Difference between revisions of "Profile Guide Step 3.2"
m (Protected "Profile Guide Step 3.2" [edit=sysop:move=sysop] [cascading]) |
|||
| Line 309: | Line 309: | ||
[[Profile Guide Step 3.3 | Go to the Next Section]] | [[Profile Guide Step 3.3 | Go to the Next Section]] | ||
| + | |||
| + | [[category:Profile Guide]] | ||
Revision as of 10:16, 30 October 2008
The second phase of MARS 2.0 Authority Control is Authority Matching. While Authority Cleanup improves authority controlled headings using custom MARS 2.0 correction tables, authority matching compares each authority controlled heading in your bibliographic records against authority record headings from any of a number of national and other authority files. Authority files available in MARS 2.0 Authority Control processing are:
National Authority Files
- NAF -LC Name Authority File
- SAF -LC Subject Authority File
- LCG -LC Genre Authority File
- ACF -LC Children’s Authority File
- MeSH -National Library of Medicine MeSH Authority File
- NLC -Library and Archives Canada Canadiana Authority File
- NLC -Library and Archives Canada Canadian Subject Headings Authority File
- AAT -The Art & Architecture Thesaurus
- GSAFD -Guidelines on Subject Access to Individual Works of Fiction, Drama, etc.
Other Authority Files
- Your local Authority File
- MARS 2.0 Auxiliary File
Contents
Goals of Authority Matching
Authority matching uses the headings in authority records to update or correct the bibliographic headings so they conform to current standards. Authority matching is also the basis for providing full authority records for your local system. The goals of authority matching are to:
- Identify invalid headings and change them to valid forms based on cross-references found in authority records (convert to the established form of heading)
- Identify headings that have incorrect spacing, punctuation, indicators or subfield codes, and change to the correct form based on matches found on authority records in the NAF/SAF
- Identify invalid higher levels of a heading and change them to their valid forms, based on cross-references found in authority records (for example, convert the name portion of a name/title heading to the established form of heading), and then compare the updated heading against the authority records
- Identify matched MARC authority records that are to be selected and distributed to the library
- Identify headings requiring manual review, if the Library requests manual review by the MARS 2.0 staff
Fields Under Authority Control
The MARS 2.0 Authority Control process corrects and updates the full range of authority controlled headings. The following bibliographic headings/tags are included in MARS 2.0 authority control processing:
Subfields Disregarded
A number of MARC subfields are disregarded during MARS 2.0 authority matching. In the following headings, the volume designations in fields 810 and 440 subfields $v, the heading linkage information in field 130 $6 subfield, and the ISSN in field 440 $x subfield, are all examples of subfield information which is not under authority control.
Examples 810 2 $aJohn Bartholomew and Son.$tBartholomew world travel series ;$v10. 130 0 $6880-01$a”Hsūūan lai his kan” his lieh. 440 4 $aRomanica Gothoburgensia ;$v12, 16$x0080-3863
Subfields Matched or Ignored
The table below shows the subfields that are included in the MARS 2.0 authority matching process:
| MARC Tag | Subfields Retained During Matching | Subfields Ignored During Matching |
|---|---|---|
| 100, 400, 700, 800 | a b c d f g h k l m n o p q r s t y z | e u v w x 2 3 4 5 6 |
| 110, 410, 710, 810 | a b c d f g h k l m n o p r s t y z | e u v w x 2 3 4 5 6 |
| 111, 411, 711, 811 | a b c d e f g h k l m n o p q r s t u y z | v w x 2 3 4 5 6 |
| 130, 830 | a d f g h k l m n o p r s t x y z | v w 2 3 5 6 |
| 730 | a d f g h k l m n o p r s t y z | v w x 2 3 5 6 |
| 440 | a n p | v w x 6 |
| 600 | a b c d f g h k l m n o p q r s t v x y z | e u w 2 3 4 5 7 |
| 610 | a b c d f g h k l m n o p r s t v x y z | e u w 2 3 4 5 6 |
| 611 | a b c d e f g h k l m n o p q r s t u v x y z | 2 3 4 5 6 |
| 630 | a d f g h k l m n o p r s t v x y z | w 2 3 5 6 |
| 650 | a b c d v x y z | e w 2 3 6 |
| 651 | a b v x y z | w 2 3 6 |
| 655 | a b c v x y z | e w 2 3 6 |
Normalization
Headings from both your bibliographic records and the MARS 2.0 authority files (LC, MeSH, etc.) are normalized before they are compared for matching. Backstage uses the NACO normalization standard. During normalization:
- Alphabetic characters are converted to uppercase
- The first comma will be retained in the ‘a’ subfield for personal name headings
- All other punctuation is removed
- Certain diacritics and hyphens are left in for CJK normalization
- All other diacritics are removed
- Special characters are replaced by an alphabetic equivalent
- Subfield codes are removed
- Subfield delimiters for all but first subfield will be left in
Example Bib record heading: $aArchitecture$zBrazil$xS’ao Paulo (State) Normalized heading: ARCHITECTURE$BRAZIL$SAO PAULO STATE
Notice the subfield codes and the diacritic in S’ao Paulo have been discarded when constructing the normalized form of the heading. Because the subfield codes are ignored during the authority matching process, invalid subfield codes do not affect the matching and may be corrected during the process.
