enter the number directly, using the drop-down menu in Advanced Search.  You can use the question mark (?) as a character masking symbol (to stand for a single character in your search) or use the asterisk (*) for left or right truncation.  (These Advanced-Search features are newly available also in Cookbook Finder.)

As in the WorldCat database from which Fiction Finder is derived, not every record has a DDC number. A DDC number displayed and searchable in the Fiction Finder database may be short because an abridged number was assigned originally. Or a WorldCat record may have a full built number, but Fiction Finder may give a shorter number because Fiction Finder gives only as much of a built number as can be found in the DDC 23 database (from which WebDewey is derived).  Why? Fiction Finder needs to be able to display a caption from the DDC 23 database with the number. Indeed, Fiction Finder shows only numbers and captions that match a schedule or built number in DDC 23 even if the number taken from the WorldCat database was assigned from another edition; that introduces inconsistencies when there have been relocations or discontinuations.  Nevertheless, the shorter or inconsistent numbers can still be useful for access, especially in combination with the many other elements available for searching in Fiction Finder.    

Search results reported here reflect searches done on May 16, 2014; results may change as the database is updated. 

To sample folk literature, let’s begin by using Advanced Search to do a title search for "Cinderella" (retrieves 886 records).  The first record retrieved, Cinderella, has 398.2 Folk literature.  Here is the start of the record:

A user who wants to branch out and consider other folk tales can click 398.2 (retrieves 3654 records) and look at the record for, e.g., Stone Soup: An Old Tale:

Another user, in response to the original title search for Cinderella, might select the third record retrieved.  (The second record has no DDC number.) The third record, Cinderella, has 791.4372 Single films.  Here is the start of the record:

 

Clicking 791.4372 retrieves 20761 records, with a variety of popular and classic films at the top of the list.  Clicking in the column at the left under "Genre" to limit to animated films retrieves 695 records.  Among the first ten is another film from Disney about a legendary character: Mulan.  

A title search for "Mulan" (retrieves 76 records) has among the first ten records a book version, Mulan Saves the Day, with the DDC number 398.2095102 Persons (Individuals)—folklore—China (built with 398.209 History, geographic treatment, biography of folk literature plus T2—51 China,  as instructed at 398.2093-398.2099 Specific continents, countries, localities, plus 0, as instructed at 398.2093-398.2099:01-08 Tales and lore on a specific topic, plus 2 from 398.22 Tales and lore of persons without paranormal powers, also as instructed at 398.2093-398.2099:01-08 Tales and lore on a specific topic).  (Note: In WebDewey the add note at 398.2093-398.2099:01-08 has been corrected to begin: "Add to 0 . . . . ") Here is the start of the record for Mulan Saves the Day:

Clicking 398.2095102 retrieves 12 records, including The Lost Horse: A Chinese Folktale and Ling-Li and the Phoenix Fairy: A Chinese Folktale.

A search for 398.2095* (using Dewey in the drop-down menu) retrieves 16 records, including The Hungry Coat: A Tale from Turkey, classed in 398.20956102 Persons (Individuals)—folklore—Turkey (built with T2—561 Turkey), and The Patient Stone: A Persian Love Story, classed in 398.2095502 Persons (Individuals)—folklore—Iran (built with T2—55 Iran).  

None of the 16 records, however, has been built with T2—52 Japan, because the DDC 23 database does not include a 398.209+ number built with T2—52 Japan. (Of course, the add note at 398.2093-398.2099 Specific continents, countries, localities allows classifiers to build 398.20952+ but the already-built number is not in the DDC 23 database and thus there is no caption for Fiction Finder to display.)  A search for Dewey 398.209* retrieves 2109 records.  Many of those records have only 398.209 History, geographic treatment, biography, because that is all that matches numbers in the DDC 23 database.  One way to identify some Japanese folk literature in that set of 2109 records is to search for Dewey 398.209 and Place Japan (if we did not already know that 398.20952+ is missing, we could search for Dewey 398.209* and Place Japan):

That search retrieves 56 records, including The Funny Little Woman and Tsunami! Here is the start of the record for Tsunami!

A searcher who wants to cast a wider net and further reduce the impact of inconsistencies in the DDC numbers taken from the WorldCat database can truncate the DDC number further, e.g., by searching for Dewey 398.2* and Place Japan; that search retrieves 132 records.
 
(Note: Information about Fiction Finder was supplied by Diane Vizine-Goetz, Senior Research Scientist, OCLC Research.)

¹Beall, Julianne and Diane Vizine-Goetz. 2006. "Finding Fiction: Facilitating Access to Works of the Imagination Scattered by Form and Format." In Knowledge Organization for a Global Learning Society: Proceedings of the Ninth International ISKO Conference (Vienna, Austria, July 4-7, 2006), edited by G. Budin, C. Swertz, and K. Mitgutsch, 295-301. Advances in Knowledge Organization, no. 10. Würzburg: Ergon.

