USING THE - Library of Congress



Library of CongressBIBFRAME Pilot 2.0 Training for CatalogersModule 2: The BIBFRAME Editor and the BIBFRAME Linked Data Store Prepared byTim Carlton, Les Hawkins, and Paul FrankCooperative and Instructional Programs DivisionLibrary of CongressJuly 2017Library of CongressBIBFRAME Pilot 2.0 Training for CatalogersModule 2: The BIBFRAME Editor and the BIBFRAME Linked Data StoreCourse SummaryThis module is part of a 2-module training curriculum, “Library of Congress BIBFRAME Pilot 2.0 Training for Catalogers.” In this module, participants will learn, hands-on, the interface of the BIBFRAME Editor, how to navigate, how to record bibliographic and authority data, how to save bibliographic data, how to search converted BIBFRAME Works and Instances, and how to work overall in the context of the BIBFRAME Pilot 2.0. Key components of the instruction are learning how to utilize the “profiles” in the Editor, and how to interact with the BIBFRAME Linked Data Store. An important additional objective is to prepare you to serve as a tester of the Editor and to provide valuable feedback through the instructors to the developers.Learning ObjectivesAt the end of this module, participants will be able to:access the BIBFRAME Editordefine the basic terminology used in the Editor navigate in the Editor interfaceselect the appropriate Resource Type and Profile to process the resource in handrecord descriptive and subject datarecord authority data for Agentsrecord data using various methods: dialogs, direct-entry, and menusutilize “AutoComplete” search functionsenter data as ‘literals’ when a search does not find appropriate existing recordsconfirm and edit the recorded dataadd multiples of the same field or elementpreview the created ‘record’save the created ‘record’retrieve the created ‘record’test the features of the Editor and contribute suggestions for refinementdiscuss the feasibility and practicality of Editor features and processessearch the BIBFRAME Linked Data Storeidentify appropriate Works and Instances in the Linked Data Storeselect appropriate Works in the Linked Data Store to link to new Instances apply an understanding of the purpose, focus, and limited objectives of the Pilotprovide feedback and suggestions to help developers improve the EditorTraining MethodsGuided exploration Duration3 hoursTable of Contents TOC \o "1-5" \h \z \u Unit 1: Getting Started PAGEREF _Toc488055522 \h 1What is the BIBFRAME Editor? PAGEREF _Toc488055523 \h 1How Do I Access the BIBFRAME Editor? PAGEREF _Toc488055524 \h 1For More Information PAGEREF _Toc488055525 \h 3Login and Authentication PAGEREF _Toc488055526 \h 3Remember -- This is a Pilot PAGEREF _Toc488055527 \h 3Unit 2: Overview of the Interface PAGEREF _Toc488055528 \h 5The Objective of This Unit PAGEREF _Toc488055529 \h 5Terminology PAGEREF _Toc488055530 \h 5The BIBFRAME Editor Workspace Screen PAGEREF _Toc488055531 \h 8Parts of the Screen – The Navigation Bar PAGEREF _Toc488055532 \h 8Parts of the Screen – The Create Resource List PAGEREF _Toc488055533 \h 11RDA Captions PAGEREF _Toc488055534 \h 12Types of Fields -- Dialog Fields PAGEREF _Toc488055535 \h 12Field Edit Buttons PAGEREF _Toc488055536 \h 13Types of Fields -- Direct-Entry Fields PAGEREF _Toc488055537 \h 14Multiples of a Field PAGEREF _Toc488055538 \h 15Black-Label Fields PAGEREF _Toc488055539 \h 16Adding Work Attributes -- The “Instance of” Button PAGEREF _Toc488055540 \h 17The Lookup Function (Scenario a) PAGEREF _Toc488055541 \h 18Typing the Work Elements (Scenario b) PAGEREF _Toc488055542 \h 18Creator of Work (RDA 19.2) PAGEREF _Toc488055543 \h 19Preferred Title PAGEREF _Toc488055544 \h 19Other Work Elements PAGEREF _Toc488055545 \h 20Reviewing and Editing RDA Work Elements PAGEREF _Toc488055546 \h 20Publication Information PAGEREF _Toc488055547 \h 20Adding Additional Instance/Manifestation Elements PAGEREF _Toc488055548 \h 21Entering RDA Expression Elements PAGEREF _Toc488055549 \h 21Scroll down the BIBFRAME Work (RDA Work Elements). PAGEREF _Toc488055550 \h 22Click on the Related Expressions BIBFRAME Work (RDA Expression) button. PAGEREF _Toc488055551 \h 22Content Type PAGEREF _Toc488055552 \h 23Language of Expression PAGEREF _Toc488055553 \h 23Previewing the Record PAGEREF _Toc488055554 \h 24A Brief Look at the BIBFRAME Output PAGEREF _Toc488055555 \h 25What About ISBD Punctuation? PAGEREF _Toc488055556 \h 27Recording and Trouble-Shooting Multiples PAGEREF _Toc488055557 \h 27Unit 3: Working in the Linked Data Store PAGEREF _Toc488055558 \h 29Background PAGEREF _Toc488055559 \h 29What is the Linked Data Store? PAGEREF _Toc488055560 \h 29How was the Linked Data Store created? PAGEREF _Toc488055561 \h 29Linked Data Store Navigation and Displays PAGEREF _Toc488055562 \h 31How to Interact with the Linked Data Store and Make it Your Friend PAGEREF _Toc488055563 \h 32The BIBFRAME Editor and the Linked Data Store PAGEREF _Toc488055564 \h 32Pre-Cataloging Searching PAGEREF _Toc488055565 \h 33Searching for BIBFRAME Works in the Editor PAGEREF _Toc488055566 \h 33Searching for BIBFRAME Works in the Linked Data Store PAGEREF _Toc488055567 \h 34How Do Voyager and the Linked Data Store Interact? PAGEREF _Toc488055568 \h 34Workflows PAGEREF _Toc488055569 \h 34The MARC 985 Field PAGEREF _Toc488055570 \h 34Unit 4: Authority Work, Administrative Metadata, and Item Information PAGEREF _Toc488055571 \h 35Authority Work for Names PAGEREF _Toc488055572 \h 35Name Authority Work for Agents PAGEREF _Toc488055573 \h 35Creator of Work PAGEREF _Toc488055574 \h 35Authorized Access Point -- Creator of Work PAGEREF _Toc488055575 \h 35Scenario a): Searching for an established Authorized Access Point PAGEREF _Toc488055576 \h 36Scenario b): Recording an unestablished Authorized Access Point PAGEREF _Toc488055577 \h 38Authority Work for Subjects PAGEREF _Toc488055578 \h 39“Traditional” Method PAGEREF _Toc488055579 \h 39“Component” Method PAGEREF _Toc488055580 \h 41Administrative Metadata for Instances PAGEREF _Toc488055581 \h 41Item Information PAGEREF _Toc488055582 \h 43Unit 5: Workflow for the LC Pilot PAGEREF _Toc488055583 \h 45Why is Workflow So Important? PAGEREF _Toc488055584 \h 45Where Do I Start? In the BIBFRAME Editor PAGEREF _Toc488055585 \h 45Then What? In Voyager PAGEREF _Toc488055586 \h 45Reporting Findings, Problems, Issues, Complaints … PAGEREF _Toc488055587 \h 46After You Leave the Classroom: What Should You Do? PAGEREF _Toc488055588 \h 46Unit 6: Help, Support, and Other Resources PAGEREF _Toc488055589 \h 47iCohere BIBFRAME 2.0 Pilot Portal PAGEREF _Toc488055590 \h 47The BIBFRAME mail box PAGEREF _Toc488055591 \h 47LC’s BIBFRAME Webpage PAGEREF _Toc488055592 \h 47Feedback Sessions PAGEREF _Toc488055593 \h 48Appendix A. RDF for BIBFRAME PAGEREF _Toc488055594 \h 1Unit 1: Getting StartedWhat is the BIBFRAME Editor?The BIBFRAME Editor (BFE) is a tool that enables input of BIBFRAME vocabulary elements.? The BFE will be frequently improved as users comment and as features are added, refined, and enhanced. ?How Do I Access the BIBFRAME Editor?Open Chrome and go to the BIBFRAME Editor by pasting in the URL: this siteYou should now see the Browse screen: We will discuss the functions of the browse screen later in the next unit. For now click “Editor” to access the BIBFRAME profiles. Click on the Editor tabThe link at the bottom leads to useful information.Click on the link to bibframeFor More InformationNote the description of BIBFRAME at the top of the page (highlighted in the image below). Most of the links on this page lead to information that is fairly technical. Among the links you might want to explore is the link to “Implementation, Tools and Downloads” (highlighted at the bottom). Login and Authentication For the BIBFRAME Pilot, there is no login procedure. Use of the Editor is strictly limited to those working in the Pilot. Once BIBFRAME is implemented, there will be access and login requirements, and other features intended to ensure the integrity and security of Library data. Remember -- This is a PilotWe know you will have questions as you use the Editor -- and suggestions. You may find it frustrating at times, but do not give up.? We need your input and expertise to improve the tools. We, the instructors, have learned through working with Network Development and MARC Standards Office (NetDev) staff members that creating tools for BIBFRAME is a lot more complex than it may appear. To get through this successfully, we all have to be flexible and patient. In the end our flexibility and patience will pay off.?Creating this ‘brave new world’ is a work-in-progress. Some features had to be simply tried -- to see