IA Packing List discussion
IA Packing List discussion - 8/15/2007
Chris Freeland, Steve, Robert Miller @ IA offices in SF.
I met with Robert & Steve (book apps engineer) to discuss storing additional title- and item-level metadata alongside each scanned book. After reviewing operations at UIUC and preliminary talks with IA, it seemed that adding these items to the Packing List would require the least amount of overall changes in existing library workflow & IA software architecture. I walked them through the 3 outcomes we hope to achieve by making these changes:
- Enabling each library to make a one-to-one match between item records in their ILS to scanned items at IA. Some libraries are doing this manually, which is manageable given a relatively low volume of scanning. Manual processes will not scale.
- Facilitating citation resolution & resource identification within the BHL Portal. The citation resolver for BHL will need as much metadata as possible about titles & items to accurately link EOL & names databases to scanned content. I walked them through how scientific literature is used, and demonstrated the connections between Tropicos & Botanicus.
- Enhancing IA search & retrieval, as well as other services built on top of OCA content, like OpenLibrary.org. The metadata we’re asking to store is fundamental for accurate description of scanned resources, and should reside in metadata stored alongside the digital objects.
As in previous discussions,
both Robert & Steve were supportive of changes but could not make a commitment to add this to the development priorities - that decision is made by Brewster. We walked through the Packing List, revised after discussions with BHL partners, and discussed the meaning of each field, how that field addressed the 3 outcomes, and the impact for making each change from a development perspective.
It became clear that the easiest, and most important, enhancement would be to record an item-level identifier supplied by the library. This could be a barcode or ILS-generated id, and would be up to each library to determine which of those fields (or another) to supply to uniquely identify the book being sent out for scanning. This field would be supplied in the Packing List, copied & pasted into the Scribe software at the time of scanning, and recorded as <ItemID>, or something similar, in one of the XML files that IA already writes out for each scanned book. This would provide the necessary information for automated matching of item records to scanned objects.
It also became clear that adding fields to describe detailed enumeration & chronology for serials and multi-volume monographs is going to be a tougher sell. The main concern from Robert was the amount of time it would take to copy & paste the multiple fields between the Packing List and the Scribe software. We discussed ways of automating the pull of information from the Packing List, which included:
- Encoding the values for fields into barcodes using standard barcode generation software and printing them on the Packing List. This way the scanner operator would just have to scan a barcode (or a small series of barcodes) to rapidly populate multiple fields in the Scribe software. More work is needed on this to evaluate barcode formats, barcoding software, and how to drop barcodes into cells in Excel.
- Finding another format for the Packing List, like XML, and scripting an ingest procedure. This, obviously, will take more IA development time and was generally considered doable, but not ideal.
Next steps
- Follow up with Brewster, Robert and Steve next week (8/20).
- Discuss at BHL meeting in September
- Come to agreements on how to get additional item-level information into BHL Portal
- Demonstrate/Evaluate encoding metadata in a barcode.
- Topic for discussion at OCA meeting?