Vendor Guide for PromptCat, OCLC Vendor Record Contribution Program & CIP Record Upgrade Project

1. General Description of Services

With the introduction of OCLC PromptCat, OCLC Vendor Record Contribution Program, and the CIP Record Upgrade Project, OCLC is able to meet the cataloging and acquisition needs of our member libraries. By OCLC establishing successful working relationships with outside vendors, our libraries are able to continue to work with their preferred materials vendor, while at the same time, consolidate their cataloging processes with OCLC.

The Internet is the major communication method that vendors and OCLC will use to communicate activities for all of these services. Vendors will send files to OCLC via OCLC's Electronic Data Exchange (EDX) Service, which uses Internet FTP. Vendors must have Internet access and FTP capability in order to participate in PromptCat, OCLC Vendor Record Contribution Program, and/or the CIP Record Upgrade Project. Internet access must be acquired independently by the vendor prior to beginning participation. The details about sending files to OCLC are explained in Chapter 8.

Participation in any or all of these services requires OCLC and the vendor to complete a number of tasks from evaluating files to signing appropriate legal agreements. These tasks may take anywhere from two to six months for completion.

1.1 PromptCat

PromptCat is a vendor-prompted copy cataloging service for non-serial titles. It links order-fulfillment by a vendor to OCLC, setting a library's holding symbol in WorldCat, the OCLC Online Union Catalog, and delivering a monographic OCLC-MARC record to the library. Libraries have the option to receive materials 'shelf-ready' from the vendor. These materials arrive from the vendor physically processed and ready to be shelved. Based on local decisions, materials may arrive with barcodes, security strips, book jackets, spine labels, cards, pockets, date due slips, property stamps, bookplates, etc.

PromptCat works from a relation file, called the Vendor Inventory Database (VID), consisting of vendor titles linked to monograph records in WorldCat. The link is the vendor's unique control number for each title and OCLC's unique bibliographic record control number. The VID is built by matching the vendor's title file to WorldCat. The vendor sends ongoing updates (at least weekly) via OCLC's EDX service for matching. OCLC does not keep the vendor record, only the vendor's unique control number. These ongoing updates continue for as long as the vendor participates in PromptCat.

The vendor sends a manifest, a formatted invoice, to OCLC when the materials are shipped to a library. Typically, for approval plans, this is weekly. For firm orders, it could be daily. The manifests are processed daily at OCLC, using the library's PromptCat profile, to create files of records according to the library's specifications. For every file of records, OCLC may edit the records, add data to the records, delete data from the records, move data within the records, de-select some matching records from the file and deliver the records in one of four mediums (EDX, OCLC PromptCat Online File, tape, and cards). Since every library's profile may differ, OCLC may process each file differently.

Manifest processing results in a PromptCat cataloging report sent to the library, listing titles for which an OCLC-MARC record is supplied. It also lists titles for which no record exists in WorldCat, or for which no record was delivered because a holding was already set, or for which the matching record did not meet the library's selected options.

To participate in PromptCat, a vendor MUST:

  • have Internet access
  • have a valid standard address number (SAN). The Standard Address Number (SAN) is a unique identifier for an organization involved in the book industry. SANs facilitate the communication of data surrounding activities such as purchasing, shipping, and billing. They are assigned by R.R. Bowker. To apply for a SAN, contact Diana Fumando at R.R. Bowker at +1-908-771-7755.
  • be able to output bibliographic records in either USMARC or EDIFACT format
  • have a unique control number for each title (vendor control number)
  • be able to send customer manifests in EDIFACT format.

1.2 The OCLC Vendor Record Contribution Program

Founded in 1967, OCLC began with the idea of combining computer technology with library cooperation to reduce the costs and improve services through online, shared cataloging. That idea of using technology to extend cooperation has grown into a worldwide organization in which almost 36,000 libraries from 74 countries use OCLC products and services.

OCLC's cataloging service is now the biggest and the busiest in the world, enabling libraries each year to catalog more than 42 million items. WorldCat (the OCLC Online Union Catalog) is the world's largest and most comprehensive database of bibliographic information. Containing the merged catalogs of libraries around the world, it makes available to libraries and their users, resources no single library could possess. WorldCat is utilized in many different ways, ranging from cataloging of bibliographic information, searching for interlibrary loan material, to identifying new titles to order to add to a library's collection.

OCLC also pursues working relationships with library material vendors for the contribution of bibliographic records into WorldCat. In particular, we are striving to enrich our database with records pertaining to non-U.S. and audio-visual materials to name a few.

Vendor records must pass through a thorough evaluation by our database specialists. Vendor records are loaded into WorldCat IF:

  • records do not match existing records after processing;
  • fewer than 2% of the records contain machine-detectable errors in MARC coding;
  • fewer than 20% of the records are duplicate records in WorldCat;
  • records meet OCLC K-level standards for completeness; and
  • records are in USMARC format.

OCLC loads vendor records through an offline batchload process. When a vendor record matches an existing WorldCat record, OCLC adds a vendor-specific 938 field to the master record. When no single unique match is found, OCLC adds the vendor record as a new record with a 938 field. In either case, OCLC sets the vendor's OCLC symbol.

The 938 field contains vendor-specific ordering data, defined as follows:

938 Vendor Specific Ordering Data
  Both indicators are blank.
Subfield Codes
‡ a Full name of vendor. Supplied by OCLC.
‡ b OCLC-defined identifier for vendor. Supplied by OCLC. These identifiers are used for indexing purposes and should not be confused with OCLC-3 character institution symbols. Identifiers currently defined for this field are:
Casalini Libri: CASA Iberbook International: IBER Touzot: TOUZ
Harrassowitz: HARR Puvill Libros: PUVL  
‡ c Terms of availability. Purchase price.
‡ n Vendor control number. A unique number assigned by the vendor to identify the item for offline ordering.
‡ s Vendor status. Indicates if item is backordered, out-of-print, not yet published, etc.
‡ z Note. General notes field in which a vendor can include order information specific to the item.

To facilitate searching for vendor records in WorldCat, OCLC indexes 938 ‡a and ‡b.

To receive more information about the OCLC Vendor Record Contribution Program, complete and submit the Vendor Interest Form.

