Serial Computer Files Under Format Integration (Crystal Graham) Ann Ercelawn 29 Nov 1995 14:01 UTC

Date: Tue, 28 Nov 1995 15:53:00 -0800
From: Crystal Graham <Crystal_Graham@UCSDLIBRARY.UCSD.EDU>
Subject: Serial computer files under format integration

Forwarded from the INTERCAT list, fyi -- ed.
________

     This is a position paper from Becky Ringler and Crystal Graham to the
     CONSER Task Force on Electronic Resources distributed for your information
     and comment.  Permission is granted to forward it to other listservs or
     individuals.

______________________________

To: Jean Hirons, Library of Congress
From: Crystal Graham (cgraham@ucsd.edu) and Becky Ringler (rringler@ucsd.edu),
University of California, San Diego
Date: October 24, 1995

          SERIAL COMPUTER FILES UNDER FORMAT INTEGRATION

We believe that serial computer files composed of textual language
material should be designated as serials with computer file characteristics
under format integration phase 2, rather than designated as computer
files with serial characteristics.   The primary material type of the record
should continue to be coded as "a," with secondary characteristics of
computer files.

                    FORMAT INTEGRATION CODING

One of the primary objectives of format integration is to represent more
than one material type.  This will be done by selecting one material type
as the primary material type (which governs the meaning of the 008
fields) and a secondary material type recorded in the new 006 fixed field.
The primary material type is recorded in Leader/06 (Type of record).
Along with this, the Leader/07 (Bibliographic level) is coded to indicate
monograph, serial, collection, etc.  The 006 byte 00 is used to indicate the
secondary material type.  See Appendix: Fixed Field Coding for a
reference chart of these data elements.

Under format integration phase 2, seriality is regarded as a secondary
characteristic for non-textual materials.  The Leader/06 (Type of record)
is coded for the primary material type (e.g., musical sound recording,
printed map), and the Leader/07 (Bibliographic level) is coded as serial.
The Leader/06 governs the meaning of the 008 fixed fields (i.e., prompting
a workform for a sound recording, a map, etc.).  The serial aspects are
recorded in the 006 field.  Note that the Leader/07 Bibliographic level
and the 006 byte 00 (called T006 on the OCLC workform) are both coded as
"s."

Example for serial sound recording:

OCLC:     NEW  Rec stat: n
Entered:19960101    Replaced:19960101 Used:19960101
Type: j   Elvl: I   Srce:     d    Audn:    Ctrl: Lang: eng
Blvl: s   Form:     Comp:     zz   AccM:fg  MRec: Ctry: cau
Desc: a   FMus: n   Ltxt:          DtSt:c   Dates:1982,9999
-----------------------------------------------------------
T006: s   Freq: a   Regl: r   ISSN:    SrTp:
Orig:     Form:     EntW:     Cont:     Gpub:
Conf: 0   Alph: a   S/L: 0
------------------------------------------------------------

For textual materials, the meaning of the 008 is governed by a
combination of the Leader/06 type code and the Leader/07 Bib lvl, (i.e., a
workform for monograph or serial is prompted by the coding of the
Bibliographic level).

                       CONTENT VS. CARRIER

We believe that, in the context of format integration planning, the
material type "a" (textual language material) has been erroneously
equated with print materials.  The essence of language material is that it
is composed of words (as opposed to sounds being the essence of sound
recordings, maps being the essence of cartographic materials, etc.)

ESSENCE OF MATERIAL           MARC FORMAT

Maps                          Maps
Media, graphics, objects      Visual materials
Sound (music or spoken)       Music/Sound recording
Executable programs           Computer file
Text (Monographic)            Books
Text (Serial)                 Serials

There is no separate microforms format for precisely that reason:  the
essence of microforms is textual language material (Leader/06 = a) with a
bibliographic level (Leader/07) such as m (monograph) or s (serial).
Details about the microform carrier are recorded in the 007 fixed field (as
well the GMD and 008/23 Form of item or 008/22 Form of original item).

While format integration eliminates separate MARC formats, the
cataloger is still responsible for choosing a primary material type, which
determines the meaning of the subsequent 008 bytes.  We believe that
the essence of language serials issued in machine-readable form is still
primarily textual.  We propose that computer file textual serials be coded
as textual serials with computer file characteristics (Leader/06 = a
(language material); Bib lvl = s (serial); 006/00 = m (computer file).
Details about the computer file characteristics would be recorded in the
007 fixed field and GMD.  If desired, 008/23 Form of item and 008/22
Form of original item codes could be defined for textual computer files.

                         EFFECT ON USERS

Since format integration validates all fields for all material types, the
choice of primary material type has no effect on the bibliographic
description.   The coding would affect users when an automated system forces
them to choose a separate file for searching serials (e.g., University of
California's Melvyl catalog, RLIN, and LC).

