MARS Authority Control

Always Load in Correct Order

March 20th, 2009

WHY IT’S A GOOD IDEA TO LOAD YOUR AUTHORITIES IN THE ORDER WE PROCESSED THEM

We have often had clients ask us in what order they should load their authorities, if they have several groups to load. Perhaps you’ve found yourself in the situation where you haven’t loaded your last set of authority updates and now you’ve sent in new bibliographic records to be processed, or maybe your next scheduled update has arrived. You find yourself looking at two or more batches of authorities and you wonder, “Does it really matter which goes first?” Absolutely! And here’s why:

* Say you had a scheduled authority update in December, but things went crazy busy and you haven’t had time to load those authorities yet. Now it’s March and you have a large group of new bibliographic records that need processing, so you send them in. Thinking you can save time by loading both the December updates and the bibliographic authorities at the same time, you wait for the new group to be returned. But now … which to load first? You should load the December updates first. Example:

* You have a heading for Doe, John,$d1955- and sometime last year LC updated that to Doe, John Joseph,$d1955- (and if we’re lucky, LC added the “old” heading as a 400 see-reference). This changed authority would deliver with the December group you hadn’t loaded yet.

* Now you’ve sent in your bib records and in there is the heading for Doe, John,$d1955- . However, between December and today’s date LC decided to edit the record again and they put out a new authority with the new heading of: Doe, John J.,$d1955- (and still kept the original “old” heading of Doe, John,$d1955- in a 400 see-reference tag).

* With the processed bibs you’d get back the very newest authority for Doe, John J.,$d1955-, which is what you’d want in your system.

* But if you decide to load the bibliographic records and associated authorities first and the December authority updates second, the middle version of Doe, John Joseph,$d1955- (from the December updates) would overwrite the newest authority sent with the bibs, and you would be stuck with an older, not current LC version of this heading.

Which is why it’s always wisest to load oldest-to-newest, when you’re working with several projects at once.

Written by: Judy Archer

Tags: Authority Records, Bibliographic Records, ILS, Loading Records
Posted in Automated Authority Control, Bibliographic Records | Comments Closed

Understanding the MARC Structure

March 16th, 2009

Underneath the MARC records that we all know and love is a somewhat cryptic structure that tells our systems how to read the record. Luckily, this structure rarely gets corrupted, but when it does, it’s good to have a basic understanding of how to read and understand a MARC Format.

The MARC format is a text based format, meaning you can open it with a text editor. It is probably a good idea to only open a few (or only one!) record in a text editor, because it’s very difficult to read otherwise.

Every MARC record starts with a leader, the leader gives your system information about the record, including how big it is and what type of a record it is. Next is what is called a directory. Just like a normal directory, it tells you what tags are in it, and where the data for each tag is located.

This is how a leader is defined for a Bibliographic Record:

Leader: 

Position | Description              | Explination
------------------------------------------
00-04 | Record Length               = This is how long the record is
05    | Record Status               = Is the record new, changed or deleted
06    | Type of record              = Authority, Book, Computer file, etc.
07    | Bibliographic level         = Monograph/Serial/etc.
08    | Type of control             = Archival or not
09    | Character coding scheme     = MARC8 or UTF8
10    | Indicator count             = # of indicators each tag has
11    | Subfield code count         = # of subfield codes that make up 1 subfield
12-16 | Base address of data        = The byte where actual record data begins
17    | Encoding level              = Level of encoding/cataloging
18    | Descriptive cataloging form = AACR2
19    | Multipart resource record level = # (blank)
20    | Length of the length-of-field = # of bytes to store how long each tag is
21    | Length of the starting-character-position = # of bytes to store where tag begins
22    | Length of the implementation-defined portion = Rarely used
23    | Undefined                   = Not used

Looking at a leader can be confusing, but it’s also the only way to find some problems.