1.3 CIP Record Upgrade Project

The purpose of the Cataloging in Publication (CIP) Program is to enhance/update pre-publication cataloging records for those books most likely to be widely acquired by the nation's libraries. This CIP data is printed in the book and greatly facilitates cataloging activities for libraries. The data are also distributed prior to the book's publication in machine-readable form via the MARC (Machine-Readable Cataloging) tapes, alerting libraries and other bibliographic services around the world to forthcoming titles.

OCLC member libraries use CIP records in WorldCat for ordering, and upgrade the records later when they have the final printed piece in hand to provide physical descriptions as well as new or corrected cataloging information. By accelerating the upgrading of CIP records in WorldCat, OCLC hopes to provide libraries with full cataloging records as soon as possible after publication. In conjunction with this project, libraries may also take advantage of the OCLC Bibliographic Record Notification service. This service automatically delivers upgraded bibliographic records to libraries.

OCLC works closely with vendors in the CIP Record Upgrade Project. Trained cataloging staff at the vendor site upgrade CIP records before books are shipped to libraries.

Vendors will upgrade CIP records to full-MARC records as newly published materials arrive, then send the records to OCLC for inclusion in WorldCat. The records then will be available to users of OCLC cataloging and reference services, including PromptCat. For additional information on the CIP Record Upgrade Project, see Chapter 2

2 Vendor Bibliographic Records

In order to create the relation file for PromptCat or load records for OCLC Vendor Record Contribution Program or the CIP Record Upgrade Project, you must send bibliographic records of the titles you make available to your libraries. The file format is negotiated and agreed upon with your company in the legal agreement signed prior to your participation. These specifications are included in each appropriate agreement, but also are contained in this Vendor Notebook as Exhibits A through E.

The types of files you send depend upon the projects/products in which you are participating. The table below outlines the types of files that are sent for each project/product. Bibliographic files are discussed in this chapter; manifest files are discussed in chapter 7.

PRODUCT SUBMITTALS FORMAT SPECIFICATIONS
PromptCat Bibliographic USMARC Should meet the minimum record requirements outlined in the OCLC Batchloading Guide, 3rd ed., Appendix B, Section 5 . (See Exhibit A)
  Bibliographic EDIFACT OCLC supports EDIFACT 96A version of the PRICAT document. (See Exhibit B)
  Manifest EDIFACT OCLC supports EDIFACT 96A version of the QUOTES document. (See Exhibit C)
OCLC Vendor Record Contribution Program Bibliographic USMARC Should meet the minimum record requirements outlined in the OCLC Batchloading Guide, 3rd ed., Appendix B, Section 5. (See Exhibit A)
CIP Record Upgrade Project Bibliographic USMARC OCLC's CIP Record Upgrade Project Specifications (See Exhibit D)

Contact OCLC to confirm that you are working with most current version of any specifications you have on file from OCLC.

2.1 PromptCat

PromptCat vendors first send a sample file of titles for evaluation purposes. An evaluation is issued by OCLC and vendors may or may not need to make changes to their file. Once the sample file meets established specifications and an acceptable match rate has been reached, PromptCat vendors may begin sending live data for processing and then ongoing weekly updates of new titles. The ongoing weekly updates must be sent via EDX. Records must be in USMARC or EDIFACT. Record requirements for USMARC files appear in chapter 5.

Depending on the size of the database, loading the vendor's file may take several weeks. The vendor cannot begin to send manifests and the user cannot receive PromptCat records until OCLC has finished matching the file and is successfully loading the weekly updates.

Since the updates contain the newest titles, and the subsequent week's approval plan shipments will contain the newest titles as well, OCLC must load the vendor's weekly updates before processing manifests that week. OCLC is committed to a 48-hour turnaround time for processing updates, and a 48-hour turnaround time for manifests.

2.2 OCLC Vendor Record Contribution Program

Records furnished by vendors for participation in OCLC Vendor Record Contribution Program must be in USMARC format. For format purposes, records must meet the minimal requirements for bibliographic records as described in the OCLC Batchloading Guide (See Exhibit A).

Vendors begin participation in OCLC Vendor Record Contribution Program by sending a sample file of titles for evaluation purposes. An evaluation is issued by OCLC and vendors may or may not need to make changes to their file. Once the sample file meets established specifications, vendors begin participation in OCLC Vendor Record Contribution Program by sending, via tape or EDX, their entire backfile of bibliographic records. After this initial backfile, vendors are asked to send weekly updates of records via EDX. Initial processing of the backfile generally takes several weeks to a month. Records are processed utilizing several matching algorithms described below.

If a record furnished by a vendor is unique and matches specific criteria to be loaded, OCLC will load such records into WorldCat. A 938 field is created for each record. This field will contain ordering information -- vendor name and symbol, purchase price, availability, etc. The vendor's OCLC-assigned holding symbol also is attached to each bibliographic record.

If a record furnished matches an existing record in WorldCat, OCLC will use standard record merging procedures and will add unique information contained in the vendor record to the existing WorldCat record, including the 938 field and holding symbol.

2.3 CIP RECORD UPGRADE PROJECT

Third party suppliers will upgrade CIP records to full-MARC records as newly published materials arrive, then send the records to OCLC for inclusion in WorldCat. The records then will be available to users of OCLC cataloging and reference services, including PromptCat.

Vendors will apply changes to CIP records as outlined by OCLC in OCLC's CIP Record Upgrade Project Specifications (See Exhibit D). Vendors also must pass OCLC's quality review process for National Level Enhance CIP record upgrade work. Updated records will be sent to OCLC via EDX using Internet FTP. Records will be upgraded and submitted on a weekly basis.

3 Match Processing

The batchloading program that matches your records with WorldCat records uses a series of sophisticated matching algorithms. In general, the algorithmic sequence can be briefly described as follows.

unique-key matching algorithm: attempts to match based upon unique numbers for the item, such as ISBN and LCCN. Records that do not find a single exact match are passed through the extended matching algorithm.

extended matching algorithm: compares selected elements of the records and determines a match based upon whether particular character strings match. If a record cannot be matched through extended matching, it is later processed through DDR.

DDR: is an abbreviation for Duplicate Detection and Resolution. DDR also compares selected elements of the records, but determines a match based upon a pre-determined ratio of likelihood of a match. That is, if the match ratio is set at 75% (required match rate for PromptCat), then only those records that DDR determines as 75% or higher likelihood of a match are actually determined to be matches.