World Digital Library and DDC

AP journalist Brett Zongker has brought to our attention that the World Digital Library (WDL) actively continues to expand.  With the latest additions of manuscripts from the Walters Art Museum in Baltimore, the collection has surpassed the milestone of 10,000 digital items.

The reason we take special note of the WDL—that is, besides how cool it is!— is that DDC is part of the WDL metadata.  Every item is assigned at least one three-digit Dewey number, and while the numbers are hidden, the captions from the Dewey summaries appear as topics—in seven languages.  For details, see this presentation.

Cookbook Finder and Dewey

"Cookbook Finder is an experimental, works-based application that provides access to thousands of cookbooks and other works about food and nutrition described in library records. You can search by person, place, topic, course, ingredient, method, Dewey and more." Cookbook Finder was developed by OCLC Research.  

Although Dewey has always been part of Cookbook Finder, it has now become visible, with assignable numbers from DDC 23 (including built numbers) appearing with captions in detailed-record displays.  To search by Dewey number, click the class number in the detailed record.

Detailed records in Cookbook Finder are long, but here is the beginning of a detailed record, showing the Dewey number:

 

If you find a work that you want, you can click the cover art (or the cover art place holder) in the detailed record to go to the record in WorldCat.org to find the work in a library.

What is included in Cookbook Finder?
•    Commercial, community, society cookbooks
•    Culinary biography
•    Literary cookbooks
•    Popular and scholarly works on food and nutrition
What formats? Print, electronic, video, audio, etc.

With Cookbook Finder, you can type "biography" into the basic search box, and the first result will be My Life in France by Julia Child, followed by biographies of other cooks—very different from typing "biography" into the search box of WorldCat.org or a typical library OPAC!

Using Advanced Search in Cookbook Finder, you can type "desserts" in the Subject box and "Italy" in the Place box to retrieve 43 works, including Sweet Maria's Italian Desserts: Classic and Casual Recipes for Cookies, Cakes, Pastry, and Other Favorites; and Dolci Toscani: The Book of Tuscan Desserts.  The detailed records display: 641.86 Desserts.  A bibliographic record may have a more specific number—641.860945 or 641.8609455 (built with 641.86 Desserts plus T1—09 History, geographic treatment, biography plus T2—45 Italy or T2—455 Tuscany region)—but those built numbers are not found in the DDC 23 database (the database distributed in WebDewey).  Only a number in the DDC 23 database will be displayed in Cookbook Finder, because the caption is needed to explain the number.  

Among the "Related Works" listed with Sweet Maria's Italian Desserts is Sweet Maria's Italian Cookie Tray.  The detailed record for that work displays 641.8654 Cookies.  Among the next steps planned for the Cookbook Finder is provision of browsing based on Dewey hierarchy; that would provide another way to get from desserts to cookies.

Using Advanced Search, you can type "Italian cooking" into the Subject box to retrieve 3863 results, including Everyday Italian: 125 Simple and Delicious Recipes and Heat: An Amateur's Adventures as Kitchen Slave, Line Cook, Pasta Maker, and Apprentice to a Dante-Quoting Butcher in Tuscany. The detailed records display 641.5945 Cooking—Italy and 641.59455 Cooking--Tuscany (Italy). In these cases, the built numbers are in the database (641.59 Cooking characteristic of specific geographic environments, ethnic cooking plus T2—45 Italy or T2—455 Tuscany region).

If you click the DDC number 641.59455 in the detailed record for Heat, you get 142 results.  If you are looking strictly for a cookbook, you can use the "Narrow Your Results" option at the left to limit a search to the genre "Cookbooks"; that leaves 7 results, excluding Heat but including The Tuscan Sun Cookbook: Recipes from our Italian Kitchen.

Also in the DDC 23 database are these numbers and captions:
    641.59451 Cooking--Northern Italy
    641.59457 Cooking--Southern Italy
    641.59458 Cooking--Sicily (Italy)
These represent another opportunity for future browsing based on Dewey hierarchy.

(Note: Information about Cookbook Finder was supplied by Diane Vizine-Goetz, Senior Research Scientist, OCLC Research.)

Navigating to Schedule Numbers in WebDewey

What if you search for the full WebDewey record for a specific schedule number, and you get too many results?  For example, what if you type 599 into the standard Search box on the WebDewey home screen, looking for the full record for 599 Mammalia (Mammals), and click Search?

The standard Search box in WebDewey searches all fields by default.  A search for 599 in all fields yields 680 records:

If you have the time and patience to look in that set of results, you will find that the record you want is number 454.

Why so many results for an all-fields search for 599?  A few of the unwanted results are caused by matches with table numbers (e.g., T2—599 Philippines), but most of the unwanted results occur because of notes that specify 592-599, especially the add footnote attached to many schedule records: "Add as instructed under 592-599."