Here is an example record, as it may appear on your screen:
001 __ 3180021
005 __ 20050216201852.0
008 __ 040805s2005 nyu b 001 0 eng
010 __ ▼a 2004018260
020 __ ▼a0415971675 (alk. paper)
035 __ ▼a(DLC) 2004018260
040 __ ▼aDLC▼cDLC▼dDLC▼dCaONFJC▼dOrLoB-B
043 __ ▼an-us—
050 00 ▼aPS374.H56▼bO73 2004
082 00 ▼a813/.5409358▼222
090 __ ▼aPS374.H56▼bO73 2005
100 1_ ▼aOrbán, Katalin.
245 10 ▼aEthical diversions :▼bthe post-holocaust narratives of Pynchon, Abish, DeLillo, and Spiegelman /▼cKatalin Orbán.
260 __ ▼aNew York :▼bRoutledge,▼c2005.
300 __ ▼aix, 209 p. ;▼c24 cm.
440 _0 ▼aLiterary criticism and cultural theory
504 __ ▼aIncludes bibliographical references (p. 193-205) and index.
505 00 ▼gCh. 1.▼t”Mauschwitz” : monsters, memory, and testimony — ▼gCh. 2.▼tFamiliarity and forgetfulness in Walter Abish’s fiction — ▼gCh. 3.▼tPinpricks on the Ars(e) Narrandi : liminality and oven-games in Gravity’s rainbow.
600 10 ▼aPynchon, Thomas▼xKnowledge▼xHolocaust, Jewish (1939-1945)
600 10 ▼aSpiegelman, Art▼xKnowledge▼xHolocaust, Jewish (1939-1945)
600 10 ▼aAbish, Walter▼xKnowledge▼xHolocaust, Jewish (1939-1945)
600 10 ▼aDeLillo, Don▼xKnowledge▼xHolocaust, Jewish (1939-1945)
650 _0 ▼aAmerican fiction▼y20th century▼xHistory and criticism.
650 _0 ▼aHolocaust, Jewish (1939-1945), in literature.
650 _0 ▼aJudaism and literature▼zUnited States▼xHistory▼y20th century.
650 _0 ▼aWorld War, 1939-1945▼zUnited States▼xLiterature and the war.
650 _0 ▼aEthics in literature.
650 _0 ▼aJews in literature.
852 0_ ▼bkr▼hPS374.H56▼iO73 2005
856 41 ▼3Table of contents▼uhttp://www.loc.gov/catdir/toc/ecip0421/2004018260.html
949 __ ▼aApproval▼b1628024-35▼c67.20▼d1▼i20027328▼jUSD▼tBook

Now, if we are to take a look at the actual MARC structure, it looks like this (when looking at an underlying MARC record, you may not be able to see all of the special characters like end of field and end of record marks, but they are there):

01872cam a2200397 a 450000100080000000500170000800800410002501000170006602000280008303500220011104000360013304300120016905000240018108200

210020509000240022610000210025024501180027126000340038930000250042344000430044850400640049150502240055560000620077960000

620084160000600090360000590096365000590102265000500108165000660113165000650119765000260126265000240128885200280131285600

7801340949005601418318002120050216201852.0040805s2005 nyu b 001 0 eng a 2004018260 a0415971675 (alk. paper) a(DLC) 2004018260 aDLCcDLCdDLCdCaONFJCdOrLoB-B an-us—00aPS374.H56bO73 200400a813/.5409358222 aPS374.H56bO73 20051 aOrbán, Katalin.10aEthical diversions :bthe post-holocaust narratives of Pynchon, Abish, DeLillo, and Spiegelman /cKatalin Orbán. aNew York :bRoutledge,c2005. aix, 209 p. ;c24 cm. 0aLiterary criticism and cultural theory aIncludes bibliographical references (p. 193-205) and index.00gCh. 1.t”Mauschwitz” : monsters, memory, and testimony — gCh. 2.tFamiliarity and forgetfulness in Walter Abish’s fiction — gCh. 3.tPinpricks on the Ars(e) Narrandi : liminality and oven-games in Gravity’s rainbow.10aPynchon, ThomasxKnowledgexHolocaust, Jewish (1939-1945)10aSpiegelman, ArtxKnowledgexHolocaust, Jewish (1939-1945)10aAbish, WalterxKnowledgexHolocaust, Jewish (1939-1945)10aDeLillo, DonxKnowledgexHolocaust, Jewish (1939-1945) 0aAmerican fictiony20th centuryxHistory and criticism. 0aHolocaust, Jewish (1939-1945), in literature. 0aJudaism and literaturezUnited StatesxHistoryy20th century. 0aWorld War, 1939-1945zUnited StatesxLiterature and the war. 0aEthics in literature. 0aJews in literature.0 bkrhPS374.H56iO73 2005413Table of contentsuhttp://www.loc.gov/catdir/toc/ecip0421/2004018260.html aApprovalb1628024-35c67.20d1i20027328jUSDtBook