Depending on the final match rate, the algorithmic sequence may be altered to achieve a higher match rate. For example, a vendor file may be processed first through DDR and then through unique-key with certain qualifiers turned on or off.

The end result of match processing depends upon the project/product in which you are participating:

  • For PromptCat-related files it is the generation of links between vendor control numbers and OCLC record numbers in the Vendor Inventory Database. PromptCat vendor records for which there is no match will be stored in a re-search file to be matched against WorldCat periodically. After 180 days, those records which do not match will be deleted. OCLC provides to you a quarterly report of matching statistics. The report will include the number of records read, matched, and not matched for the quarter. It will also include the number of defective records and the delivery rates to your libraries.
  • For OCLC Vendor Record Contribution Program files it is the loading of unique records or matching of pre-existing records in WorldCat. If unique records are loaded, the vendor's OCLC symbol is placed in the 040 subfield c, and a 938 field is created detailing ordering information. As a final step, the vendor's OCLC holding symbol is added to the bibliographic record. If a vendor record is matched to an existing WorldCat record, a 938 field is created and the vendor holding symbol is attached. The vendor's OCLC symbol also is added to the 040 subfield d. OCLC provides to you a quarterly report of matching statistics. The report will include the number of records read, matched, not matched, and records added to WorldCat for the quarter.
  • For CIP Record Upgrade Project files it is the loading of these upgraded records into WorldCat. Records are matched to existing WorldCat records on LCCN plus title. Incoming records have an Encoding Level of "blank" and the current WorldCat record will be an Encoding Level 8. If there is a one-to-one LCCN match and the WorldCat record is Encoding Level 8, the WorldCat record is replaced. Also, record status changes to "c" for corrected record, institution symbol of the Third Party Supplier is added to 040 ‡d, and the "replaced" date is updated to show the date of the replace. The amount of bibliographic data in a vendor's record, its form and accuracy, and content designation affect the ease of matching and the overall match rate. OCLC is willing to discuss each vendor's capabilities and to work with vendors to achieve the most matches possible.

4 Vendor Control Numbers -- PromptCat

The Vendor Control Number (VCN) is a unique control number that you assign to each publication. It is the most important PromptCat data element that you supply to OCLC. When OCLC matches one of your records to a WorldCat record, a link is created between the OCLC record number and your VCN. This link is the basis for all future matching of that particular publication when sent to OCLC in your customer manifests. If you do not currently assign a unique identifier to each publication in your inventory database, OCLC requires that you do so for your participation in PromptCat.

If it is impossible to enhance your database with this information, you may consider using the ISBN as the VCN. However, there are disadvantages to this option:

  • You may track items in your database that have not been assigned ISBNs.
  • You must apply the ISBN for set to a record describing a set and an ISBN for a single volume when describing a single volume within a multi-volume set.
  • Although intended to be unique, publishers may assign ISBNs to more than one item.

If your system does not lend itself to the maintenance of a vendor control number, please contact OCLC. We will be glad to work with you to determine how best to identify your materials

5 USMARC Record Requirements

If you are supplying USMARC records (which is mandatory for OCLC Vendor Record Contribution Program and the CIP Record Upgrade Project), records and fields must conform to the standards specified in OCLC-MARC Tape and Export Record Formats, USMARC Format for Bibliographic Data.

The logical record length cannot exceed 6,144 bytes.

  • Each bibliographic record must have:

    1.1. A 245 field (Title)

    1.2. Valid codes in the Leader for:

    1.2.1. Record status (Leader byte 5)

    1.2.2. Record type (Leader byte 6)

    1.2.3. Bibliographic level (Leader byte 7)

    Uppercase letters are invalid in these Leader elements.
  • A 008 field (fixed field) containing exactly 41 characters, including the end-of-field (EOF) character.

In addition, each record that you submit for unique-key matching only must have at least one of the following for matching to occur:

  • An OCLC control number
  • An LCCN
  • An ISBN
  • An ISSN

The following tables contain a detailed description of minimal-MARC record requirements. These requirements are necessary for batchloading to accurately match records. The requirements include the following:

  • Information that must be present and valid in the minimal-MARC record for the record to be processed
  • Information used during unique key matching (both optional and required)
  • Information used during extended key algorithm (both optional and required)

You do not need to supply an OCLC number, LCCN, ISBN or ISSN. However, unique keys provide for efficient matching. Additional information may be found in Exhibit F; Considerations for Vendors Sending USMARC Data.

MINIMAL MARC REQUIREMENTS

Field/Subfield Matching Algorithm Position of Bytes Format Used Comments
Leader:* U/E 0-23 all 24 characters of data
Logical Record Length* U/E 0-4 all  
Record Status* U/E 5 all Although a valid value needs to be present, it doesn't have to be accurate if you cancel your symbol from all the records on the tape with a cancel-holdings project
Type of Record* U/E 6 all  
Bib Level* U/E 7 all  
Blanks** U/E 8-9 all  
Indicator Count** U/E 10 all Contains the character 2
Subfield Code Count** U/E 11 all Contains the character 2
Base Address of Data* U/E 12-16 all  
Encoding Level** U/E 17 all Although a valid value needs to be present, it doesn't have to be accurate
Descriptive Cataloging Form* U/E 18 all  
Linked Record Code** U/E 19 all Contains a blank
Length of Length-of-Field** U/E 20 all Contains the character 4
Length of Starting-Character Position** U/E 21 all Contains the character 5
Length of the Implementation-Defined Position** U/E 22 all Contains the character 0
Undefined** U/E 23 all Contains the character 0
Directory* U/E   all  
001* U/E   all Source system unique control #
008* U/E 0-40 all 41 bytes total: 40 characters in positions in 0-39 plus end-of-field in position 40. You can insert blanks in bytes that are not listed in this table if accurate information is not available. For information requirements refer to OCLC-MARC Tape and Export Record Formats
  E 6 all  
  E 7-10 all  
  U/E 23 bks, sco  
  E 33 med  
010 LCCN ‡a U/E all    
020 ISBN ‡a U all    
028 Music Publisher Number E   sco, rec  
035 Local System Number U   all When an OCLC number is available, enter the OCLC number in field 035, subfield ‡a, preceded by ( OCoLC), including the parentheses. The OCLC number is 1 to 8 digits. Leading zeros are optional. Both indicators are blank. All other 035 fields are ignored.
1xx or 1st 7xx E   all For all formats except serials, only the 1xx or the first 7xx field is used. For serial records, all 1xx and 7xx fields are used for matching