how they work, and ultimately to enable refinement and fine-tuning.? A significant part of your role is to test the Editor -- to tell us what works (or doesn’t), what is user-friendly (or isn’t), what could be less confusing or more clear. Your help is needed to design the best tool possible.?BIBFRAME is an entirely new way of doing what you have been doing for a long time – for some of you, for decades. Your mastery of MARC -- like Rome, as the saying goes -- was not built in a day. We’ll start slow, and build -- and do a lot of repetition.Unit 2: Overview of the InterfaceThe Objective of This UnitIn this overview, our objective is limited: to become familiar with the terminology, look, and feel of the Editor, and gain some experience using the fields, buttons, boxes, on-screen features, and functions. We won’t worry too much about the logic of some of the elements we play with. We will focus more closely on creating descriptions of the resources you catalog. TerminologyDescription:the term description is used in this manual to refer to the cataloging of the resource in hand and the output from the BIBFRAME editor. Profile:an online template for creating a description of a resource or concept. The profiles used in the BIBFRAME Pilot 2.0 are: monograph, notated music, serial, cartographic resource, sound recording, and BluRay/DVD Workspace:the area of the Editor in which you utilize dialogs, free-text fields, menus, buttons, and look-up functions to create the BIBFRAME record. Also, the Browse link takes you to a workspace where your BIBFRAME records are savedFields:separate spaces on the template in which you enter cataloging dataDialog Field:a field represented by a button; clicking on the button displays a dialog in which you record data in text boxes; these dialogs sometimes involve separate elements (e.g., publication data) and can utilize either “Lookup” or the recording of “literals”Direct-Entry Field:a field represented by a text box in which you enter data directly, e.g., for transcribed fieldsField Edit Buttons:buttons which appear after you ‘set’ data; clicking on the ‘pen’ allows you to revise the data; clicking on the ‘trash can’ allows you to delete the dataLookup:(also known as “AutoComplete”) a function which searches the database as you type and displays a menu of matching terms, from which you select the term to be used in authorized access points or other ‘controlled’ fields (e.g., “Language of the Expression,” “Relator Role,” etc.) Data Boxes:a box that displays the specific data you have entered in a field; in some cases, this boxes contains default data which you can change as neededPublishA function in the BIBFRAME editor used to send your completed resource description to the BIBFRAME database.Searching Searching can be done in two places:The BIBFRAME EditorThe Linked Data StoreThere are advantages and disadvantages to both. Whichever search strategy you use, the lookup is going to the BIBFRAME 2.0 database, called the Linked Data Store. Searching will be covered in greater detail in Unit 3, but since searching is enabled through the BIBFRAME Editor, let’s take a quick look at that, and come back for more detail in Unit 3. In the BIBFRAME Editor, you can search by name/title combination and by title with a “type ahead” function. Name title searches must be entered in author title form as in the example below. Search for Work:Result:When the desired Work is found, press the Save changes button. The Work is now associated with the Instance you are cataloging. The search interface in the BIBFRAME Editor is limited to the type ahead with exact matching on the author/title or title. Additional search parameters will be added to the lookup as the pilot progresses.The Linked Data Services (LDS) interface on the other hand provides extended search parameters. The LDS (sometimes referred to as Linked Data Store – this is how we will identify it for the purposes of the BIBFRAME Pilot) is an online platform for linked data versions of Library's bibliographic and authority data. It may be used to determine if the Instance you have in hand has already been cataloged and has a Work associated with it.Working in the Linked Data Store will be covered in Unit 3. The BIBFRAME Editor Workspace Screen Open Chrome and go to the BIBFRAME Editor by pasting in the URL: this siteYou should now see the Browse screen:Parts of the Screen – The Navigation BarThe Navigation Bar at the top of the interface includes tabs to:BrowseEditorLoadThe Browse interface is where your BIBFRAME descriptions are temporarily saved before they are finally published to the BIBFRAME database. All of the descriptions catalogers contribute are saved in a temporary holding file. The file may searched from the Browse interface and descriptions retrieved to further revise before final publishing to the BIBFRAME Database.The Publish function will be discussed later in the manual. The development of the Publish function has not been completed. We will let you know when it is available and how to use it later in the pilot. To search the Browse interface, use any part of the the id number (far left below), the “name” assigned to the description by the editor, parts of the title or the catalogers Windows id (under the comment header below) to retrive cataloged resources. Use the Edit button (far left below) to retrieve and revise your work. Click on the Load tabThis takes you to the Load interfaceThis is where URIs for Works in the BIBFRAME 2.0 database can be retrieved and edited. When fully functioning, the URI for the Work will be inserted and linked to your Instance description. The development of the load function has not been completed. We will let you know when it is available and how to use it later in the pilot. Click on the Editor tabThis takes you to the Editor workspace0-952500Parts of the Screen – The Create Resource ListThe “Create Resource” list at the left is the starting point for cataloging using the BIBFRAME Editor. The drop-down menus under each resource type – Monograph, Notated Music, Serial, Cartographic, Sound Recording: Audio CD, Moving Image: BluRay DVD, Moving Image: 35mm Feature Film, Prints and Photographs, and Authorities – allow you to access a profile for two of the three primary BIBFRAME core classes (Instance and Work); or for a concept that has a relationship to the core classes (Authorities). Just as in RDA, in BIBFRAME 2.0, authorities for persons, families, corporate bodies, meetings, and jurisdictions are termed Agents. For now, we are just looking at the Interface to observe how the Editor behaves. We will type some ‘test’ data. To simplify your first experience with the Editor, we will choose the simplest profile – which is not always the most appropriate for the resource you have in hand. From the “Create Resource” list, click on MonographClick on Instance0-381000Parts of the Screen – The WorkspaceNote the legend at the top: A BIBFRAME “Instance” is roughly equivalent to an RDA “Manifestation.” This profile would be used, for example, if you do find a BIBFRAME Work description for the resource you have in hand – for an already-cataloged different manifestation, say.For the moment, we will skip discussion of the “Instance of” area at the very top.RDA CaptionsThe blue captions to the left of each field are active links to that element in the RDA Toolkit. The first time you click on a caption, you will need to log into the Toolkit (i.e., the LC account); the Toolkit will then open directly to that specific instruction. Subsequent clicks will open the Toolkit in another browser tab, again to the specific instruction. This is a really useful feature! Types of Fields -- Dialog FieldsThe first ‘type’ of field used in the Editor consists of a shaded button, labeled with an instruction (e.g., “Instance Title,” “Instance Title Variation,” Publication Activity,” etc.) or a ‘shorthand’ for the element. When you point to one of these buttons, it becomes shaded more darkly and the pointer becomes a ‘hand.’ If you click on one of these “Dialog Fields,” you see a dialog in which you record the appropriate data.Click on the Instance Title button next to the caption “Title Information”The cursor is already in the dialog’s text boxType Linked data for cultural Click on the + iconClick on Save changesField Edit Buttons-1143003429000Beneath the button, a Data Box displays the content, and two buttons can be used to edit or delete the field.Oops -- the title proper is actually Linked data for cultural heritage.Click on the yellow pen to get to an editable dialogClick again on the yellow pen -- i.e., on the next screen -- to actually edit the dataClick in the text box and add heritage at the end of the existing title. Click on the + iconClick on Save changes190503810000The edited field now looks like this. You can repeat as necessary to edit the data.For now, we won’t click on the red trash can to delete the data.Note:Instead of clicking on the + and Save changes buttons in the steps above, you can press the <Enter> key on the keyboard for each step. Because your hands are already on the keyboard, you might find this alternative method slightly easier, ergonomically.Let’s take a quick look at our very sketchy record so far:Scroll to the bottom of the screen and click on PreviewNote the Save and Publish buttons. The Save button is used to save your record in the temporary browse file where you can retrieve it for revision. As mentioned above the Publish button allows you to push your final description to the BIBFRAME 2.0 database. Remember, development of the Publish function has not been completed. After a brief discussion of what you see, click on the Back Arrow to continue creating the recordTypes of Fields -- Direct-Entry FieldsThe second ‘type’ of field consists of a text box, labelled faintly with the RDA name of the element. If you click in these “Direct-Entry Fields,” the border ‘glows’ blue and you simply type the ‘literal’ data directly into the text box, and then click on the “Set” button. We will see in the next exercise that some of these “Direct-Entry” fields offer menus of controlled terms, from which you simply select the appropriate option. Click in the Statement of Responsibility Relating to Title Proper (RDA 2.4.2) direct-entry text boxType Ed Jones and Michele SeikelClick on the + icon or press the <Enter> keyOops!You realize that you recorded only part of the statement of responsibility. A function was given on the resource: Ed Jones and Michele Seikel, editors. Hint:If the Data Box does not display the full content of the field, ‘hover’ the cursor over the Data Box; in a few seconds, the full data will display in a ‘tooltip.’To edit the data:Click on the yellow pen under the Statement of Responsibility boxChange the text to Ed Jones and Michele Seikel, editorsClick on the + icon or press <Enter>‘Hover’ the cursor over the Data Box to see the revised content in the tooltipDiscussion Point:What about the period at the end of the Statement of Responsibility? Multiples of a FieldIf you need multiples of a field, simply type one field, click on the + icon, and then repeat the procedure. You should see a ‘double set’ of Data Box and Field Edit Buttons for each field. If you need to edit one of them, just make sure to click on the proper ‘edit pen.’Black-Label FieldsWhat do you notice about the fields that are labelled in black?Correct! These are not specific RDA elements. They are either aggregates of elements or fields containing locally-defined elements. Adding Work Attributes -- The “Instance of” ButtonRemember that we originally selected the “Instance” profile. We could have instead selected the “Work” profile to record all the data related to the Work. We knew that a Work record already exists for Linked data for cultural heritage – and we could search for that Work record using ‘AutoComplete’ or search in the LDS to incorporate it into our Instance. We chose the “Instance” profile, which includes only those fields related to the recording of instance/manifestation data. You can use the “Instance of” button to help with one of two scenarios: Lookup. If you know there is a related instance already in the database – e.g., for an earlier edition – you can ‘lookup’ the Work elements for that instance;OR Typing the Work Elements. If you want or need to explicitly add the Work elements at this time, you can input the data. This scenario applies when a Work description does not exist in the BIBFRAME 2.0 database. In the BIBFRAME Instance (RDA Manifestation) profile, click on the BIBFRAME Work (RDA Work Elements) button: Note the two parts of this dialog, separated by “OR,” related to the scenarios described above:01968500Above the OR is the ‘Lookup’ function, used to find and ‘import’ existing Work data, if a name/title record exists in the converted database.Below the OR is the alternate way to use this dialog – Typing the Work Elements directly.You should recognize these elements as the FRBR/RDA elements of a Work – Creator or Work, Form of Work, Date of Work … etc. Right?You should also recognize that some of these are Dialog Fields and some are Direct-Entry Fields.The Lookup Function (Scenario a)In the Lookup field, we are looking for the Work RDA and serials cataloging, by Ed Jones. In this lookup, we will try a left-anchored search. Think about the RDA authorized access point for the work. Since there is a creator, the creator’s name, in its established form, precedes the Work title. So we will key in Jones, Ed, 1951- RDA and serials catalogingClick on Save changes‘Hover’ the cursor over the Data Box to see the BIBFRAME link to the Work The ‘Lookup’ function (our scenario a) allows you to find the Authorized Access Point (i.e., the Creator/Preferred Title combination) and ‘import’ it into the BIBFRAME record in one step; but this only applies if the Work IS in the database.The next function (our scenario b) applies to those cases when the Work does NOT already exist; you record each individual part of the Authorized Access Point (this is a variation of the scenario we will work with in our next Unit)Typing the Work Elements (Scenario b)Click on the red trash can to delete the “Instance of” element (you would not normally do this; we are doing it in training to allow you to practice the full functionality)Click again on the BIBFRAME Work (RDA Work Elements) buttonIn this scenario, we’ll use the fields below the “OR” …Creator of Work (RDA 19.2) Click on the Agents Contribution button next to the caption “Creator of Work”Click on the Person buttonIn the “Lookup” text box, type Jones, Ed and wait to see a menu of matches Select the menu choice Jones, Ed 1951- ‘Hover’ the cursor over the Data Box What does that string represent?Right! – the “Authorized” form for the CreatorNote that the searching for and selecting a creator ‘populates’ the Data box.Click on Save changesIn the “Search Relationship Designator” text box, type autSelect AuthorClick on Save changesA Note About Searching Techniques:-- the comma and space after the surname are critical -- or you will get very different results-- you can truncate the search -- but be sure to type enough to yield a manageable set of results; unless the surname is extremely unique, you will probably want to type some of the forenameWe have now completed the “Creator” portion of the Authorized Access Point. We will stay on this same dialog to complete the “Preferred Title” portion of the AAP. Preferred TitleClick on the Title button next to the “Preferred Title” labelNote:This is not a ‘Lookup’; you are recording data, so you must record the exact Preferred Title, spelled and capitalized precisely according to RDAIn the text box, type RDA and serials cataloging Click on the + icon or press EnterClick on Save changesIf necessary, ‘hover’ the cursor over the Data Box for the Preferred TitleOther Work ElementsThere is no conflict between our Authorized Access Point and another already in the database, so we don’t need to record Form, Date, Place, or Other Distinguishing Characteristic. Also there is no Other Agent Associated with a Work in this case. But we can add a few more Work elements. In this brief exercise we will not add any other attributes of the Work.Scroll down to the bottom of the dialog and click on Save changesReviewing and Editing RDA Work ElementsClick on the yellow pen under the “Instance of” button. Note that the tooltips display the identifiers, not the human-readable data strings.If you do need to edit one or more of these elements:You may need to click on several ‘pens’ to drill down to the specific value you need to edit. If you review a field but decide not to edit it, click on “Cancel”You can now continue where you left off and add more Instance elements.Publication InformationScroll down and click on the Publication Activity buttonThis dialog allows you to enter the sub-elements of the Publication Statement. Of course, none of the fields are ‘Lookup’ – and this makes sense. You must transcribe these elements exactly as on the source. You are not looking for an ‘authorized form.’Type ChicagoClick on the + sign or press <Enter>Type ALA Editions, an imprint of the American Library AssociationClick on the + sign or press <Enter>Type 2013Click on the + sign or press <Enter>The