And here is a sample analysis of this record:

Leader: 

Position | Description              | Data
------------------------------------------
00-04 | Record Length               = 01872 : confirmend in a hex editor the record is this length
05    | Record Status               = c
06    | Type of record              = a
07    | Bibliographic level         = m
08    | Type of control             = # (blank)
09    | Character coding scheme     = a
10    | Indicator count             = 2
11    | Subfield code count         = 2
12-16 | Base address of data        = 00397 : confirmed in hex editor this is correct
17    | Encoding level              = # (blank)
18    | Descriptive cataloging form = a
19    | Multipart resource record level = # (blank)
20    | Length of the length-of-field portion = 4
21    | Length of the starting-character-position portion = 5
22    | Length of the implementation-defined portion = 0
23    | Undefined                   = 0

Directory: 

Tag | Field length | Starting character position
001 0008 00000
005 0017 00008
008 0041 00025
010 0017 00066
020 0028 00083
035 0022 00111
040 0036 00133
043 0012 00169
050 0024 00181
082 0021 00205
090 0024 00226
100 0021 00250
245 0118 00271
260 0034 00389
300 0025 00423
440 0043 00448
504 0064 00491
505 0224 00555
600 0062 00779
600 0062 00841
600 0060 00903
600 0059 00963
650 0059 01022
650 0050 01081
650 0066 01131
650 0065 01197
650 0026 01262
650 0024 01288
852 0028 01312
856 0078 01340
949 0056 01418| (56+1418 = 1474) 1474 = Ending field terminator below

Fields:
(For our convenience in readin the record we have replaced certain non-displayed fields with graphical representations)
| = field terminator (also used to terminate the directory)
$ = subfield code delimiter
^ = record terminator
_ = utf8 characters have been replaced by an undersore so the postions line up in a non-hex display

3180021|20050216201852.0|040805s2005 nyu b 001 0 eng | $a 2004018260| $a0415971675 (alk. paper)| $a(DLC) 2004018260| $aDLC$cDLC$dDLC$dCaONFJC$dOrLoB-B| $an-us—|00$aPS374.H56$bO73 2004|00$a813/.5409358$222| $aPS374.H56$bO73 2005|1 $aOrb__n, Katalin.|10$aEthical diversions :$bthe post-holocaust narratives of Pynchon, Abish, DeLillo, and Spiegelman /$cKatalin Orb__n.| $aNew York :$bRoutledge,$c2005.| $aix, 209 p. ;$c24 cm.| 0$aLiterary criticism and cultural theory| $aIncludes bibliographical references (p. 193-205) and index.|00$gCh. 1.$t”Mauschwitz” : monsters, memory, and testimony — $gCh. 2.$tFamiliarity and forgetfulness in Walter Abish’s fiction — $gCh. 3.$tPinpricks on the Ars(e) Narrandi : liminality and oven-games in Gravity’s rainbow.|10$aPynchon, Thomas$xKnowledge$xHolocaust, Jewish (1939-1945)|10$aSpiegelman, Art$xKnowledge$xHolocaust, Jewish (1939-1945)|10$aAbish, Walter$xKnowledge$xHolocaust, Jewish (1939-1945)|10$aDeLillo, Don$xKnowledge$xHolocaust, Jewish (1939-1945)| 0$aAmerican fiction$y20th century$xHistory and criticism.| 0$aHolocaust, Jewish (1939-1945), in literature.| 0$aJudaism and literature$zUnited States$xHistory$y20th century.| 0$aWorld War, 1939-1945$zUnited States$xLiterature and the war.| 0$aEthics in literature.| 0$aJews in literature.|0 $bkr$hPS374.H56$iO73 2005|41$3Table of contents$uhttp://www.loc.gov/catdir/toc/ecip0421/2004018260.html| $aApproval$b1628024-35$c67.20$d1$i20027328$jUSD$tBook|^