245**

‡a Title Proper

‡b Remainder of Title ‡n Number of Part/Section of a work

‡p Name of Part/Section of a work

U/E   all If a tape is submitted for unique-key matching only, the 245 field can contain any combination of alphanumeric characters (for example, Title) in subfield ‡aFor the records to be processed using the extended matching algorithm, the information in the 245 field must accurately reflect the title of the work. For serial records subfields ‡n, ‡p, and ‡c are also used in extended matching

260

‡a Place of publication

‡b Name of publisher

E   all  
255 Scale E   map  
300 Physical Description E   bks, med, map, sco, rec  

U=Unique-Key

E=Extended

U/E=Both

*Required; must be valid

**Required;need not be valid

6 EDIFACT Bibliographic Record Requirements (PromptCat only)

If you are supplying bibliographic data in the EDIFACT format, data must conform to the following standards and specifications:

  • A valid SAN is required. The Standard Address Number (SAN) is a unique identifier for an organization involved in the book industry. SANs facilitate the communication of data surrounding activities such as purchasing, shipping, and billing. They are assigned by R.R. Bowker. To apply for a SAN, contact R.R. Bowker at +1-908-771-7755.
  • Each record must contain a unique control number. If a vendor control number is present, this will be used as the unique control number. If no vendor control number is present, the ISBN will be used as the unique control number. If neither a vendor control number nor an ISBN is present, the publication number will be used as the unique control number. If none of the above are present, an error will be received and the record will not be processed.
  • All records will go through LCCN validation.
  • All records will go through ISBN validation.
  • Title is a required field for all records.

7 PromptCat Customer Manifest Files

The manifest is the vehicle the vendor will use to notify OCLC which records to send to the PromptCat subscribers. The manifest will contain the titles sent to a library, the vendor control numbers of those titles, some library information, and vendor-supplied data to be inserted by OCLC into the OCLC-MARC record. As appropriate, the manifest also may contain information to be used to generate a label file. Vendors may send information to be included in barcodes, enumeration, location, sublocation, and notes for each title. The format for the manifest is EDIFACT.

Since the weekly update contains the newest titles, and the subsequent week's approval plan shipments will contain the newest titles as well, OCLC must load the vendor's weekly updates before processing manifests that week. OCLC is committed to a 48-hour turnaround time for processing weekly updates, and a 48-hour turnaround time for manifests.

Vendors prompt OCLC to activate batch cataloging procedures by sending an EDIFACT formatted invoice of all items that have been shipped to the user. OCLC uses the data in the invoice to set the library's holdings and output an OCLC-MARC record for each title according to the library's profile. The manifest will contain all information the library wishes to have added into their delivered cataloging record from OCLC. The library and vendor agrees upon the details of information to be transmitted and added to the cataloging records.

When a library is profiled to receive shelf-ready information, OCLC also will provide a label file to the library, or to the vendor, depending upon how the library is profiled. This label file is used to generate the specific spine and pocket label options requested by the library. The label file may be sent to the vendor so they can do the physical processing requested for each title delivered to the library. See Exhibit G for further details on shelf-ready (processing options) data elements. Optionally, the library may choose to receive the label file to complete the physical processing.

A cataloging report for the user will be available electronically with each shipment of records. It lists the titles for which an OCLC-MARC record is supplied, titles for which there is no record, and titles for which no record was delivered because holdings were already present or for which the matching record did not meet the library's selected options

7.1 OCLC's Validation of Manifest Data

PromptCat validates certain data elements of every manifest processed. This table presents the validation points and the consequences of an error. Any errors in the following areas are written to a report that is forwarded to OCLC development and product support staff. The effect on processing of the manifest depends upon the type of error encountered. While OCLC development staff currently contacts the vendor directly whenever an error is detected within a manifest, this process will eventually be automated.

VALIDATION POINT IF ERROR DETECTED:
The vendor SAN is unknown to OCLC. OCLC does not maintain a list of all SANs; we only maintain a list of SANs of our trading partners. Halts processing of the entire manifest.
The Vendor SAN/Vendor Customer ID combination is not profiled in OCLC's PromptCat profiling system.

OCLC maintains a profile for each unique vendor SAN/vendor customer id combination. This is based on the order form the library completes when the decide to use PromptCat. The customer id the vendor sends in the manifest must exactly match the customer id that is in the PromptCat profiling system.

Prevents processing of the particular invoice.
The Vendor SAN/Transmission Set Control Number (message number, if EDIFACT) combination has already been sent. Prevents processing of the particular invoice.
A line item does not have a matching control number. The processing of this line item depends on other EDI segments sent in the manifest. The line item may not be processed, or it may erroneously be processed with the previous control number.
The control number/line item number is duplicated The particular line item is not processed.
The Vendor SAN/Vendor Customer ID/Site ID is invalid for the particular institution. OCLC also maintains a profile for each unique vendor SAN/vendor customer id/site id combination based on the order form the library completes when the decide to use PromptCat. The customer id/site ids the vendor sends in the manifest must exactly match the customer id/site ids that are in the PromptCat profiling system.

If a vendor does not send a site id for a line item, the system uses the customer id as both the customer id and site id. For some vendors and customers this is a valid situation.

The particular line item is not processed.
A line item does not have a title. The particular line item is not processed.
A line item does not have a VCN. Not only must the VCN be transmitted, but also it must exactly match the VCN that was included in the vendor's bibliographic record for the item. (see 4 –vendor control numbers for further information about this subject) The particular line item is not processed.
An EDI field code is invalid. The particular data element is not added to the record.

A manifest audit report is sent each time a vendor sends a manifest to OCLC. These reports are computer generated and are emailed back to the vendor, at an address specified by the vendor. If a vendor sends a manifest and does not receive this audit report, it is highly likely that OCLC either did not receive the manifest, or was unable to process it. If the vendor does not receive this audit report, the vendor should contact OCLC immediately. We will look into the problem and let the vendor know if the manifest should be resent.

7.2 PromptCat's Use of Processing Keys

The key to PromptCat processing resides with the use of Processing Keys. A processing key is the combination of a vendor SAN, a vendor customer id and a site id. Each processing key is linked to an OCLC four-character holding library code. The vendor SAN was described in Section 1.1. A vendor customer id is the code that identifies the library to the vendor. A site id is a code which is sent by the vendor for eat title.