dialog should now look like this:Note each Data Box, displaying the data string for that element If you need to edit or delete an element, click on its yellow pen or red trash canClick on Save changesAdding Additional Instance/Manifestation Elements Further down, default values already appear for “Mode of Issuance,” “Media Type” and “Carrier Type.” Since we are cataloging a “book,” we will accept the defaults of single unit, unmediated and volume. We did not add the “Content Type” yet because it is in the BIBFRAME Work (RDA Expression Elements) profile.Add an LCCN in the field on this page.Note the field at the very bottom: “Your Windows ID and Comments.” This will serve a purpose similar to the MARC 955 field.Entering RDA Expression ElementsThe profile for adding expression elements is accessed through the BIBFRAME Work (RDA Work Elements) profile.Select the Instance of” button on the Instance profileScroll down the BIBFRAME Work (RDA Work Elements).Click on the Related Expressions BIBFRAME Work (RDA Expression) button.The Expression profile will open:Content TypeScroll down to the Content Type fieldNote that the Content Type “text” has been supplied by default.right000We will accept this default, but just to see how the drop down menu works we will select an additional Type for photographsPut your cursor in the text boxSelect still image from the drop down menuThe profile shows both Content Types have been recorded:Language of ExpressionThis is an RDA core element, currently recorded as a code in the MARC 008 fieldClick on the Language of Expression fieldThe following popup appears:Type eng in the language of content boxSelect English from the menuClick on Save changesFor this exercise we will add no more Expression elements. To save and close both the Expression and the Work profiles:Click the Save changes on the BIBFRAME Work (RDA Expression) profile.Click the Save changes BIBFRAME Work (RDA Work) profile to return to the Instance profile.Previewing the RecordAt the bottom of the BIBFRAME Instance (RDA Manifestation) pageClick on PreviewNext, you can either “Save” the record (if the description is correct) or click on the Back arrow to edit it. You should start learning how to interpret the RDF data you see on the next pages.A Brief Look at the BIBFRAME OutputWe now see two views of the BIBFRAME data: an RDF view (rendered in Turtle- Terse RDF Triple Language) and a JSONLD view. We will look briefly at parts of the RDF view. Remember that we entered only a few elements.In this top portion of the record, we see (highlighted)RDF namespace declarationsRDA Carrier Type (value 1049 = (‘volume’)RDA Mode of Issuance (value 1001 = ‘single unit’) and RDA Media Type (value 1007 = ‘unmediated’) BIBFRAME Work URLPublication ActivityTitle ProperAnother way to view the RDF triples produced by your description is to use the Visualize feature at the bottom of the preview page. Scroll to the bottom of the page to use this feature.The visualization represents the data you input as RDF triples. Click open the bubbles to reveal more of your data and how they are rendered in RDF.What About ISBD Punctuation?MARC is being replaced; that is one reason for the intended transition to BIBFRAME. However, ISBD punctuation is not going away – at least not entirely. What do you do about punctuation in the BIBFRAME context?Generally:The ISBD punctuation that comes before a MARC field or subfield and has a direct one-to-one correspondence with an entry box in the BIBFRAME Editor is not necessary. ?So, for example, you do not need to record the slash (/) that precedes the Statement of Responsibility (a system that chooses to display BIBFRAME data with ISBD punctuation could supply that punctuation on display, if so desired).The ISBD punctuation that has no direct correspondence to MARC subfield coding will continue to have significance in conveying meaning to humans.? So, for example, you do need to record the semicolon ([space];[space]) that separates multiple statements of responsibility entered in the same entry box in the BIBFRAME Editor.From a Linked Data perspective, however, it is better to record separate statements, rather than stringing two or more together using ISBD punctuation. Each statement creates one RDF “triple.” RDF technologies can parse individual triples much more efficiently than multiple statements in one triple. The BIBFRAME Editor will help you here, as you would not need to add punctuation that would appear between input boxes on the form. You would only need to supply punctuation within an entry box on the profile.For example, the colon (:) used in the Publication Statement in ISBD to separate the Place of publication from Publisher’s name does not need to be input.? Therefore, you should record (using our earlier example): Cambridge, England in the “Place of Publication” field and Cambridge University Press in the “Publisher’s Name” fieldNOT Cambridge, England and : Cambridge University PressAND NOT Cambridge, England : and Cambridge University Press, Recording and Trouble-Shooting MultiplesIn some cases, we have recorded multiples of the same field (e.g., “Note on Manifestation,” “Media Type,” “Carrier Type,” “ISBN”, and some others). It is important that you understand the proper techniques for doing this – and how to trouble-shoot any problems that may arise.After you record one of the fields, simply click again in the box or on the button and record another field (or, for fields such as Content, Media, or Carrier Type, select another option from the menu).Remember to “Set” or “Save” the new field.Generally, do not click on the yellow pen icon to edit the existing field, unless that is in fact what you really want to do (for example, if you made a typographical error in a literal field such as a note). As a last resort, click on the red trash can icon and re-do the field.Unit 3: Working in the Linked Data StoreBackgroundWhat is the Linked Data Store?In order to create a complete cataloging environment for the BIBFRAME Pilot, there needs to be a BIBFRAME database of bibliographic descriptions to search against. The BIBFRAME database needs to be a “live” database as well, so new BIBFRAME descriptions that you create can be added and edited. So, the BIBFRAME database not only is used for searching, it is used as a repository for all the work done in the Pilot. That is why it is called the Linked Data Store. In the first phase of the Pilot, the Linked Data Store was static. It could be searched and consulted, but existing descriptions could not be retrieved for editing, and new descriptions could not be added. In this phase of the Pilot, there are considerable enhancements to the Linked Data Store. These enhancements bring the cataloging environment much closer to that goal of recreating a real-world cataloging scenario. 7899403333750 HYPERLINK "" HYPERLINK "" was the Linked Data Store created?The Linked Data Store was created from two different sources – the LC/NACO Authority File, and the LC bibliographic database. You are familiar with both these databases through your cataloging work in Voyager in MARC.Just as a MARC bibliographic record includes attributes of different RDA entities – for example, in a MARC bibliographic record, the 1XX may be the creator of an RDA Work, the 245 may be the Title Proper of a Manifestation, and a 5XX note may refer to any of the RDA entities Work, Expression, Manifestation or Item – the MARC bibliographic record also includes attributes of the three BIBFRAME core levels of abstraction: Work, Instance, and Item. So converting the MARC bibliographic database from Voyager to BIBFRAME was not a simple task. Remember also that in the LC/NACO Authority File, there also are BIBFRAME Work (and RDA Work and Expression!) descriptions. The Linked Data Store was created by first identifying all the BIBFRAME Works in the LC/NACO Authority File. These works would appear as titles (think MARC 130) or as name-titles (think MARC 1XX $a + $t). All Works from the LC/NACO Authority File were converted to BIBFRAME Works first. Then the LC bibliographic database was converted to BIBFRAME. In this case, BIBRAME Work information and BIBFRAME Instance information needed to be “teased out” of the MARC bibliographic descriptions. So whereas the conversion of the titles and name-titles from the LC/NACO Authority File was a 1:1 process, the bibliographic records needed to be split apart, and then a deduplication process needed to be applied to make sure that there were no duplicate Work descriptions. In the Linked Data Store, you can identify the source of the descriptions from the label to the right of the brief