Ending field terminator = 1474
Record terminator position = 1475

Summary
———————————
Ending field position = 1475
Base address of data = 397

1475 + 397 = 1872 (Record Length)

Tags: MARC, Structure
Posted in Bibliographic Records | Comments Closed

The Significance of the Subfield w

February 19th, 2009

One of the things we’re asked about occasionally is what the meaning of a $w (subfield w) is in See References in Authority Records (4XX tags). We hope this will help explain what the $w does.

The $w is a control subfield with up to four character positions. One-character alphabetic codes are used to describe special instructions that apply to the display of 4xx and 5xx fields. The definitions of the codes in subfield $w are dependent on the character position in which they occur. The coding of any character position mandates that each preceding character position contains a code or a fill character (|); succeeding character positions need not be used. For example, if $w/0 and $w/1 are not applicable but $w/2 is applicable, then $w/0 and $w/1 each contain code n (Not applicable) or a fill character (|). The n is used more often than the fill character.

We will use the following positions and codes to identify the references that our process adds to the LC authority records.

/2 – Earlier form of heading. We will use either of the following codes depending on which authority file (LC, MeSH, etc.) we found the former heading in.

e – Earlier established form of heading (national authority file) – We will use this for headings found in the LC Authority File

o – Earlier established form of heading (other authority file) – This will be used for headings found in other authority files.

/3 – Reference display. The following codes will be applicable in our process.

a – Reference not displayed

n – Not applicable – reference would display (If code n is applicable, subfield $w/3 need not be coded.)

Client libraries will be able to choose if they would like these added references to display or not and we will code the added reference accordingly.

$5 – Institution to which Field Applies. This subfield contains the MARC code of the institution or organization that has added a tracing, reference, or linking entry field to an authority file record. It is used for institution-specific information that may or may not apply to the universal use of the authority record.
We will identify these added references by adding a $5 with our institution code to them. i.e. $5UtOrBLW

So, a sample 4xx from a match made by our new process for a client that does NOT want the reference to display would look like:

4xx 0 $wnnea$aSome subject$xSubfield too$5UtOrBLW

It could look like this if they did want the reference to display:

4xx 0 $wnne$aSome subject$xSubfield too$5UtOrBLW

For more infomation on $w and $5, see the MARC21 Authority Format under the Tracings and References-General Information section at the following link:

http://www.loc.gov/marc/authority/adtracing.html

Tags: See References
Posted in Authority Control | Comments Closed

Open Ended Subject Dates

February 5th, 2009

What do you do with open ended subject dates, when the Library of Congress has created authorities with closed dates and no open ended see references? This was the question we sought to answer over the last week.

Take for example, you have a bib record with the following heading:
650 _0 $aWorld politics$y1955-

The authority for this heading has the following authorized and see references:

sh 85148226
150 _0 $aWorld politics$y1955-1965

Although these records are related, in an automated process they wont be linked, and the ending date will not be supplied. We decided there had to be a better way to link these up, and the process wasn’t very difficult, in fact, we’ve been doing it for names forever!

The answer is in the hierarchy. If we allowed our hierarchy system to include the closing date for subjects, we could provide the open ended bib dates a form of hierarchy to match against.

In essence, our Authority record above would create 2 possible heading matches. They would be (after following our normalization routine):
$aWORLD POLITICS$1955-1965
$aWORLD POLITICS$1955-

Now your bib heading $aWorld politics$y1955- could match and update against the proper LC Authority record. When we started researching this, we discovered another caveat. In some cases there are 2 authorities, one with an open ended date and one with a closed date. Our conclusion is that it’s obviously better in this case to trust what the original cataloger selected and match the open ended authority, rather than the closed authority. An example of this would be the authorities:
sh 85007061
151_0$aArgentina$xPolitics and government$y1810-
sh 85007060
151_0$aArgentina$xPolitics and government$y1810-1817

In this case, our bib record heading of 651 _0 $aArgentina$xPolitics and government$y1810- would match the first Authority.