The source of the site id may be one of four elements sent in the manifest: fund code, location code, subaccount code, or account code. The library, with the help of the vendor, identifies one or more site ids. There may be multiple site ids for each vendor customer id. Site ids are not case sensitive when used as part of the processing key. When they are used to construct the processing key, they are temporarily converted to upper case. If a site id, such as a fund code, also is added to the bibliographic record, it is added in whatever case in which it is sent in the manifest. All site ids, except location, may be mapped to the bibliographic record.

8 Vendor EDX: Submitting Files to OCLC via Internet FTP

1. Overview

As a participating vendor in PromptCat, OCLC Vendor Record Contribution Program, or CIP Record Upgrade Project, you will be using OCLC's Electronic Data Exchange service (EDX). You will be sending files via Internet ftp using format and submittal specifications provided by OCLC. The file format is negotiated and agreed upon with your company in the legal agreement signed prior to your participation. Each file type is sent separately from any other file type. An EDX Assessment Form is included as Exhibit H. The types of files you send depend upon the projects/products in which you are participating.

PRODUCT SUBMITTALS FORMAT SPECIFICATIONS
PromptCat Bibliographic USMARC Should meet the minimum record requirements outlined in the OCLC Batchloading Guide, 3rd ed., Appendix B, Section 5. (See Exhibit A)
  Bibliographic EDIFACT OCLC supports EDIFACT 96A version of the PRICAT document (See Exhibit B)
  Manifest EDIFACT OCLC supports EDIFACT 96A version of the QUOTES document. (See Exhibit C)
OCLC Vendor Record Contribution Program Bibliographic USMARC Should meet the minimum record requirements outlined in the OCLC Batchloading Guide, 3rd ed., Appendix B, Section 5. (See Exhibit A)
CIP Batch Upgrade Bibliographic USMARC OCLC's CIP Record Upgrade Project Specifications. (See Exhibit D)

Contact OCLC to confirm that you are working with most current version of any specifications you have on file from OCLC.

Before you may send a file to OCLC, you must complete and return the OCLC EDX Assessment form. The information you provide in this form about your Internet ftp environment will help our support staff diagnose problems and offer help if you happen to have problems with a transfer. Complete an EDX Assessment form.

2. File Header Specifications

USMARC Bibliographic and CIP Upgrade data files must begin with a seven-element header in ASCII. The header must begin and end in angle brackets (< and >). Other than the first and last elements (which have angle brackets), each element of the header contains the following data in this order:

  • One or more optional spaces
  • A label in ALL CAPS
  • A colon
  • A single space
  • The value of the label
  • A semicolon

Format the header as in the following example. Begin it in column 1.

<OCLCSYM: FUG; INSTNAME: Falstaff Books; DATESENT: 951230; FORMAT: M; DATATYP: P; RECCNT: 147; RUNOPT: E;>

Label Meaning
OCLCSYM Your company's OCLC symbol
INSTNAME Your company's name
DATESENT The date that you are sending the file, in yymmdd format
FORMAT Always M, for MARC format
DATATYP Use P, for PromptCat and/or OCLC Vendor Record Contribution Program submittals

Use CIP, for Upgrade submittals

RECCNT The number of logical records not including the header
RUNOPT The run-option code that OCLC assigns to you

These headers are not required for EDIFACT files.

3. Submitting Sample Files

You should notify the OCLC Vendor Liaison at OCLC when you are ready to submit a sample file. The only prerequisite to sending a sample file is completion and return of the EDX Assessment form.

OCLC does not support anonymous ftp. Sample files are transferred through a special OCLC account, the password for which is changed after each session. When you notify the OCLC Vendor Liaison that you are ready to submit a sample file, the Liaison will pass all necessary information to OCLC User & Network Support (UNS). UNS staff will then contact you to perform the transfer under the special account.

Once you have successfully transferred the sample file, the Liaison will confirm with you that the file has been received and is being evaluated.

Results of the evaluation of a sample bibliographic file are sent to you in a Batchload Status Report Memorandum via email. Results of an evaluation for an Upgrade project are sent to you in a CIP Enhance Evaluation Memorandum via email.

Results of the evaluation of a sample customer manifest file are usually reported to you in a telephone call or via email.

If additional sample files are required, please contact the OCLC Vendor Liaison each time you are ready to submit a new sample file. The Liaison will notify UNS staff, who will, in turn, contact you to perform the transfer.

Sample bibliographic (USMARC) files should be sent to:

EDX.PCAT.xxx.TEST.FTP

Sample EDIFACT bibliographic files should be sent to:

EDX.PCAT.xxx.TEST.EDIBIB.FTP.DYYMMDD

Sample manifest data should be sent to:

EDX.PCAT.xxx.TEST.EDIMAN.FTP.DYYMMDD

(xxx = OCLC symbol)

4. OCLC EDX Account

Once your sample files have been evaluated as compatible with OCLC's load programs, OCLC will establish an EDX account for your company. This account is basically a special authorization to OCLC's IBM 3090 mainframe. You will be assigned an account name and given a default password to use for your first logon. You will be notified via e-mail or letter when your account is ready.

The ftp address to use for sending all EDX files is:

EDX.OCLC.ORG

In most cases, you will need to change to binary or image mode to ftp your data files. You must enclose the OCLC filename in single quotation marks when you specify it for the file transfer.

4.1 Password Changes

OCLC's system will force you to change your password the first time you log on and every 90 days after that. If you forget to change your password after 90 days, the system will inform you that your password has expired. You will then only be able to access your account after installing a new password.

Passwords are alphanumeric and between 4 and 8 characters long. They are not case-sensitive. To change your password, enter the following string at the password prompt:

<old password>/<new password>/<new password> <RETURN>

Only you will know your password. Should you forget your password, OCLC will have to reset your account. Please keep your password confidential. Any abuse of access privileges will result in your account being revoked.

OCLC's IBM security program may refuse to let you set certain character strings as passwords. It prefers a mix of alpha and numeric characters to a recognizable word. The program also tracks your passwords and will not permit a password to be re-used for about 8 years.

4.2 Hours of Support

You may send files Monday-Sunday, except for the following hours which OCLC reserves for scheduled maintenance:

Monday through Saturday, 2:00AM-3:00AM (U.S., EST); Sunday, 2:00AM-6:00AM (US, EST).