OCLC and many local systems also offer the ability to limit by material
type.  Ideally, systems will map records with multiple material types to all
relevant files (e.g., serial and computer file).  Realistically such a
capability is unlikely to be promptly implemented in the wide variety of
local systems used in libraries.

The latest trend in preservation is the creation of machine-readable
images of brittle materials.  Just as microform serials are intended to
substitute for the original paper items, these scanned images serve as
surrogates for the originals.  Users with a specific citation are seeking the
publication and the fact that it is on paper, microform, or computer file is
secondary.  Separating the record for the computer-file format from the
other textual records may cause the user to miss the digitized record due
to a search of the wrong file or use of the limiting function.

We oppose differentiating among digitized serials created through
optical-character recognition, image scanning, word-processing files used to
create the print version, and manual keying of articles.  Such distinctions are
difficult to make and not useful for users.  We believe that all textual
computer file serials should be coded in the same way, namely, as serials
with computer file characteristics.

                        MULTIPLE VERSIONS

The need to collocate records multiple versions of the same publication
has been widely accepted by the library community.  The ALCTS
Guidelines for Bibliographic Description of Reproductions, in particular, offers
a method of organizing and displaying cataloging records for different
versions.  Various alternatives have been suggested in terms of the storage
of this data: a hierarchical structure (as proposed in the Guidelines),
separate linked records brought together in a single display, or a
composite record.  (The latter technique is under discussion as a cost-effective
way to indicate that a digitized version of a serial is also available
on the Internet, with the inclusion of an 856 in the bibliographic record).

The goal of all of these techniques is to present data about multiple
versions in a single display for the user.  Encoding the bibliographic
records for digitized serials as a different primary material type from the
printed version will likely preclude the possibility of implementing a
multiple versions approach.

                       CATALOGING WORKLOAD

In the absence of a multiple versions approach, catalogers are creating
separate bibliographic records for print and computer file versions of the
same serial.  This onerous task has been streamlined by the "cloning"
techniques offered by the bibliographic utilities and most local systems. This
function allows the cataloger to create a new bibliographic record while
automatically copying the data from an existing record.  If digitized serials
are regarded as primarily computer files, catalogers will not be able to use
this technique since most systems will not copy records to a
different material type workform.  Catalogers will have to use more
cumbersome methods if they have to create a new record on a different
workform.

                       CONSER IMPLICATIONS

Since CONSER's policy heretofore has been to use the serials workform
for computer file serials, the workload for implementation of format
integration phase 2 will be much less under the proposed scenario.  If
computer file textual serials were designated as primary material type
computer file, all the CONSER records would have to be converted to a
new workform.  Not only does this involve the complex changing of
material type in the CONSER database itself, but it would mean a
complex mechanism of deleting and recataloging for local systems.  Under
our plan, the existing CONSER records could simply be enhanced with
006 and 007 fields.

                        WIDER IMPLICATIONS

This proposal applies only to serials that are textual language materials.
In a case like WordPerfect's Disk of the Month, which consist of executable
computer macros issued serially, the primary material type would be
computer file with the bibliographic level and 006 set to serial.

While we began this discussion in the context of coding remote-access
serials, we found that the arguments for recording textual materials as type "a"
to be applicable to electronic journals, and direct-access serials (e.g.,
CD-ROMS and floppies).  In fact many of them also apply to textual file
monographs.

We hope that CONSER will reconsider the plan to designate serial
computer files as primary material type computer file.  We further suggest
that the library community consider limiting use of the primary material
type "m" (computer file)  to executable computer programs, games,
numeric data files, etc., and designate all textual computer files as primary
material type "a" (language material), be they monograph or serial.

EXAMPLE

Example #1 shows the coding for a computer file serial with computer file
as the primary characteristic, as suggested in Format Integration and Its
Effect on Cataloging, Training, and Systems.   Example #2 shows our
proposed coding, with seriality as the primary characteristic.  The same
bibliographic information is recorded in both.  We believe user
considerations, cost, and theoretical framework all argue for the
alternative approach.