display. The four descriptions above represent BIBFRAME Works. Let’s take a closer look at one of them, Buffalo Jones’ adventures on the Plains. Linked Data Store Navigation and DisplaysWhen you click on the brief display for Buffalo Jones’ adventures on the Plains above, here is what you see: This is a MARC XML display, and not a BIBFRAME display. You might not recognize it as a MARC view because it uses label names instead of MARC tags. But let’s look at the View Tagged DisplayThis is the MARC view that probably will look more familiar to you. Remember, it is not a BIBFRAME display either The doc link takes us more into a BIBFRAME world. This is a true BIBFRAME view of the complete MARC record, including both Work and Instance information Here is the Instance that is associated with the BIBFRAME Work record we started with. The view is in RDF and only includes attributes of the BIBFRAME Instance. There is no BIBFRAME Work information in this display The four links here allow you to access different serializations of the BIBFRAME Work description. These show only the BIBFRAME Work descriptionThe Bookmark This Item link might be the most important link on this screen. Don’t be confused by the phrase “Bookmark This Item” – “Item” in this case means the BIBFRAME Work description, not a BIBFRAME Item description. If you were cataloging a new Instance of this Work, you would be able to copy the Bookmark link and paste it into the BIBFRAME Editor in order to complete the description of the new Instance. Because the Linked Data Store is a work in progress (and in this case, that is a good thing, because it means that changes that you request can be considered and implemented!) the interface will be changing over the span of the pilot. So what you see in the images in this manual might change over time. Remember, flexibility. How to Interact with the Linked Data Store and Make it Your FriendWe walked briefly through the Linked Data Store display above, but now we want to look at how you will interact with the store. The BIBFRAME Editor and the Linked Data StoreThe majority of your work in the Pilot will be done in the BIBFRAME Editor. But the Editor and the Linked Data Store are synchronized so you can lookup Work and Instance descriptions in either the Editor or the Linked Data Store. There are advantages to both methods (and also some disadvantages)! Let’s talk through them. Ideally, you would want to search the Linked Data Store through the Editor. This makes a lot of sense since you will not need to toggle from one site to another. Here is an example of searching the Linked Data Store through the Editor:You will notice that only the title is searchable when using the Editor to search a description in the Linked Data Store. This might be difficult if you are searching for a common-sounding title, where many results are likely. You would need to be certain that the Work title you see in the result screen is the correct one to attach to your new Instance. The Editor, at least right now, does not give you the ability to look more in-depth at a Work description. So you may want to search for your Work description first directly in the Linked Data Store. Looking up a Work in the Linked Data Store opens up many more search strategy options. So even though you need to toggle to a different site, you can be assured that you are identifying the correct Work with no doubts, and then you can copy the Work title to the Editor so you can complete your Instance description. Here the same search from 1) above, but done in the Linked Data Store as an “Everything” search: Do you see the advantage of searching in the Linked Data Store? Not only do we see the Work description that we want, we also see that there is an Instance description as well. By selecting “Everything” in the search, we allow all descriptions to be displayed. So there is a good chance that the Instance we have in hand has already been described – and we have a duplicate copy. In the Linked Data Store, there is a much more robust searching capability than in the Editor. If you find the Work you need using the Editor lookup, all you need to do is click on it to save it, and then add your Instance description. But what if you find the Work you need in the Linked Data Store? You need to copy the title, and paste it into the Editor lookup box. Pre-Cataloging SearchingDuring the Pilot, you will need to catalog first in BIBFRAME and then in Voyager. Voyager will remain the database of entry, though. Let’s walk though some possible cataloging workflows.Searching for BIBFRAME Works in the EditorYou have a BIBFRAME Instance (an RDA Manifestation) in hand. You want to search to see if a BIBFRAME Work exists for this Instance. If a Work exists, you can select it in the Editor, and add your specific Instance information. In the Editor, a Work search will search on title or name only. The default search is keyword, not left-anchored. So it is a good idea to search directly in the Linked Data Store. Searching for BIBFRAME Works in the Linked Data StoreYou have a BIBFRAME Instance (an RDA Manifestation) in hand. You want to search to see if a BIBFRAME Work exists for this Instance. If a Work exists, you can select it and copy it to the Editor, and proceed with the description of the Instance you have in hand. How Do Voyager and the Linked Data Store Interact?WorkflowsThe work that you do in Voyager will remain the catalog of record. The work that you do in BIBFRAME will be used for testing in order to further refine the BIBFRAME cataloging environment. The Linked Data Store will be refreshed daily with new and updated MARC records from Voyager. That means all the bibliographic work that was done in Voyager on the previous day. So for a non-BIBFRAME Pilot participant, this is a good situation, although it might not be something that non-Pilot participants are aware of. But for you as a Pilot participant, you want the work you do in BIBFRAME, even though it does not represent the catalog of record, to not be overlaid by the work you do in Voyager. The MARC 985 FieldFor that reason, when you do your Voyager bibliographic work – fortunately, you do not need to do this for your authority work – more on that in the next unit – you want to identify the work you do to prevent it from migrating back to the Linked Data Store in the overnight distribution. So you should add a 985 field to the bibliographic work you do in Voyager – whether it is for a new description or an update to an existing description – with this text:Unit 4: Authority Work, Administrative Metadata, and Item InformationAuthority Work for Names In the BIBFRAME Pilot Phase One, the primary concentration was on how the BIBFRAME model could be used to represent RDA resource descriptions. This meant that the bibliographic aspects of cataloging were front and center. In BIBFRAME Pilot Phase Two, more experimentation is being done with authority work. Name Authority Work for AgentsName authority records are not a part of the Linked Data Store. However, there is a Linked Data Service, id., that contains name authority records and subject authority records.In this second BIBFAME pilot, as in Pilot Phase One, you will be linking out to the authority records in id., although the process will be transparent. However, in this pilot, you can input authority descriptions for Agents. Just as with your bibliographic work that is also done in Voyager, you will need to do your authority work in Voyager as well. There is a difference, though: the authority work you do in Voyager will not migrate to the Linked Data Store. That is why you do not need to add the 985 field (see the previous unit) to your Voyager authority work. Let’s look at the BIBFRAME Editor and name authority work for Agents in greater detail: Creator of WorkAuthorized Access Point -- Creator of WorkYou search the Linked Data Store and discover that a Work you need is not there. So you need to create a description of the Work using the Editor. 