You may encounter problems with the transfer of a file such as being denied access by the system or experiencing more than 3 occurrences of interrupted transmission for the same file. When these or other problems occur, you may call OCLC User & Network Support directly at +614-764-6000 (ext. 3750). US vendors may call toll free (800) 848-5800. Hours of support are 7:00AM to 9:00PM (US, EST), Monday through Friday.

You should try to transfer a file at least 3 times, and should double-check your Internet link before calling OCLC to report a problem. If there is a problem in transferring the file, OCLC User & Network Support will check to verify that the OCLC Internet link is functioning properly and that the OCLC host computer is functioning properly. Other than confirming these components of the transfer process, OCLC cannot provide support for telecommunication trouble over the Internet.

5. Submitting Live Data

Before you can submit live data files, the following steps must be completed:

  • The formal agreement between OCLC and your company must be signed by both your company and OCLC.
  • You must have completed and returned the OCLC EDX Assessment form.
  • You must have submitted sample files that were evaluated as compatible with OCLC's load programs.
  • You must have received notification from OCLC that your EDX Account is active.

5.1 USMARC Bibliographic & CIP Batch Upgrade Files

OCLC processes all files submitted in these two categories at 2:00AM (US, EST) each day, 7 days a week. You may send multiple files on any given day. When the files are collected at 2:00AM (US, EST), multiple files of the same type will be concatenated into a single file for processing.

To maximize the customer manifest match rate, USMARC files must be submitted to OCLC at least 24 hours prior to the transmission of any manifests which contain orders for the items represented in the bibliographic files.

Remember that USMARC files must include the OCLC-defined 7-element ASCII header detailed in Section 2 of this document.

The file name structure you will use depends upon the type of bibliographic file you are sending. In both of the following file name structures, xxx represents the 3-character symbol OCLC assigns your company. The (+1) in the name string generates a unique dataset extension each time a file is sent during a given day.

For submittals of USMARC Bibliographic files for PromptCat and/or OCLC Vendor Record Contribution Program, put your files to the following file directory/name structure

EDX.PCAT.xxx.FTP(+1)

For files submitted for Batch Upgrade Projects, use:

EDX.UPGRADE.xxx.FTP(+1)

5.1.1 USMARC Submittal Notification

When you send USMARC files of bibliographic records through EDX, you may request an e-mail report on the initial pickup of your file by OCLC. Each morning at about 2:00AM (US, EST), OCLC runs a routine that scans all of the EDX-submitted files and routes them into their appropriate processing streams. This routine includes a file diagnostics and repair routine that corrects some of the more common problems (such as missing header information) and will extract irreparably flawed records before the file is passed on to the processing stream.

If you send multiple files on a single day, you will receive a summary report of all the files you sent.

If any of the files are corrected by our diagnostics routine, you will receive an additional e-mail message that presents a high-level description of the changes that were made to the file so that it could be processed. These corrected files are reported as 'warning' files and contain the string WARN in the file name structure.

If any of the files you send contain errors that cannot be corrected by our diagnostics routine, you will receive an additional e-mail message that presents a high-level description of the error along with the number of records contained in the error file. The rejected records are written to a file that contains the string PASS in the file name structure. When you receive a report that a PASS file has been created, please log in to your EDX account and get the file so that you can correct the records and re-submit them with a later file.

If you are interested in receiving this notification report via e-mail, please send to the OCLC Vendor Liaison an e-mail message that includes your 3-character OCLC symbol, up to 3 contact names, and the e-mail addresses to use for the notification. Please note that OCLC will not re-distribute messages that do not make it through the Internet because of downed systems, connections, etc. If you are expecting to receive a notification message and do not, you will need to contact OCLC about the status of your file(s). Also, if you ever need to change any of the names on the notification list, contact the OCLC Vendor Liaison.

5.2 EDIFACT Bibliographic Files

For EDIFACT bibliographic files, put your files to the following file directory/name structure:

EDX.PCAT.xxx.EDIBIB.FTP(+1) or

EDX.PCAT.xxx.EDIBIB.FTP.DYYMMDD

5.3 Customer Manifest Files

Customer manifest files should be submitted to OCLC at least 24 hours after any bibliographic files that contain titles corresponding to the order. Currently, OCLC processes manifest files at 10:00PM (US, EST) each evening, 7 days a week.

You may send more than one customer manifest file during the day. However, multiple manifests are concatenated into a single file for processing.

To submit EDIFACT Customer Manifest files, put the file to:

EDX.PCAT.xxx.EDIMAN.FTP(+1)

6. Sample Session

The following instructions are intended as an introduction to Internet ftp with OCLC. They are based upon a simple login session from a UNIX environment to OCLC's IBM environment to submit bibliographic files for PromptCat or OCLC Vendor Record Contribution Program. Depending upon your local Internet environment, or special ftp software you may be using, your session may look very different. If you have trouble, please contact OCLC User & Network Support.

  1. Give the ftp command: ftp edx.oclc.org

    The OCLC system responds with the name prompt.

  2. Enter your account name.

    The OCLC system responds with the password prompt.

  3. You must change your password the first time you log on. Passwords are alphanumeric and between 4 and 8 characters long. To change your password, enter the following string at the password prompt. (NOTE: Keystrokes are hidden, so type carefully !!!)

    <default password>/<new password>/<new password><RETURN>

    The system responds with the message:

    Password has been changed.

    Txxx1 is logged on.

    ( where xxx is your OCLC symbol)
  4. Change to binary mode. In some systems, this is a binary command, in other systems it's referred to as an image command. You will have to know which command is required for the system you are using for ftp. As an example, the binary command might look like this:

    bin <RETURN>

  5. The filename structure in the IBM functions as a directory structure. User & Network Support staff recommend that you change directories using the cd command before you give the ftp put command. Enter the following command using your 3-character OCLC symbol. Be sure to enclose the directory structure in single quotes.

    CD 'edx.pcat.xxx'

  6. Give the put command for the data file you want to send. For example,

    put EDX.PCAT.xxx.FTP(+1) <RETURN>

    The first message you should expect to see displayed by the system is:

    Port Request OK

    The system then responds by telling you how long the transfer is taking and when transfer is complete.

  7. Once you have transferred the file, logoff the system:

    quit <RETURN>

    The system responds with the message:

    Quit command received. Goodbye.