For this example, the normalized form of the established heading (1XX) in the authority record is:
Authority heading: $aArchitecture$zBrazil$zS~ao Paulo (State) Normalized heading: ARCHITECTURE$BRAZIL$SAO PAULO STATE
Since the normalized forms of the bibliographic and authority headings are the same, the established form in the authority record replaces the form in the bibliographic record. The bibliographic heading then contains the correct diacritic (~ instead of ‘) and the subfield code for “S’ao Paulo” has been corrected to $z.
Step 3.2 Question
Normalization of Name Headings
Normalization of the name headings is different. All information in the personal name heading is considered when determining matches. Only $a of the name heading is removed (i.e., the delimiter and subfield, not the data within $a). Name headings are defined as bib tags 100 (along with 240/243), 600, 700 and 800. Death date and other information in the bib heading date will always be considered. A bib heading with a death date will not match against an authority heading without one.
Normalization of a name heading is as follows:
| Example | Bibliographic Record | Authority Record |
|---|---|---|
| Un-normalized form | $aCampbell, James,$d1826-1900 | $aCampbell, James,$d1826- |
| $aCampbell, James,$d1826-1900 | ||
| $aCampbell, James,$d1826-1910 | ||
| Normalized form | CAMPBELL, JAMES$1826-1900 | CAMPBELL, JAMES$d1826 |
| CAMPBELL, JAMES$d1826 1900 | ||
| CAMPBELL, JAMES$d1826 1910 | ||
| Match | $aCampbell, James,$d1826-1900 |
When multiple matches occur on personal name headings, the bib heading will not be replaced with any of the matches. In the case with multiple potential authority matches, the program cannot know which one to take over the others and so the bib heading remains unchanged. The multiple potential authority matches will be reported to the user in an optional report R10 - Heading Matches Established Headings in Multiple Authority Records (see Step 5 of the Planning Guide).
During the final phase of Authority Control, the incorrect form of the heading in the bibliographic record is replaced by the established form of the authority record. The corrected heading contains missing or changed diacritics.
Bibliographic Text Replaced
When a heading matches a 1XX authorized heading or a 4XX unauthorized heading in an authority record, MARS 2.0 Authority Control processing:
- Replaces the bibliographic heading text with the text from the 1XX field of the authority record
- Replaces the second and third characters of the bibliographic tag with the second and third characters of the 1XX authority heading tag, as appropriate (see “Tag changes” in Step 3.2 of the Planning Guide)
This replacement process corrects errors in capitalization, punctuation, spacing, subfield coding and some tagging errors. MARS 2.0 ensures that the appropriate ending punctuation is supplied in the bibliographic heading.
Matching of Name Headings
Currently if you have a ‘generic’ name heading (no date or full form of name) in your bib record and a match is found on a See reference (400) of an LC Personal Name authority record, MARS 2.0 will automatically flip the authority heading to reflect the new authority.
From the example below, if your bib authority heading was $aCampbell, James, your new bib authority heading would become:
100 1 $aCampbell, James,$d1826-1910
This would link to the authority below:
001 n 2001026796 005 20010614121312.0 008 010614n| acannaabn |n aaa 010 $an 2001026796 100 1 $aCampbell, James, $d1826-1910 400 1 $aCampbell, James 670 $aDirect your letters to San Jose, 2000:$bt.p. (James Campbell) CIP data (Campbell, James 1826-1910)
Choosing to flip on ‘generic’ name headings will give you a greater match rate, but it can cause mismatched headings. The MARS 2.0 software allows for the flip of all name authority headings including the ‘generic’ heading if there is a match to a See reference (400) in the LC authority record. If you do not want the generic name heading in the 4XX to match, choose No—Do not Flip Generic Name Headings.