8382001460500Click on the Agents Contribution button Click on the Person buttonThere are two options, depending on whether a NAR for the creator already exists:If the creator is already in the database, “Lookup” the Authorized Access Point.If the creator is not already in the database, create the Authorized Access Point using the data input section under the “OR” breakScenario a): Searching for an established Authorized Access Point In the “Search …” box, type Carlton, TimSelect the correct authorized access point*** Did you notice that you cannot access the entire authority record from this display? Here is a case where you want to be sure to know in advance the correct authorized access point – Voyager will help! ***The name is now ‘recorded’ as the Creator (Searching + Selecting populates the “Record” box).Click on Save changesNow we will assign the appropriate Relationship DesignatorClick in the Agent Role boxType autSelect AuthorBoth the Creator and the Relationship Designator are now ready to be “Saved.”Click on Save changesScenario b): Recording an unestablished Authorized Access PointMore likely, a NAR for you does not already exist – and to practice with the full functionality of the Editor, we will now follow this scenario. Now you are on your own. You create the description of the Agent using the attributes that you would record in an authority template in Voyager. Remember the resource from Unit 2 – Linked data for cultural heritage? One of the editors was Michele Seikel. Let’s imagine that an authority record does not exist for this Agent. Here is some information you can use in order to create the authority record:Michele Seikel?is a tenured professor on the library faculty at Oklahoma State University. She has held positions at Norman Public Library, the University of Oklahoma, and Stanford University, and served as a professional librarian at Oklahoma Panhandle State University and at Oklahoma State University. Her primary professional focus is in cataloging, and she has published several research papers in technical services journals. In the ALA, she has cochaired the Cataloging Norms Interest Group and the Cataloging and Metadata Management Section’s Policy and Planning Committee. Currently, she chairs the ALCTS Planning Committee, and is a member of the editorial board of the journal?Library Resources and Technical Services.Michele Seikel?is a tenured professor on the library faculty at Oklahoma State University. She has held positions at Norman Public Library, the University of Oklahoma, and Stanford University, and served as a professional librarian at Oklahoma Panhandle State University and at Oklahoma State University. Her primary professional focus is in cataloging, and she has published several research papers in technical services journals. In the ALA, she has cochaired the Cataloging Norms Interest Group and the Cataloging and Metadata Management Section’s Policy and Planning Committee. Currently, she chairs the ALCTS Planning Committee, and is a member of the editorial board of the journal?Library Resources and Technical Services.Authority Work for Subjects“Traditional” Method Scroll back up to the “Work” area of the template; near the bottom of those elements is the label “Subject of the Work”.Sometimes you know that an established heading exists for a LCSH heading + subdivision combination, but sometimes, because of LCSH’s free-floating subdivision policies, an established heading does not exist. We need to consider both scenarios when working in BIBFRAME. Although you can do some name authority work in BIBFRAME, as described above, for this pilot, all LCSH work will be done according to existing policies and practices. You will only record subject information for Works in BIBFRAME, you will not do any actual authority work. Let’s look at an example:You are cataloging a resource about the Library of Congress and its buildings. You want to record Library of Congress—Buildings as a subject. Search Library of Congress in the Lookup under SubjectNote that the lookup searches both LCSH and the Name Authority FileThere is not an established heading Library of Congress—Buildings, since –Buildings is a free-floating subdivision under corporate bodies according to the Subject Headings Manual (SHM) instruction sheet H 1105, Corporate BodiesBecause the heading + subdivision is not supported by an LCSH authority record, we need to input the heading + subdivision as a “literal”Optionally, you can record the source of the heading + subdivision by searching and selecting LCSH in the source lookup box:“Component” MethodAnother way to record an LCSH heading + subdivision combination is thought he componentList search. You can search each element individually with this method, although the lookups are only to LCSH, not to the LCNAF – so the search we did above, with the Library of Congress, would not work with this method. Administrative Metadata for InstancesJust as more experimentation is taking place with Agents in this second BIBFRAME Pilot, experimentation is also taking place with Administrative Metadata for Instances. Although we are moving away from MARC and try not to make too many comparisons between BIBFRAME and MARC, it might help you to think of BIBFRAME Administrative Metadata as the type of information you might record as a MARC 008 bibliographic element, a MARC Leader element, or even a MARC 040 field element. At this point, all of this is at the purely experimental stage, and changes will take place throughout the pilot. One nice thing about this is that many of the values are preset, according to the resource format and the cataloging institution. 1876425-889000Only some of the boxes need to be filled in! Note that many of them are pre-populated. Your cataloger ID (Windows ID)Creation date Change date (if you are updating an existing description)Encoding levelProjected publication date (for a CIP description)Profile – for now, this needs to be input manually. It is the profile you are working with Item InformationThis is another new area of experimentation. Item is the third BIBFRAME core level of abstraction – Work, Instance, Item.You are familiar with Items in RDA as well. Information about BIBFRAME Items will be recorded in the Item part of the Editor. Item information is linked from the Instance area of the Editor for each type of resource. 2675890174752000 Don’t worry about all the boxes in the Item view – this is a big area of development and experimentation at this point. The idea is that you become familiar with some of these elements, and think about how you might record them when describing a resource. The circled elements might be ones that you will use.Unit 5: Workflow for the LC Pilot Why is Workflow So Important?BIBFRAME Pilot participants must create bibliographic descriptions for the same resource using both BIBFRAME (using the BIBFRAME Editor) and MARC (using the Voyager ILS). Where Do I Start? In the BIBFRAME Editor Select a resource to be describedFirst, search for the BIBFRAME Work in one of these two ways:Directly in the Linked Data Store: you find the work, you can copy the URL and paste it into the BIBFRAME Editor, in the “Instance of” box. Using the title or creator name as a left-anchored search directly in the BIBFRAME Editor, in the “Instance of” box.If you find the Work this way, you can just click on the title and the URL will be saved in the “Instance of” box. *** Note: please refer back to Unit 3 – searching in the Linked Data Store will give you a much better picture of the cataloging situation. *** If you find a BIBFRAME Work and save it in one of the ways outlined above, you only need to complete the BIBFRAME Instance description in the EditorIf you do not find a BIBFRAME Work in the BIBFRAME database, you will need to create both the Work and the Instance in the BIBFRAME Editor. Once you have completed the description, Preview your work to make sure it is correct.Save your work – this saves your work to the “local” workspace.Publish your work – this migrates your work to the Linked Data Store. Then What? In Voyager Catalog the resource in the Voyager ILS, following current section practices and policies:Do any necessary name authority work in the Voyager ILSDo any necessary subject or classification-related authority work in MinaretAdd a 985 “Record History” field with subfield $c BibFramePilot2 and subfield $e [date]Use the MACRO! Save the record following normal routinesComplete end-stage processing, as requiredReporting Findings, Problems, Issues, Complaints …Have we said that we need your help?Although you will be interacting with the BIBFRAME Editor in some very complex ways, the primary purpose of this BIBFRAME Pilot is very simple: To determine whether the BIBFRAME Editor is a viable mechanism for describing resources You will discover many features in the BIBFRAME Editor that are confusing, do not work as you expected, could work in a much more efficient way, need to be added, or may not even be necessary. There are surely many other things you will encounter, as well. The purpose of a pilot is to test, experiment, and (ultimately) improve. It may take several pilots to realize the far-reaching benefits of describing library resources in BIBFRAME using the BIBFRAME Editor. So your