When you stumble upon too many results, instead of looking through the long list of results, one simple response is to insert the index label "dd:" for Dewey Numbers in front of the number to override the default choice of index, then click the Search button immediately to the right of the index selection box (which still reads "All Fields") to search for dd:599

Then you retrieve three records, and the one you want is number 2:

(For a complete list of index labels, look in WebDewey Help under Search WebDewey for Search indexes. Or look at "Introduction to WebDewey 2.0" on the Dewey Training Courses site.)

Alternatively, you can get the same results as a search for dd:599 by using Advanced Search and searching for 599 in Dewey Numbers.  To do that, you do not need to click Advanced Search to go to the main Advanced Search input screen; instead, you can click the pointer to the right of the index selection box that says "All Fields" and change "All Fields" to "Dewey Numbers" then click the Search button immediately to the right of that box:

Another alternative approach is to browse Dewey Numbers (with Captions).  You do not need to click Browse at the top of the screen to go back to the main Browse input screen; instead, you can use the radio buttons at the left to switch from Search to Browse. The default browse field "Dewey Numbers (with Captions)" will appear in the index selection box at the right. Then you need to press <Enter> or click the Search button to the right of that box to start the browse:

With the Browse approach, the number that you want is the second number in the list (the immediately preceding number is given for context). You don’t see the Manual record for 599 in the browse results, but you can easily get to it via the See-Manual reference in the full record display for 599:

 

 

WebDewey Hierarchy Displays and Schedule Numbers that End with Zero

In the hierarchy box for the WebDewey record for 500 Natural science and mathematics, there appear at first glance to be three different entries for the number 500:

A closer look reveals that the 500 just above the entry for 500 Natural science and mathematics has the last zero grayed out, and only 50 is in blue and underlined as a link.  The 500 at the top has both zeros grayed out, and only the 5 is in blue and underlined as a link.  Why?  The three different entries are used to display different hierarchical relationships. In the hierarchy box for the WebDewey record for 500 Natural science and mathematics shown above, the entries in the downward hierarchy all have the format 500.X: 500.2, 500.5, and 500.8.  In the bar at the top of the hierarchy box, the focal notation is displayed as 500.  That notation matches the notation in the 153 field of the underlying MARC record (click the MARC button to see the underlying record):

153 ## 500 $e 50 $j Natural sciences and mathematics

In the hierarchy box, click the entry just above the record for 500 Natural science and mathematics, and the downward hierarchy in the format 50X will be displayed:

In the bar at the top of the hierarchy box, the focal notation is displayed without the grayed-out zero: 50.  That notation matches the notation in the underlying MARC record:

153 ## 50 $e 5 $j Science

Click the top entry inside the hierarchy box, and the downward hierarchy in the format 5X0 will be displayed:

In the bar at the top of the hierarchy box, the focal notation is displayed without the grayed-out zeros: 5. That notation matches the notation in the underlying MARC record:

153 ## 5 $j Science

Why do the two entries with grayed-out zeros have a shortened caption—Science—by comparison with the full caption for 500: Natural science and mathematics?  The captions for entries with grayed-out zeros come from DDC Summaries: "The headings associated with the numbers in the summaries have been edited for browsing purposes and may not match the complete headings found in the schedules."

Like entries in DDC Summaries, the entries for numbers with grayed-out zeros have no notes, no Relative Index entries, and no mapped terminology.

What about number-span entries, e.g., 580-590 Natural history of plants and animals in the hierarchy box for 5 Science above?  There are no number-span entries in DDC Summaries, and no modified headings for them.  The zeros are not grayed out, and clicking that entry in the hierarchy box leads to the full record with notes.

Overlapping Spans in WebDewey

In WebDewey, when a number is displayed in the "hierarchy box," above it are displayed its parent, its parent's parent, its parent's parent's parent, etc., while below it are displayed all the numbers for which it is the parent.  Sometimes you might be a little surprised by what you see or don't see there.  This is particularly the case when spans are involved -- and more particularly the case when overlapping spans are involved.

For example, when 220-290 is displayed, we find above it 200; we find below it 220, 230-280, and 290.

When 230-280 becomes the focal notation, we find above it 200 and 220-290; we find below it 230-270 and 250-280.

When 230-270 becomes the focal notation, we find above it 200, 220-290, and 230-280; we find below it 230 and 240.

But when 250-280 becomes the focal notation, we find above it 200, 220-290, and 230-280; we find below it 250, 260, and 270-280.

The situation at 230-270 (as compared to the situation at 250-280) may seem odd, but it is, in fact, correct.  After all, there is an entry for 230-270 Specific elements of Christianity, and its parent is 230-280, as shown.

