----- Original Message -----
From: Mary Mastraccio [mailto:[log in to unmask]]
Sent: Tuesday, April 24, 2012 05:54 PM
To: [log in to unmask] <[log in to unmask]>
Subject: Re: [PCCLIST] Alternates for the RDA conversion

Comments from the Univ. of Washington:
Goals worth achieving: 
G-1. headings not suitable under RDA are clearly identified, [ in a machine environment:  the computer will not know which headings are usable]
G-2. all records that may receive 046s and 378s get them, [ it will be very important to populate the file with needed data elements, and the 046 is one of them.  It would be unfortunate to fall back on partial and manual addition of this field, when it can be done programmatically by machine.]
G-3. and all records suitable for use under RDA are coded as such [ in a machine environment:  the computer will not know which headings are usable; The authority file also needs clear coding on the status of records.  That is, AACR2 headings suitable for use under RDA need to be coded as RDA. ]

Univ. of Washington view:
V-1. the real problem is the capacity of the NACO distribution pipeline.  It was adequate in the past, but not in today's environment. 
Rather than discussing inadequate alternate scenarios, we would rather see the discussion shift to how the pipeline can be brought up to a capacity that meets current demands.  

V-2 In our view, a minimum capacity for the pipeline is 10% of the file per day, which would be about 800,000 records/day.  This would allow a complete reload in 10 days, which is still a long time, but acceptable.  Surely future plans to "pour it into a new data structure" will require this kind of capacity, so we may as well start now.

V-3. From the point of view of a local library, changes to large numbers of records are not a problem as long as the heading (1XX) itself stays the same.  We can load hundreds of thousands of records quickly if the only changes are added note fields or changes to coded values.

V-4. Alternate 5:  this looks acceptable if it can be carried out in the week described.  However, if pipeline capacity is an issue, we do not understand how this is possible.

MLM Response: 
G.2. I agree that the Alternate Options do not give a clear picture of when and how some of the changes to authority records would take place but if I understand them correctly, the phase to add 046 and 378 will still be an automated process. I would say, all of the automated processes in the original option, except for adding 667, should be on a project list for phased implementation. The timing could be spread out over the next year to relieve workflow issues. No need to wait until Day 1 for many of these.

G.1 and G.3. I do not think it is necessary to code records as valid/not valid for RDA use. Neither computers nor users care if the heading/record is "RDA". To assume that it is necessary to prevent use of AACR2-only records in an RDA catalog implies that all current NARs which cannot automatically be converted to RDA will be "deletes"--if they aren't valid for use, unless you have a non-RDA catalog, they basically function as delete/dead/defunct records. That is 225,000 headings that libraries have been using which suddenly are invalid---NOT a realistic option. It is a viable option to identify these internally for future projects to "enhance" these to full-RDA records.

V.1 - V.3. It is a good idea to start thinking about changes that are needed to facilitate heavy data flow. However, it is not just the NACO pipeline where the issue lies. True, if most of the changes do not impact indexable terms then the loading/re-indexing in local libraries may be faster than anticipated but libraries already receive many changed records that only have changes in coded values or notes and they DO notice the impact of loading a larger file, even for these non-significant changes. We have received many calls/emails from customers concerned about how they will be able to handle such large files if the phased implementation is too aggressive and repetitive.

V.2 & V.5. IF it is necessary to redistribute most/all LC records, it may work best for everyone to have a target date when they need to load and re-index. Some local systems will have to be down during reindexing so it is necessary to have a well-advertised long-lead time to prepare for this.  Although the NACO-node was mentioned, from all I've heard from local libraries, the concern with any phased implementation is that it not just distribute records without thought to the impact on local systems. I believe it is better to spread out the redistribution and only include records that have real changes--not a just a flag indicating "not valid for RDA".

Mary L. Mastraccio
Cataloging & Authorities Manager
San Antonio, TX 78265