role in this BIBFRAME Pilot is crucial, even as it may be frustrating at times. After You Leave the Classroom: What Should You Do?Continue to catalog your resources TWICE – first in BIBFRAME, and then in VoyagerContinue to make note of your experiences, both good and bad --- although the ultimate goal is improvement, it is always nice to hear about things that already work well!Share your findings with everyone else in the pilot by posting on the iCohere LMS site, or by sending a message to bibframepilot@Share your thoughts at the periodic feedback sessions that will take place during this phase of the PilotPlease!We WANT your feedback. Provide it using either the iCohere LMS – there is an Error Report form there, too – or bibframepilot@, or at the feedback sessions. Unit 6: Help, Support, and Other ResourcesiCohere BIBFRAME 2.0 Pilot PortalYou can post questions You can read about what other Pilot participants are experiencing You can submit Error ReportsYou can access documentationYou can check on upcoming BIBFRAME classes *** Think of it as a virtual classroom ***The BIBFRAME mail boxbibframepilot@ You can ask questionsYou can report errorsThe trainers access and respond to your messagesLC’s BIBFRAME Webpage SessionsThe trainers and staff in the Network Development and MARC Standards Office (NDMSO) will schedule periodic Feedback sessions. They may not have all the answers – they may not have many, early on – but they pledge to try and find out for you. We are all ‘learning on the job.Announcements about these Feedback Sessions will be posted to the iCohere site.Appendix A. RDF for BIBFRAMEThis article was authored by Ray Denenberg, Senior Network Engineer, Library of CongressNetwork Development and MARC Standards Office. We are including it in this appendix to help you become more comfortable with the terminology and concepts, and better able to interpret the ‘records’ – i.e., ‘graphs – you create during the BIBFRAME Pilot.* * *This paper is written for the benefit of catalogers who will need to create BIBFRAME descriptions. They do not need to be RDF experts but they do need to understand RDF to some degree. This is not an attempt to dumb-down RDF (that does not work well) but rather to avoid complexities that don’t seem necessary, as well as to try to offer an intuitive approach. ??Of course, intuitive, to some extent, goes hand in hand with dumbing-down and this paper becomes more abstract (and less intuitive) as you proceed through the sections, but the first section, about RDF triples, probably strikes the right balance. RDF TriplesRDF is about triples.? Think of a triple in terms of a basic grammatical sentence; it has three parts (1) subject, (2) verb, (3) object. ??For example, the sentence, “Pepper is a cat” has three parts (1) subject: “Pepper”, (2) verb “is a”, (3) object “cat”. We often refer to a triple as a “statement”.The “RD” in “RDF” is “Resource Description”.? Essentially, we start with a resource, and we describe it with statements.? A resource is anything identified by a URI. Notes:Virtually anything can have a URI, so virtually anything can be a resource. At this point, people usually ask, “what is the difference between a URI and URL.” For purposes of this discussion, there is no substantive difference whose explanation would be of any value, so assume they are the same.The resource description will be?a set of triples (Or sentences. Or statements.)Consider the novel, Gone with the Wind.? It’s a resource, in fact it is a BIBFRAME resource, specifically, a BIBFRAME Work. As such, it has a URI: . ?In constructing triples describing this resource, this URI is used as the subject for these triples. Here are some examples (this is in plain English, not RDF syntax): aBIBFRAME Work . aBIBFRAME Text . has titleGone with the Wind . creatorMitchell, Margaret, 1900-1949 . subject Plantation life—Fiction . subjectWomen--Georgia—Fiction . classificationPZ3.M69484 . classificationPS3525.I972 .Let’s examine these statements.? First, (1) and (2):? ?Both state a type for this resource. Any resource may have more than one type, and in this case it is a BIBFRAME Work and it is also a BIBFRAME Text (as opposed for example to a Moving Image).? These types are called “classes” in RDF, and BIBFRAME Text is a subclass of BIBFRAME Work (which means that any BIBFRAME Text is a BIBFRAME Work.? The terms Work and Text, class names in the BIBFRAME vocabulary, are referred to as bf:Work and bf:Text to denote that they are being used within the context of the BIBFRAME vocabulary, and not some other vocabulary that might also use these terms but with a different meaning. (And the prefix ‘bf:’?is bound to the BIBFRAME?vocabulary through an association statement within the RDF description, not shown.)? Also the verb expression “is a” will be abbreviate as just “a”; this is well-known within all of RDF to denote “is of type” or “is of class”.? Finally,?URIs within RDF triples are enclosed with angle brackets.So (1) and (2) are rewritten:< .< .For the remaining statements:? ?“has title”, “has creator”, “has subject” and “has classification” are simply “title,” “creator”, “subject” and “classification” in BIBFRAME. Hence bf:title, bf:creator, bf:subject, and bf:classification. ??So, verbs don’t always sound like verbs; in these four cases, the “has” part is implied.? When constructing?an RDF verb, you can make it sound like a verb (hasCreator) or not (creator); there are recommended practices governing this.?? Anyway, the verb part of an RDF triple is referred to as a “property”, “predicate” or “relationship”.? These three things have slightly different meanings; “property” and “relationship” are used for different types of objects (discussed later) and “predicate” is a generalization of those two, so we’ll use “predicate” when we mean “either a property or a relationship”? (And note the convention: class names begin with uppercase; predicates begin with lower case).One more thing: if we consider the set of these eight triples to be a description, note that all eight triples have the same subject.? That subject does not need to be repeated and instead there is a convenient way to denote that subject of the next triple is the same as that of the current triple:? You end the triple with a semi-colon (“;”) rather than with a period (“.”).? ??When you get to the last triple for the subject, end with a period. So the description, consisting of eight triples, is re-written as:< >abf:Work ;abf:Text ;bf:title“Gone with the wind” ;bf:creator “Mitchell, Margaret, 1900-1949” ;bf:subject “Plantation life—Fiction” ;bf:subject“Women--Georgia—Fiction” ;bf:classification“PZ3.M69484” ;bf:classification“PS3525.I972” .And we can go even a step further.? When consecutive triples not only have the same subject, but also have the same predicate, you don’t have to repeat the predicate, instead you can list the subject and predicate once and separate the objects with commas (“,”).? So we can rewrite the description as:< ,? bf:Text ;bf:title“Gone with the wind” ; bf:creator“Mitchell, Margaret, 1900-1949” ;bf:subject “Plantation life--Fiction” , “Women--Georgia—Fiction” ;bf:classification“PZ3.M69484” , “PS3525.I972” .Although we’ve reduced the description to five lines, these are still eight triples. RDF as a Graph Well, we said RDF is about triples in the previous section. But some experts feel that RDF is more about graphs. An RDF description can be expressed as a set of triples or as a graph. But whether you like graphs or triples better, the fact is that in order for a computer to understand the description it has to be decomposed (serialized) into triples. (Serialization is discussed in the next section) Describing it by a graph may be useful for visualization but it is never necessary. Let’s look at an example. Consider the following description: “Pepper is a female, gray cat, born in 1999.”This description is expressed by four triples:Pepper a Cat ;color gray ;sex Female ;yearBorn“1999” .It can also be expressed by the following graph:177101530480Cat00Cat112585511049000120967563500a 00a 279082531115Female00Female185737555880sex00sex120967522288500295275137160Pepper00Pepper12096751682750021431252540color00color2943225104775Gray00Gray998220228600014382750yearBorn00yearBorn161925076201999001999Different brains visualize things differently, some visualize graphs more easily than triples, and others visualize triples more easily. In either case, if you need to commit the above description to RDF, you are going to have to decompose it into triples. If you visualize graphs more easily than triples, create the graph first and then decompose it into triples. If you visualize triples more easily (or as easily) then there is no need to create a graph. Please note that there is some hand waving in the above example, in the creation of the graph/triples. For example, you can’t just make up predicates, like “color” or “yearBorn”. They have to come from an established RDF vocabulary. RDF SerializationsSo you’re given the description “Pepper is a female, gray cat, born in 1999.” (or the graph shown in the previous section that represents this description), and you might want to turn it into an RDF description. That means turning it into triples. In order for those triples to be consumable by a computer, the triples have to be represented in a commonly understood format that a computer can recognize and process. There are a number of these formats in use; we call them “RDF serializations”.The most straightforward serialization (though not the prettiest) is called N-Triples. It’s just raw triples – we’ll just call it “triples”.Here, each triple in the RDF description is written out in full –< .