9 Working With Libraries

Libraries that are interested in subscribing to PromptCat and/or OCLC Vendor Record Contribution Program should contact their OCLC-affiliated Regional Network. A listing of these Networks is attached as Exhibit H.

For PromptCat, OCLC together with the Network and vendor review the PromptCat Order and Options form for each PromptCat library. See the ordering forms. Once a profile is filled-out, it must be approved and signed off on by all parties: OCLC, vendor, Network and Library. Profiles are completed based on the following specifications:

Vendor-supplied data

Standard data description: Terms libraries should use to indicate which vendor-supplied data elements they want added to their PromptCat records. This is a dynamic list and represents a compilation of all the data elements offered by all vendors and where, in the EDIFACT manifest, they should be placed. None of the vendors offer all of the data elements listed. The PromptCat Order and Options form itself has a free text area for libraries to fill-in the vendor-supplied data elements they prefer. Libraries need to discuss this with their vendor(s) prior to completing the PromptCat Order and Options form.

Example Of a Quotes Message

Note that for clarity each segment is shown as starting a new line. In fact, there are no carriage returns or line feeds at the end of a segment. In the following example an ' is used as a segment terminator, a + as a DE separator, and a : as a component DE separator.

UNB+UNOC:2+123-4567:ZZ+322-2748:ZZ+970627:0856+1234' Interchange header
UNH+M0576+QUOTES:D:96A:UN' Header
BGM+31B+R0250' Type of message
DTM+137:19970626:102' Message date
RFF+IV:IN391987' Invoice number
DTM+3:19970623:102' Invoice date
NAD+DP' Mandatory segment
RFF+API:ABCD' Identifies ABCD as the customer id assigned by the vendor
LIN+1' Line item 1
PIA+5+02968580:VN+0789000393:IB' Vendor control number and ISBN
IMD+F+050+:::Marketing in Central and Eastern Europe' Title
QTY+1:1' Quantity
GIR+C01+C058195816:LAC+BUSI:LLO' Barcode and location
PRI+INF:29.95:SRP' List price
PRI+INF:23.96:NTP' Net price
RFF+BO:APP123456' Approval code
RFF+SLI:PO567667' po number
DTM+4:19970620:102' PO date
RFF+ADE:ACC82494' account number
ALC+A+++SAA' shipping and handling identifier
MOA+8:2.40' shipping and handling amount
TAX+7+GST' tax identifier
MOA+124:1.20' tax amount
LIN+2' Line item 2
PIA+5+030956646:VN+0881322458:IB' Vendor control number and ISBN
IMD+F+050+:::Cooperating with Europe's monetary union' Title
PRI+INF:11.95:SRP' List price
PRI+INF:11.95:NTP' Net price
RFF+BO:APP569890' Approval code
RFF+SLI:PO45738475' PO number
DTM+4:19970620:102' PO date
RFF+ADE:ACC82494' account number
ALC+A+++SAA' shipping and handling identifier
MOA+8:1.20' shipping and handling amount
TAX+7+GST' tax identifier
MOA+124:0.60' tax amount
LIN+1+++1:2' Sub line item 1 for line item 2
QTY+1:1' Quantity
GIR+C01+C047527586:LAC+BUSI:LLO' Barcode and location
LIN+2+++1:2' Sub line item 1 for line item 2
QTY+1:1' Quantity
GIR+C02+C087452896:LAC+ECON:LLO' Barcode and location
UNS+S' Summary section
CNT+2+2' Number of line items
UNT+44+M0576' Number of segments
UNZ+1+1234' Interchange trailer

Changes to a Library's Profile

We recently went through the process of making a change to one of our PromptCat library's profiles and thought it would be useful to review the process for submitting and approving PromptCat profile changes.

  1. When the library requires a change to their PromptCat profile, the library should request the change through their OCLC-affiliated Regional Network. Depending on the type of change, the library may be asked to complete a new order form.
  2. The network will forward the change to OCLC.
  3. OCLC will make the change to the user's active profile and set profile to pending.
  4. OCLC will send a copy of the new profile containing the change to the appropriate PromptCat vendors.
  5. The vendor should review/approve the change and respond to OCLC.
  6. After the vendor, library, and network have all approved the change, the change is activated at OCLC and the profile is reactivated.
  7. The vendor should then make the appropriate changes in their library profiles.

Exhibit A

Specifications for Bibliographic Data; PromptCat EDIFACT PRICAT Definitions

OCLC PRICAT Implementation

Exhibit B

Specifications for Manifest Data; PromptCat EDIFACT QUOTES Mapping Definitions

PromptCat Quotes Implementation

Exhibit C

CIP Record Upgrade Project Specifications

CIP Record Upgrade Project Specifications (PDF)

Exhibit D

Considerations For Vendors Sending USMARC Data

While OCLC staff realize your records may have been created for acquisition and collection development purposes rather than cataloging, we need information about your records to predict the success of processing them. The questions are intended to help you identify the information OCLC needs to plan for loading your records. The information will help us achieve the most and best quality matches to WorldCat.

The questions do not take the place of the evaluation of your records that will be performed by OCLC Tapeloading & Database Services staff. Also, the questions are not a substitute for the information found in the OCLC Batchloading Guide.

I. General Question

  1. Have your records changed in format or data over the years? Since we are loading all your records, we may need to select more than one matching method depending on how your practices have changed over time.

II. MARC Format Questions

  1. What are the sources of MARC records in your database?
  2. What numbers of records come from the different sources?
  3. What are the levels of cataloging? Full, minimal, brief? Describe the differences among these categories.
  4. Do all the records conform to USMARC? If not, which do and which don't. For example, we need to know if fields are subfielded correctly. (245 ‡a To the Lighthouse ‡b a novel ‡c by Virginia Woolf, not: 245 ‡a To the lighthouse a novel by Virginia Woolf)
  5. What MARC formats do you have in your database (serials, sound recordings, etc.)?
  6. If your database is not in MARC format, can you identify types of materials for outputting to separate MARC formats (e.g. scores, sound recordings, maps)?
  7. Can you identify types of materials for exclusion from the MARC output (e.g. pre-publication book information, blank books, calendars)?