But why are 250, 260, and 270 not shown as children of 230-270?  Because a DDC notation can have only one parent designated in the MARC classification record (153 field, subfield e).  The two spans, 230-270 and 250-280, have equal claim as candidate parents for 250 and 260; the two spans, 230-270 and 270-280, have equal claim as candidate parents of 270.  But in each case only one of the candidates can be designated the parent.  It so happens that 250-280 has been designated the parent of both 250 and 260, while 270-280 has been designated the parent of 270.

The problem arises because of overlapping spans.  The graphic below shows the overlapping of spans that occurs in the 200s:

As can be seen, 230-270 overlaps with both 250-280 and 270-280.  Fortunately, there are very few cases of overlapping spans.  But when you see something in a hierarchy that at first glance looks odd, consider whether any spans are involved, and remember that any DDC number or span of numbers can have only one parent.

Is This All They Have? Part 3

A previous post introduced the possibility of using DDC-based signage and QR codes to help patrons find works on the same phenomenon but from a different disciplinary perspective.  The post also issued an invitation for a few libraries to participate in a pilot project to evaluate the usefulness of such signage.  (See the post on “requirements” for participating in the pilot project.)  In the meantime, we share a few additional QR codes we suspect some users might find useful.

Imagine a patron looking at the materials on postal service in 383.  Would the following signage help?

Or imagine a patron looking at the materials on sports in 796.  Would the following signage help? 

DDC-Based Signage and QR Codes Pilot Project: What's Needed from You

A previous post introduced the possibility of using DDC-based signage and QR codes to help patrons find works on the same phenomenon but from a different disciplinary perspective.  The post also issued an invitation for a few libraries to participate in a pilot project to evaluate the usefulness of such signage. 

We have been asked what the requirements are for participating in the pilot project.  Here’s our sense:

1. You need to have a collection shelved by DDC numbers

2. You need to have an OPAC that can be searched by call number (Note 1. Sometimes such a search is possible, even if it isn’t advertised to users.  Note 2. Of course you want users to be able to read the search results on their mobile device.)

3. You need buy-in from colleagues and the powers-that-be to put up signage in your stacks

4. You need to be willing to scan two documents we have put together of (a focused subset of) DDC numbers that are related by see or see-also references and to identify which links to related numbers and subjects make most sense for your patrons (Note. If your DDC number usage is not aligned with DDC 23, you may need to make some modifications in the relationships shown.)

5. You need to determine what you want your signage to look like and will be responsible for its production (we can give you some guidance, if needed, on how to produce the QR codes)

6. We would ask that you gather patron comments about the signage—what works, what doesn’t work, what can be improved, etc.—to share with us so we can share with others

7. If the signage is helpful to your patrons, we would love for you to promote this search enhancement through whatever channels make sense

There is no need for all the libraries in the pilot to stick to the same schedule, so we invite you to join the pilot whenever the time is right for you!

Is This All They Have?

How often do your patrons (1) find a set of resources in the collection that are relevant (or maybe only semi-relevant) to their interests, but (2) end up wondering if there are other relevant (or more relevant) resources somewhere else in the collection?  The disciplinary basis of Dewey has as an intended consequence that works about the same phenomena may be located at different places in the classification system.  Shelving materials by DDC notation will then cause resources on the same phenomena to be subject to some degree of physical scatter as well.  While sophisticated users will appreciate the attendant disciplinary context and specificity that this provides, more casual users may not. 

Is there some way to cater to disciplinary and phenomenon perspectives simultaneously?  The DDC offers several tools that facilitate finding where phenomena are treated from different perspectives.  The Relative Index is one such tool; see and see-also references in the schedules are other such tools.  “Get real!” we hear you saying.  “We don’t make copies of the DDC / access to WebDewey available to our users.  Besides, how many of them would even be interested?” 

Ah, but what if you (with a little help from your friends, that is, us) were to identify the relationships found in the Relative Index and/or see or see-also references that are most important to your users and display them as signage in your collection?  And if you have an OPAC that is searchable by DDC number, what if your signage were to include a QR code that let users immediately see what you have at other locations in the collection?

For example, let’s say you have a patron looking at the materials on retirement at 306.38.  S/he wonders, “Is this all they have?”  And then they notice nearby something like the following:

Go ahead—whip out your cell phone and scan away.  Nifty, huh?  (Of course, the actual results would reflect the holdings of your institution and how search results from your OPAC display on a mobile device.)

We are interested in working with a few libraries who would be interested in piloting such a signage program based on relationships in the DDC.  If you are interested, please email me (Rebecca Green) at [email protected] (“DDC signage pilot” in your subject line would be a great help).  Meanwhile we will periodically present other DDC-based relationships and associated QR codes as enticements.  (Our thanks to Christine Burroughs, who helped us think through the possibilities here while she was at Shaker Heights Public Library.)