The NACO FTP process
There are two mechanisms for processing records via File Transfer
Protocol (FTP) currently in use by LC and its NACO partners. These
mechanisms are used to facilitate participation in the NACO component
of the Program for Cooperative Cataloging. This NACO FTP process
is often referred to as the NACO "record exchange" process or as
the NACO "contribution/distribution" process with the FTP partners
often called "NACO nodes."
NACO FTP
- Distribution recipients
- Note that before this option can take place interested
parties must reach agreement with LC for testing, scheduling,
coordination of changes to format elements, etc.
- Distribution recipients must hold a copy of the 5.5 million-record
LC Name Authority File (NAF), and they must keep it current
by loading daily NACO distribution files from LC's Cataloging
Distribution Service (CDS).
- Distribution recipients must also retrieve and process
daily response files that contain error messages that may
pertain to records that were contributed to LC the previous
day.
- Because distribution recipients hold current copies of
the NAF, they can contribute updates as well as new records
to the master NAF housed at the Library of Congress.
- Distribution recipients may also be known as "NAF copy
holders", "full nodes", "NACO record exchange partners" or "participants
in the LC (or NACO) distribution/contribution process."
- At present, there are three distribution recipients: OCLC,
RLG, and the British Library.
- Distribution recipients generate their own LCCNs, each
with a distinguishing prefix (i.e., n = LC; nb = BL; no =
OCLC; nr = RLG).
- Contribution-only sites
- A parameter of NACO participation is the ability to contribute
via a "full node" (see no. 1 above). This option
may be exercised only at LC's discretion and is not generally
available to NACO participants.
- Contribution-only sites do not hold a copy of the 5.5
million-record NAF, but must have search access to a current
copy of the NAF via a NACO distribution recipient (cf. no.
1 above).
- Contribution-only sites can only contribute new
records (i.e., MARC 21 Authority Format Leader/05
of "n" (Record status).
- Contribution-only sites cannot contribute updates to the
NAF (Record status = c,a,d,s,x).
- Contribution-only nodes use pre-assigned LC numbers (i.e.,
prefix = "n" only).
- There are currently two contribution-only nodes: 1) the
National Library of Medicine and 2) LC's Overseas Offices*.
Record Transfer
In the beginning of the Linked Systems Project (LSP), 1985-1995, the Record
Transfer Protocol was used to exchange name authority records. RLG and OCLC
would connect with LC's server and retrieve authority records hourly (usually
loading them that evening); LC would gather contribution authority records
every four hours and load them immediately into MUMS.
FTP
Upon migration to the File Transfer Protocol (FTP), January 12, 1995, the schedule
for loading the files was changed to incorporate an additional 24-hour difference.
Instead of retrieving distribution records throughout the day and loading
them that evening, distribution recipients now retrieve a single file (via
FTP) the following morning. That file contains all records that were created,
updated, loaded, or deleted in the LC system during the previous day. In
addition, each day LC retrieves a single contribution file from each distribution
recipient and processes the files immediately. These files contain both newly
created and changed name authority records.
Note that subject authority records are not part of this process.
This is a simplified picture of the contribution/distribution process.
Contribution
These contribution files are retrieved by LC each morning via FTP, which means
that records created in the bibliographic utilities by NACO libraries on
a Monday are put in a queue and delivered to LC on Tuesday morning. The files
are then split into two smaller distinct files: one for new name authority
records and another for changed name authority records. The new records are
bulk imported into the LC Database using a duplicate detection profile to
ensure that LC Control Numbers (010 fields) are not duplicated. The changed
records are processed and bulk imported to merge with existing records if
the 005 field on the incoming records also matches the 005 field on the
record in the database.
Version Control
In order to insure version control of records loaded into the LC master file
a program check is set on the MARC 21 005 field (Date and time of latest
transaction). The 005 field contains the date and time (down to a tenth of
a second) of the latest transaction, it ensures that the change coming into
the database has used the latest version of the record in the master copy
of the name authority file at LC. This assures that no information is lost
because an earlier version of a record will not replace a later version of
the same record.
From 1985 to August 1999, LC's automation office handled NACO distribution.
Distribution files were created circa 1:30 am and contained all versions
of update records. For example, if a record was updated six times
during the previous 24-hour period, all six copies were in the distribution
queue and distributed to the master-copy holders. Since LC implemented
its integrated library system (LC-ILS) in August 1999, CDS has handled
NACO distribution. CDS files are deduped (based on the information
found in the 005) and as a result only the latest version of a record
updated multiple times is distributed. The number of errors involving
unequal 005 field (known as "C2" errors) has increased only slightly
since FTP became the record exchange protocol. When NACO participants
receive notices of this type of error, they must apply their changes
to the updated record which has recently been loaded by the distribution
recipient.
Distribution
Once contribution file records are loaded into the LC Database they become
part of the master copy of the LC NAF. This means that all NACO contributed
records together with the LC-generated records may now be distributed to
all distribution recipients. To do this LC creates a distribution file that
contains all name authority transactions for an entire day. The distribution
file, normally created before 8:00 am each day, is retrieved by all three
distribution recipients via FTP and loaded before the following morning.
This means that contribution records created on Monday are loaded into the
LC Database on Tuesday, distributed on Wednesday, and generally loaded by
the distribution recipients before Thursday morning.
*Both NLM and the LC Overseas Offices contribute a limited number
of bibliographic records as well as authority records. The Overseas
Offices now primarily contribute via OCLC; however, once or twice a
month a batch of new records is loaded at LC. It is expected that this
activity will be eventually phased out.
|