III. Data in the MARC Format

  1. Do the records contain valid, appropriately assigned control numbers (i.e. vendor unique number in 001, LCCN, ISBN, ISSN)?
  2. Have you created any control numbers (LCCN, ISBN, ISSN) for your own purposes? Can they be identified and removed if need be?
  3. Do you have any practices that might affect matching algorithms that are predicated on the MARC format (e.g. series titles in the 245 instead of 440 or edition information in the 245 instead of the 250)?
  4. Do you catalog according to AACR2 for all records or only some? Which ones? In particular, we have concerns about the following areas:
    1. Publication dates: Do you use spurious dates, if so how, why? (e.g. reprints in which the date is a century later than the original publication date).
    2. Publisher vs Distributor: do you use the distributor in the 260 ‡b or the publisher when they are both known?
    3. Do you abbreviate certain data, such as publisher names?
    4. Do you enter complete pagination, including pre-paging (i.e. xiii, 245 p.)

IV. When you contract to be a PromptCat and/or OCLC Vendor Record Contribution Program provider, OCLC will request the following information:

  1. A list of the fields you use to transfer vendor/local information and what data goes in those fields. Include indicators, subfields, and what is or is not repeatable.
  2. A count of titles in your database by format or material type.
  3. The number of records you add to your database on a weekly basis.

Exhibit E Processing Options Data Elements

Overview

If a material vendor wishes to participate in PromptCat shelf-ready, they may do so completely or partially. A vendor may wish to:

  • provide local data to be mapped to a bibliographic record (such as barcode, sublocation, enumeration, etc.)
  • provide local data to be mapped to a label (sublocation or enumeration)
  • print OCLC created label files and send labels to or affix labels for the library

Processing Options Data Elements

Enumeration

Enumeration is a nonchronological scheme used by the publisher on a multi-part bibliographic unit to identify the parts of the unit and to show the relationship of the parts to the unit as a whole. OCLC also includes "copy" as part of the enumeration. Enumeration for the labels is sent separately from enumeration for the bibliographic record.

Enumeration for Labels

The enumeration data will be sent to OCLC in the EDIFACT QUOTES message in the GIR segment. For PromptCat the filing suffix (LFS) in the GIR segment is used to transmit the enumeration. Each separate enumeration line will be transmitted as a separate LFS. This data will include any captions that need to be included. The first data element will be added to the label right after the last line of the call number. OCLC will not do any editing on the data. The second and subsequent data elements, if transmitted, will be added to the label on a new line after the preceding data element data.

Examples:

If the enumeration for an item is v.1, it will be transmitted as follows:

GIR+C01+v.1:LFS'

If the enumeration for an item is v.1, pt.3, where the data between commas is to be on a separate line, it will be transmitted as follows:

GIR+C01+v.1:LFS'

GIR+C01+pt.3:LFS'

Enumeration for bibliographic records

For enumeration data to be added to a bibliographic record, it must be sent separately from data for the labels.

Data sent in the ENUM fields will not be added to the bibliographic record. If a library requests data elements to be mapped to their MARC record, these will have to be transmitted separately in the QUOTES messages in either a GIR segment or an IMD segment.

The various types of copy and holdings data that are supported in PromptCat and where they will be transmitted in the QUOTES message are:

Data Element

QUOTES Designation

copy number

GIR03 = LCV

edition

IMD02 = 100

part

IMD02 = 080

volume

IMD02 = 194

Additional data elements will be defined as they are needed.

These data elements will be added to the library's bibliographic record as profiled by the library in the same manner that all other data elements are added to the bibliographic record.

Location

Location information is transmitted when the vendor uses the location as the site id within the library. For example, the vendor identifies the locations within a library as MAIN, PHYS, BIOL, etc. OCLC uses the site id to profile PromptCat options at that level. This location data is transmitted within the GIR segment with GIR02=LLO.

The actual location text to appear in a record or on a label depends on the site's profile with OCLC. Each site can have different location text for what is to appear in the record and for what is to appear on the label. This is all based on the text transmitted as the site.

Sublocation

Sublocation is defined as a data element which provides additional information regarding the location of the item. A library site may be profiled for a particular location to appear on all their labels, yet periodically a particular item may have the need for additional data. For example, the site is profiled to receive CHEM as the location on all their labels. A particular item may need a sublocation of ‘REF'. Sublocation for the labels is sent separately from sublocation for the bibliographic record.

Sublocation for labels

This information will be transmitted within the GIR segment where GIR02 = 'LLS'. The placement of the sublocation in the label in relation to the call number will be determined by profiling at OCLC. The sublocation will always be on a separate line.

Sublocation for bibliographic records

This information will be transmitted within the GIR segment where GIR02 = 'LRS'. The mapping of the sublocation to the bibliographic record will be determined by the library and profiled at OCLC.

Label Files

Using call numbers from records copied by PromptCat, OCLC creates electronic files of labels. Label files may be picked up by the library, or by a vendor that wishes to print labels and affix them to materials prior to shipping. Label files are created in three formats: OCLC sl6, OCLC slb, and coded generic file.

OCLC sl6 and slb formats are two of the most popular label formats familiar to our users. The sl6 format is one spine label and two card and pocket labels separated with blank lines. The slb format is one spine label and one card and pocket label. The OCLC Cataloging Label program is available for use by vendors to print labels in either the sl6 or slb formats. Information on the Cataloging Label program is available at the Connexion website. If you wish to obtain the Cataloging Label program for use with PromptCat label files, you will be asked to sign an agreement. There is no cost for this software. Please contact me if you wish to use this software.

Vendors participating in the pilot have used the coded generic files of labels to print labels. Each vendor created its own software to print the coded generic files. OCLC formats the data in the coded generic files as follows:

  • Data is sent in "stream" format -- one record starts in the byte immediately after the previous terminator.
  • Each label record consists of several different types of fields -- record id, spine formatted call number, author, and title. The variable label record is terminated with an ASCII x '1D' and each field within the record will be terminated with an ASCII x '1E'.

See OCLC PromptCat Coded Generic Label File: Technical Specifications for further information.

Exhibit F EDX Assessment Form

Online form: https://www.oclc.org/forms/edx-assessment.en.html

Contact VendorRecords@oclc.org to receive form.

Exhibit G OCLC-Affiliated Regional Network Contacts

For technical support, contact your regional service provider or OCLC support using the form on the Contact Us page.

Your OCLC Vendor liaison is:

Beth Fogler
T: +1-614-761-5236
T: (800) 848-5878, ext. 5236 (US & Canada)
F: +1-614-718-7212
E: foglerb@oclc.org