正解The correct answer is ア: “a-Use the current one, b-Use the current one.” This question asks how identification codes should be handled when migrating without stopping the system, using the old and new methods together (parallel migration). The decisive clues are the two premises set out in the question. First, it has been decided that the file (table) for managing the lending history will continue to use the one from the current system as-is. Each lending history record points to which book was borrowed and which user borrowed it by means of the book identification number and the user number on the user card. If the numbers were renumbered into a new scheme along with the migration, the numbers in the history records already accumulated and the numbers in the records to be added from now on would become different things, and once mixed in the same table it would no longer be possible to trace the same book or the same user. One could also create a mapping table between old and new numbers and convert all past data at once, but that departs from the premise of using the current file as-is, and it invites the risk of conversion errors and of operations coming to a halt. Second, during the migration period the current barcodes and the newly introduced RFID are used together. Attaching RFID to several hundred thousand books cannot be finished all at once, so books that already have RFID and books that still have only a barcode will be mixed together for a long time. To make sure the same book is recognized no matter which medium is read, it is natural to make the identification number written into the RFID the same as the current identification number that the book’s barcode represents. The same applies to the user number: even when switching to new user cards with built-in RFID, if the user number is carried over unchanged, the user of an old card and the user of a new card are linked as the same person in the history, so operations can continue without worrying about the order in which cards are replaced. In addition, if the numbers are left unchanged, the labels on the shelves, the notices to users, and the links with other systems also need no modification, which keeps migration cost and confusion to a minimum. Therefore both a and b should use the current ones, and ア is correct.
×Incorrectイ
Using the current one for a is correct, but replacing the user number with a new number for b is wrong. The lending history table records the user number, so if the numbers are reassigned, it becomes impossible to tell from the table that the old user number left in the pre-migration history and the new user number recorded after the migration refer to the same person. The premise of using the current file as-is collapses, and operations that span the migration period — such as processing the return of books that are out on loan, or managing overdue items and reservations — break down. Also, until cards with new numbers have reached every user, both old and new cards must be accepted, and on that point too, switching the numbers is disadvantageous.
×Incorrectウ
Using the current one for b is correct, but replacing the book identification number with a new number for a is wrong. Since barcodes and RFID are used together during the migration period, if a number different from the current one is written into the RFID, the same single book ends up being treated as a different number via the barcode and via the RFID, and at every loan and return one would have to worry about which number the transaction was processed with. Furthermore, it would no longer match the past book identification numbers left in the lending history table, so the current file could not be used as-is. The work of renumbering all several hundred thousand books itself involves re-labeling and data conversion, which runs directly counter to the aim of migrating without halting operations.
×Incorrectエ
This replaces both a and b with new numbers, and is the least appropriate. Because the numbering scheme changes on both the book side and the user side, it becomes impossible to know which book and which user in the new system the records with old numbers accumulated in the lending history table point to, and the premise of using the current file as-is does not hold at all. Considering also that barcodes and RFID, and old and new cards, are mixed together during the migration period, an extra mechanism for constantly managing the correspondence between old and new numbers becomes permanently necessary. Far from migrating smoothly without halting operations, this would greatly complicate both the migration work and the operation that follows.
ここがポイント
In a system migration, when existing data or files (tables) are carried over, the identification codes that serve as the keys for referring to that data (book identification numbers, user numbers, and so on) are in principle kept as the current ones. If the code scheme is changed, past data and new data can no longer be matched up, and the work, risk, and cost of data migration and conversion rise sharply. In designing a migration approach, the key point is to think separately about what changes (here, the reading medium) and what does not change (here, the numbering scheme).
RFID is a mechanism that reads and writes information wirelessly without contact. Unlike a barcode, it can read many tags at once and can read through obstacles, so it can make library lending, returning, and stocktaking more efficient. However, with RFID only the means (the medium) of identification changes; the identification number written into it can be the very same value as the existing barcode. In a parallel migration that uses old and new media together, making sure the same object is identified by the same number is the key to migrating without halting operations.