See From Tracings
The MARS 2.0 Authority Matching software looks at both the authorized (or established) form of a heading in the authority record’s 1XX field and the unauthorized form in the 4XX fields (See From tracings).
Example 650 0 $aPhilippine American War, Philippines, 1899-1902. 651 0 $aPhilippines$xHistory$yInsurrection, 1899-1901. 651 0 $aUnited States$xHistory$yPhilippine Insurrection, 1899-1902.
The normalized headings from the bibliographic records are compared to the normalized authorized (1XX) and unauthorized (4XX) authority record headings. The headings above match headings contained in LC subject authority record sh 85100792 (only portion of the authority record are shown below):
005 19980416161019.4 010 $ash 85100792 151 0 $aPhilippines$xHistory$yPhilippine American War, 1899-1902 450 0 $aFilipino American War, Philippines, 1899-1902 450 0 $aPhilippine American War, Philippines, 1899-1902 450 0 $aPhilippine Insurrection, Philippines, 1899-1902 451 0 $aPhilipines$xHistory$yInsurrection, 1899-1901 451 0 $aUnited States$xHistory$yPhilippine Insurrection, 1899-1902
The three unauthorized headings in the bibliographic record would all be replaced by the established form of the heading:
151 0 $aPhilippines$xHistory$yPhilippine American War, 1899-1902
The MARS 2.0 software will then merge or de-duplicate the three identical authority headings to form one heading:
650 0 $aPhilippines$xHistory$yPhilippine American War, 1899-1902.
Hierarchy
Unauthorized headings (See From tracings) are often included only in the authority record for the highest level of a heading. The authority record for Apples, for instance, contains the older form of the heading, Apple. The authority record for the fuller heading, Apples—Marketing, however, does not contain a tracing for Apple—Marketing.
Using all hierarchy levels during Authority Matching, i.e., the highest level Apples plus all fuller forms such as Apples—Marketing and Apples—Washington (State)—Marketing, allows a greater number of headings to be updated.
In the following examples, normalized headings are constructed for each level of hierarchy in the heading:
Example 1 Original Form of Heading in Bibliographic Record: 650 0 $aElectronic digital computers$xProgramming$vCongresses. “Levels of Hierarchy” for this Heading: 650 0 $aElectronic digital computers 650 0 $aElectronic digital computers$xProgramming 650 0 $aElectronic digital computers$xProgramming$vCongresses Partial Match to 4xx Heading in LC Subject Authority Record “sh 85107310”: 150 0 $aComputer programming 450 0 $aElectronic digital computrs$xProgramming Corrected Form of Heading in Bibliographic Record (a partial “flip” or conversion): 650 0 $aComputer programming$vCongresses.
Example 2 Original Form of Heading in Bibliographic Record: 100 02 $aGregory,$cof Nazianzus, Saint.$tOrations.$nOration 27-31.$lGerman & Greek. Partial Match to 4xx Heading in LC Name Authority Record “n 90662896”: 100 00 $aGregory,$cof Nazianzus, Saint.$tTheological orations. 400 00 $aGregory,$cof Nazianzus, Saint.$tOrations.$nOration 27-31 Corrected Form of Heading in Bibliographic Record (a partial “flip”): 100 02 $aGregory,$cof Nazianzus, Saint.$tTheological orations.$lGerman & Greek.
Constructing Multiple Levels of Hierarchy
When constructing multiple levels of hierarchy for a heading, the shortest form of the heading that is matched against the authority file is referred to as the highest level of the heading.
Example Full Heading: 700 12 $aFreud, Sigmund,$d1856-1939.$tSelections.$f1980 Highest Level Heading: 700 12 $aFreud, Sigmund,$d1856-1939.
Hierarchy Record Delivered to MARC File
When a hierarchy match is found not only is the authority heading replaced with the appropriate authority record, but the hierarchal version of the authority record is delivered to your MARC authority file allowing appropriate cross references to be connected to your new heading.