< .<“Gone with the Wind” .<“Mitchell, Margaret, 1900-1949” .< “Plantation life--Fiction” .<“Women--Georgia--Fiction” .<“PZ3.M69484” .<“PS3525.I972” .We also showed that the above can be reduced to the following: < ,? bf:Text ;bf:title“Gone with the wind” ;bf:creator?“Mitchell, Margaret, 1900-1949” ;bf:subject“Plantation life--Fiction” ; “Women--Georgia--Fiction” ;bf:classification“PZ3.M69484” , “PS3525.I972” .This serialization is called “Turtle” (Terse RDF Triple Language). The rules are simple; as we have shown by example: (1) when the subject is the same for two or more triples, it need not be repeated and instead the predicates/object pairs are listed separated by semicolon with a period at the end of the last triple. (2). When both the subject and predicate are the same for more than one triple, they need not be repeated and the objects are separated by commas.In addition, literals are enclosed in quotes while resources (discussed later) are enclosed in angle brackets. There is one additional feature that has not yet been illustrated. It deals with blank nodes, which we’ll cover later.But first, to complete the discussion on serialization we must mention two additional serializations in common use: RDF/XML, and JSON-LD (JSON for Linked Data).For this resource: , the following serializations are available:N-Triples: RDF/XML: : If you have trouble making sense out of any or all three of these, don’t worry, we aren’t going to discuss these any further. We’ll limit discussion of serialization to turtle.Properties and Relationships: Literals vs. ResourcesConsider the following triple from the “Gone with the Wind” example. <“Mitchell, Margaret, 1900-1949” .This triple expresses that the creator of the subject resource is “Mitchell, Margaret, 1900-1949”, which is the object of the triple.But the object could have been expressed differently:< >??bf:creator<; .Where is the web page for the LC NAF record for Margaret Mitchell, and is available in RDF (). It provides much more information about the object than just the literal. Since it is available in RDF, it is referred to as an “RDF resource”. Note: the formal definition of “RDF resource” is not established, but it is often used as just described and we assume this is its meaning, for purposes of this paper.So, the object of a triple is either a literal or an RDF resource, as in the above two triples. It is important to note, the subject of a triple is always an RDF resource while the object is a literal or RDF resource. Furthermore, the subject must be represented by a URI. If the object is an RDF resource, it may be represented by a URI or it may be a blank node (discussed later).We mentioned above that the verb in an RDF triple is referred to as a “predicate”, “property” or “relationship”, where the latter two have slightly different meanings and “predicate” is a generalization of those two. Well, “property” is used when the object is a literal. “Relationship” is used when the object is an RDF resource (because it states a relationship between two RDF resource).Blank NodesA blank node is used as the object of a triple when that object is structured information, rather than either a single piece of information represented as a literal, or an RDF resource represented as a URI.First we’ll try an intuitive approach to blank nodes, and then a more formal example.We said that RDF is about triples: subject, predicate, object. Well, the object itself can be a bunch of triples. In the triple, “Pepper is a cat”, the object, “cat” is a single piece of information. But an object may be several pieces of information, each representing a triple. Suppose I want to say “Pepper has sibling pet Shadow, a dark gray cat born in 2000”, then the subject would be “Pepper”, the predicate “has sibling pet” and the object “Shadow, a dark gray cat born in 2000”. This object is itself a set of triples. The entire statement can be represented in triples as:PepperhasSiblingPetxyz .xyzaCat ; nameshadow ;color darkGray ;sex“Male” ;yearBorn2000 .in this example, “xyz” is called a blank node because its name is irrelevant. Its purpose is to serve as the object of one triple and the subject of a set of triples, and it serves to connect the first triple to the remaining triples. it could have been called anything; for example you could change “xyz” to “abc” (in both places above) and it would have exactly the same meaning.Now a more formal example… For the triple in the previous section:< >??bf:creator URI representing the object is an RDF resource, represented by a URI. What that means is, if you follow that link, you’ll get an RDF description of the object. In this case, the URI represents the person, Margaret Mitchell, and the RDF description you will get is at (or one of the other serializations). That’s more information than just the string “Mitchell, Margaret, 1900-1949”; it included structured information parsing out her first and last names and dates of birth and death.Before we discuss blank nodes, since the above example is probably too complicated to fully appreciate – let’s construct a simpler, fictitious example, just for the sake of illustration. Let’s say is an RDF description of Margaret Mitchell, including her first and last name, dates of birth and death, and places of birth and death. Thus,<; aPerson ;firstName“Margaret” ;lastName“Mitchell” ;yearBorn“1900” ;yearDied“1949” ;placeOfBirth“Atlanta” ;placeOfDeath“Atlanta” .Further assume that is an RDF description of the book, Gone with the Wind. Thus:< in the previous section, here we have an example of a triple where the object is an RDF Resource, represented by a URI. This is the linked data approach, where all of the information conveyed by that resource is external and you get to it by following a link.In contrast, the object could instead be expressed with the data internal to the description without a URI – it’s right there, with no link to follow. Let’s illustrate this with a graph …2352675-114300Book00Book3436620106680“Gone with the Wind”00“Gone with the Wind”181356015240 a 00 a 162496517780002350135103505title00title-64008099060 00 4838065123190…….00…….185737513970 a 00 a 43815009969500104394024765001318260114935004232275114935002857586360Person00Person323850031115003600450245110yearBorn00yearBorn49377607620 etc. born00 etc. born148145571755firstName00firstName2537460122555lastName00lastName451231093980“1900”00“1900”253746068580“Mitchell”00“Mitchell”32766015240“Margaret”00“Margaret”In this graph, is declared to be a Book, its title is “Gone with the Wind” and it has a creator expressed by a predicate whose object, represented by the blank circle -- hence “blank node” -- in blue. Remember, we are going to use Turtle serialization whenever possible, but it is useful to understand that if we were to use N-Triples instead of Turtle to represent a blank node, that node would be given an identifier (as “xyz” in the informal example above) – not a URI but a simple node identifier. That node, which is the object for arrows pointing to it, becomes the subject for the outward arrows, and the purpose of the node identifier is used to identify the subject of these new triples. But enough said about N-Triples. By serializing with Turtle we avoid that whole mess with node identifiers. With N-Triples, to serialize the blank node, we now introduce the last rule: Blank nodes are enclosed with square brackets.So the above graph is serialized as: aBook ;title “gone with the wind” ;creator[aPerson ;firstName“Margaret” ;lastName“Mitchell” ;yearBorn“1900” - ] . ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download