[Previous] [Contents] [Next]

How Much Chocolate Does Database Conversion Require?

Coordinator/Presenter:
Ann Bradshaw, Dallas County Community College District
Presenters:
Rob Kairis, Kent State University, Stark Campus
Barbara Cummings, Innovative Interfaces, Inc.


Ann Bradshaw spoke first, outlining the process her library used to review their previous system [NOTIS], and what they wanted when they moved to the new system [INNOPAC]. Chocolate was used as incentive. Beginning with the catalogers and round tables where chocolate chip cookies were served, she and the catalogers discussed their database, its flaws and what they needed/wanted in a new system. Discussions began a full year before the contract for the new system was signed.

Dallas County Community College District (DCCCD) staff analyzed each field tag used which is something very difficult to do in NOTIS. They looked for methods for improving data and the access to the data. This was an opportunity to make changes, eliminate unwieldy and unnecessary data. It was also an opportunity to add fields not currently used. DCCCD had two special collections which they needed to track. The 9xx fields were used to add data to records at the time of conversion, so it could be traced.

According to Ann Bradshaw, DCCCD’s conversion went off with very few problems or errors. Staff members from all sections of the library were involved with the process. As the level of staff participation increased, so did the quality of the chocolate offered as incentive. Ann recommends that you double the amount of chocolate during TestPAC, moving to premium chocolate during database load, and providing gourmet chocolate for the testers. Double the chocolate if things are going well, triple it if the conversion is going well.

Dallas Community College built in room for growth and change. They have approximately 500,000 item records, 280,000 bib records and 76,000 patron records. Cleanup took about three weeks. Ann had a six-page exception report from INNOPAC as opposed to hundreds of pages from their conversion to NOTIS. This report has already been cleared up.

During this process, DCCCD also utilized the services of a conversion consultant. Someone who had knowledge and experience moving data from NOTIS to INNOPAC. This person acted as a go between for DCCCD and Innovative.

Rob Kairis from Kent State University, Stark Campus, was their consultant. All correspondence was done by e-mail. The DCCCD staff met Rob in Nashville for the first time.

His job was in two parts: getting the data off NOTIS and loading the data onto INNOPAC.

Rob talked about the need to off-load the data in USMARC format, and profiling the new system to handle the current system’s off-loaded values.

Rob indicated that there were two methods for off-loading the data. The first was to off-load all the data for each bib record and all of its attached records into one large USMARC record. The second method was to load all the records into separate files.

Method 1 meant adding a 945 for each item, a 946 for every lib has, and a 947 for every order record attached to the bib record. With NOTIS, Rob noted you could encounter the following problems.

Method 2 meant using separate files for each type of record. Using method 2 you would follow these steps:

Rob drew a flow chart to demonstrate the flow of data as it was matched and attached. Item records are evidently the most difficult to match. Records must have a unique identifier to match up. [Time for more chocolate.]

Once the data was converted out of NOTIS and loaded into INNOPAC, Barbara Cummings, the implementation consultant took charge of the discussion. Barbara has excellent handouts in the IUG conference notebook. Please refer to these sheets for the bulk of her discussion.

After Innovative receives the data, it is preprocessed. The consultant is looking to see if the data will convert. What if any problems appear along the way, and do the data conform to the definition of USMARC used by Innovative? Innovative’s definitions are very strict.

Barbara checked the DCCCD data for anomalies. She found a few. Barbara then described the steps she used to analyse the data from DCCCD. She began by looking at the data informally, and doing both mental and actual checklists. She reviewed the filing indicators, the subject headings etc. When anomalies occurred, she asked investigatory questions:

Once the questions were answered she did a test load of the data and looked at it. If changes needed to be made, she could do them [Innovative will charge for this] or the library could make the changes. In this case Rob as the consultant made the changes.

Barbara then went on to talk about the two most common problems she sees as an implementation coordinator - call numbers and location codes. She described the methods she uses for deriving call numbers, and location codes. These are outlined on her handouts in the conference notebook.

In summary – DCCCD’s conversion went very smoothly thanks to a lot of advance planning on the part of the college and the cooperation between Innovative, the college and their consultant. While not something anyone wants to do on a regular basis, with enough chocolate, you can get through a system conversion.

Yes, the group from Dallas Community College did bring samples.


Recorded by: Judith Cerqua, State Library of Ohio

[Previous] [Contents] [Next]