Subfields Representing Highest Level Heading
The MARS 2.0 software drops the right-most subfield(s) in generating a higher-level heading and it stops (it has reached the highest level heading) when it encounters any subfield code in the following table:
| MARC Tag | Subfield Codes Indicating Highest Level Heading |
|---|---|
| 100, 400, 600, 700, 800 | a b c d q |
| 110, 410, 610, 710, 810 | a b |
| 111, 411, 611, 711, 811 | a b |
| 130, 430, 630, 730, 830 | a d |
| 440 | a |
| 650 | a c d |
| 651 | a |
| 655 | a |
Name / Title Headings
Name/Title headings present a unique challenge. Name/Title headings are often coded as a subfield $a / $t combination in either a 7XX or 6XX field. In the following examples, LC authority records containing the full Name/Title headings have been established and would be matched during normal MARS 2.0 authority matching.
Example 600 10 $aTwain, Mark,$d1835-1910.$tInnocents abroad. 710 1 $aCanada.$tTreaties, etc.$d1992 Oct. 7.
Name / Title Split and Temporary Headings
At other times, however, the Name/Title heading is split between 2 fields, a 1XX Main Entry and a 240 Uniform Title. In order to correctly process 1XX/240 Name/Title headings, temporary headings are built whenever a record contains both a 1XX and 240 field.
Example Bibliographic Fields 100 0 $aAristotle. 240 10 $aWorks.$lEnglish.$f1984 become: 100 0 $aAristotle.$tWorks
Using normal matching, the 100 field above would match the authority record for Aristotle. There is no authority record, however, for the uniform title found in the 240 field. The two fields (100/240) are combined during the matching process only.
The temporary MARS 2.0 authority heading would then match the correct LC authority record "n 8242241”: 100 $aAristotle.$tWorks.$lEnglish.$f1984. 100 $aAristotle.$tComplete works of Aristotle.$f1984
Name / Title Tracing
Inclusion of LC authority records in a library’s record distribution is an important MARS 2.0 service. Most local systems use the 400 See From tracing in the above authority record to guide patrons from the unauthorized to the authorized heading using a display similar to the following:
Aristotle—Complete works of Aristotle—1984 See Aristotle—Works—English—1984
Name / Title Hierarchy
Use of all levels of hierarchy also affects matching using MARS 2.0 authority Name/Title headings. If there is no authority record containing the complete Name/Title heading, any authority records for the higher levels of the heading will be matched and provided.
Example 1 110 1 $aUnited States.$bCongress.$bHouse.$bSelect Committee on Hunger. $tProgress report of the Select Committee on Hunger (1992) Closest match (LC n 82270415): 110 10 $aUnited States.$bCongress.$bHouse.$bSelect Committee on Hunger.
Example 2 100 0 $aAristotle.$tDe pomo. Closest match (LC n 79004182): 100 0 $aAristotle.
Usage Codes
The matching process checks the authority record heading use codes from the authority record (008 bytes 14-16 and the type of series code 008 byte 12) to ensure that the bibliographic heading has been used appropriately:
008 [12] – Type of series 008 [14] – Main or added entry 008 [15] – Subject added entry 008 [16] – Series added entry
If the bibliographic heading has been used inappropriately, an error report is produced and no conversion will occur. For example, if a heading is used as a series added entry in a bibliographic record but the authority record indicates that the heading is not valid for use as a series, the heading will not convert.
A report of the matches in which the bibliographic usage and the authority record authorized usage codes conflict is available (see Step 5, report R11 for details).
Reference Records
LC has created reference authority records, which contain a heading in the 1XX field, which are not authorized for use in bibliographic records. These records are identified by a value b or c in byte 9 of the 008 field. The MARS 2.0 programs do not match bibliographic headings against an LC reference record. No heading is converted from reference authority records.
MARS 2.0 Auxiliary File
MARS 2.0 programs also match bibliographic headings against authority headings in the MARS 2.0 Auxiliary File. The MARS 2.0 Auxiliary File contains records with additional validated headings and additional cross-references not present in national authority files. The additional cross-references convert incorrect or obsolete forms to the authorized form of the headings. Authority records from this file are used during MARS 2.0 Authority Matching only and are not distributed to libraries.
For example, concentration camps were previously tagged as 651 headings, with the authority records in the LC SAF. When the tagging was changed to X10 and the authority records moved to the LC NAF, no 4XX See From tracings were added to the LC authority records for the obsolete form of the heading. The MARS 2.0 staff created records in the MARS 2.0 Auxiliary File that make the following conversion:
Example Bibliographic Heading: 651 0 $aHusum-Schwesing (Germany : Concentration camp) Converts to current form: 610 20 $aHusum-Schwesing (Concentration camp)