                            EXAMPLE #1
                   FORMAT INTEGRATION PHASE 2
                 SERIAL COMPUTER FILE: PLANNED

OCLC: 1234567  Rec stat: n
Entered:19960303    Replaced:19960303 Used:199603030
Type: m    Elvl: I Srce: d Audn:   Ctrl:     Lang: eng
Blvl: s    File: d       Gpub: f   MRec:     Ctry: dcu
Desc: a                       DtSt: c   Dates: 1993,9999
---------------------------------------------------------
T006: s    Freq:    Regl: x   ISSN: 1   SrTp: p
Orig:      Form:    EntW:     Cont:     Gpub: f
Conf: 0    Alph: a  S/L: 0
----------------------------------------------------------
007    c $b c $d a $e g $f
022    1066-8829
040    CUS $c CUS
110 2  Library of Congress. $b Cataloging Directorate.
245 00 LC cataloging newsline $h [computer file] : $b
online newsletter of the Cataloging Directorate,
Library of Congress
246 1   Library of Congress cataloging newsline
246 1   LCCN
310     Irregular
362 0   Vol. 1, no. 1 (Jan. 1993)-
538     Mode of access: internet email, telnet, and
gopher connections.  For online subscription, send to
listproc@loc.gov, the message: subscribe lccn
[firstname, lastname].
500     Description based on printout of online
display; title from title screen
516      Electronic journal in ASCII text.
856 0   loc.gov  $ :CCM  $h listproc  $i subscribe $z
Email subscription
856 2   marvel.loc.gov  $l marvel  $z telnet connection
to gopher server
856 7    $u gopher://marvel.loc.gov:2070/
11/services/cataloging/lccn  $2 gopher

                        EXAMPLE #2

                  FORMAT INTEGRATION PHASE 2
                SERIAL COMPUTER FILE: OUR PROPOSAL

OCLC: 2387655    Rec stat: n
Entered:19950407    Replaced:19950407 Used:19950507
Type: a     Elvl: I Srce: d Gpub:f  Ctrl:    Lang: eng
Blvl: s     Form:   Conf:0    Freq:   MRec:  Ctry: dcu
S/L:  0     Orig:   EntW:     Regl: x ISSN: 1     Alph: a
Desc: a     SrTp: p Cont:     DtSt: c Dates: 1995,9999
--------------------------------------------------------
T006: m     Audn:        File: d       Gpub: f
----------------------------------------------------------
007    c $b c $d a $e g $f
022    1066-8829
040    CUS  c CUS
110 2  Library of Congress. $b Cataloging Directorate.
245 00 LC cataloging newsline $h [computer file] : $b
online newsletter of the Cataloging Directorate,
Library of Congress
246 1   Library of Congress cataloging newsline
246 1   LCCN
310     Irregular
362 0   Vol. 1, no. 1 (Jan. 1993)-
538     Mode of access: internet email, telnet, and
gopher connections.  For online subscription, send to
listproc@loc.gov, the message: subscribe lccn
[firstname, lastname].
500     Description based on printout of online
display; title from title screen
516      Electronic journal in ASCII text.
856 0   loc.gov $f :CCM $h listproc $i subscribe $z
Email subscription
856 2   marvel.loc.gov $l marvel $z telnet connection
to gopher server
856 7   $u gopher://marvel.loc.gov:2070/
11/services/cataloging/lccn  $2 gopher

                  Appendix:  Fixed Field Coding

(Adapted from: Highsmith, Anne L.  "Format Integration: An Overview,"
Format Integration and Its Effect on Cataloging, Training, and Systems.
Chicago, ALA, 1993, p. 4-5)

                If an item is basically textual:

Type of record (Leader/06)
a (Language material) or
p (Mixed material) or
t (Manuscript material)

Bibliographic level (Leader/07)
a (Monographic component part)
b (Serial component part)
c (Collection)
d (Subunit)
m (Monograph)
s (Serial)

008 = Book or serial or archival control

006 = Any that apply to main item or accompanying materials.  Code 00
(the first position) defines the type of material described in the 006 field.
Codes for new field 006/01-17 are identical to those in fixed fields
008/18-34 for the same type of material.

              If an item is basically non-textual:

Type of record (Leader/06)
c (Printed music) or
d (Manuscript music) or
e (Printed map) or
f (Manuscript map) or
g (Projected medium) or
i (Nonmusical sound recording) or
j (Musical sound recording) or
k (Two-dimensional nonprojectable graphic) or
m (Computer file) or
o (Kit) or
r (Three-dimensional artifact or naturally occurring object)

Bibliographic level (Leader/07)
a (Monographic component part)
b (Serial component part)
c (Collection)
d (Subunit)
m (Monograph)
s (Serial)

008 = Music or map or visual materials or computer file

006 = Any that apply to main item or accompanying materials.  Code 00
(the first position) defines the type of material described in the 006 field.
Codes for new field 006/01-17 are identical to those in fixed fields
008/18-34 for the same type of material.

                     006 Field Tree Structure

006/008 Type   Byte 00 of 006 Material Type

Books          a              Language material
               t              Manuscript language material
Serials        s              Serial control
Maps           e              Printed maps
               f              Manuscript map
Music          c              Printed music
               d              Manuscript music
               i              Nonmusical sound recording
               j              Musical sound recording
Visual mat.    g              Projected medium
               k              Two-dimensional nonprojectable
                              graphic
               o              Kit
               r              Three-dimensional artifact or naturally
                              occurring object
Computer files m              Computer file
Archival ctrl  p              Mixed material