After figuring that out, we discovered a real surprise, headings that were exactly the same, but had different closing dates. For example, France — History — 1789 has the following Authorities:

sh 85051348
151_0$aFrance$xHistory$y1789-
sh 85051347
151_0$aFrance$xHistory$y1789-1900
sh 85051346
151_0$aFrance$xHistory$y1789-1815
sh 85051314
151_0$aFrance$xHistory$y1789-1793

Obviously there are different meanings for different time periods starting from 1789. Now, if the bib record had an open ended date, it would take the open ended authority. And, if the bib record had any of the closed dates, it would take the authority with the matching date.

Our focus here, has been making sure that open ended authorities take the best possible match, and if there are no open ended authorities, matching it to the correct closed date authority, without creating false positive results.

Tags: Heading Matching, Open ended dates
Posted in Automated Authority Control | Comments Closed

Announcing a new Product

January 22nd, 2009

Bibliographic Record Update Service

Due to popular demand, Backstage Library Works is now offering a third ongoing service product that we call Bibliographic Record Update Service. This service updates your bibliographic records during our Notification Service.

Over the years Backstage has offered two ongoing products for automated authority control. For all of your new bibliographic records we provide a product called Current Cataloging. Current Cataloging is designed to pick up where you left off with the original automated authority process. We take all of your new bibliographic records and process them exactly like we processed your original bibliographic database. With this process we return to you newly validated bibliographic records with updated authority headings that match the new and updated authority records we return to you. We retain in your master authority file all new and updated authority records.

To complement our Current Catalog product Backstage offers Notification Services. Notification Services matches your library’s master authority file against the National Databases on pre determined intervals to give you updated and new authority records that replace the old authority records on your database.

Both of these products together are a very efficient and cost beneficial way to maintain your authority database. For most libraries this is sufficient. However, with the Notification product it is left to the Library’s ILS system to reconcile new authorities delivered by Backstage with existing authority headings in their bibliographic records. The process of importing and updating the Library’s authority records is straightforward. Control numbers are matched up and the newly imported record overlay the old authority record. However, for some ILS systems the updating of the authority heading is done by a string match on the 1XX or the 4XX of the authority record. If neither are the same as the old bibliographic authority heading the record will not automatically reconcile. Typically, this is done through reports but the effort to reconcile or to populate the authorized heading in the bibliographic record can be partially manual and can be time consuming depending on the ILS and the library.

How big of a job is this? It really depends. Up until the advent of the closed death date Backstage did not receive a call involving the amount of time you spent on reconciling new authority records to existing authority headings but that has changed. Recently, we have noticed an interest in having this part of the ongoing authority process further automated. As a result we have developed a complimentary service we call Bibliographic Record Update Service.

What we will do during our Bibliographic Record Update Service is maintain your bibliographic records along with your authority records. With this service, Backstage will have a master bibliographic file as well as a master authority file for your library which we will maintain for you. When Notification Services run on your database, our Bibliographic Record Update Service will also run. Backstage will update or reconcile your bibliographic database as well as your authority file before we send records back to you. We will deliver to you updated bibliographic records along with your updated and new authority files.

How does this help you? The time spent reconciling your new and updated authority records to their associated bibliographic records will disappear. Your library will simply need to upload updated bibliographic records along with new and updated authority records. The process will be identical to what you do for Current Cataloging.

To maintain your bib file we will need any changed bibliographic records sent to Backstage prior to your Bibliographic Update service.

If you are interested in this new service please contact your sales representative or the MARS authority control team for more information.

<-->

Tags: Bibliographic Records, New Product
Posted in Automated Authority Control, Bibliographic Records | Comments Closed

Heading to ALA

January 21st, 2009

ALA Midwinter

Coming to Denver for ALA Midwinter meeting held in Denver? We’ll be there too!

Drop in and see us at booth 2244. We’d love to talk about our Digitization services, Cataloging, Retrospective conversion, or of course our Automated Authority Control, including a big new announcement about a new product we’ll be offering to help make your authority control even easier.

Tags: ALA, Announcements
Posted in Automated Authority Control, Conferences | Comments Closed