Barcode Requirements (See also catalog format example)
- Unique Identifiers: The barcode must uniquely identify the specimen within the collection. It is important that methods are in place that ensures that no two specimens receive the same barcode identifier, which is particularly important for collections that make their own barcodes using a barcode printer.
- Must be stable: Ideally the barcode should never be modified. Therefore, ensure that barcode is truly unique at the time of assignment so that it doesn’t have to be reassigned.
- Where to Place: The final location for the barcode should be a location that is easy to scan without the need of opening a packet or disturbing the specimen any more than is needed. Remember, that the most important role for barcodes is that they will supply an easy and reliable method for identifying a specimen when preforming curatorial management tasks. If one is processing a group of specimens for a loan, one should be able to quickly go through a stack and scan each specimen without much trouble. Note that OCR returns of text immediately to the right or left of the barcode can be problematic. In order to reduce OCR “noise” that a barcode can create, it is preferable if the barcode is above or below the label with no adjacent text on a horizontal plane.
Barcode Recommendations and Comments
- Global Unique Identifiers (GUID): Ideally, the barcode identifier would uniquely identify that specimen relative to all other specimens found worldwide. The current TDWG recommendations for creating unique herbarium identifiers are to use: <institution code>:<collection code>:12345678. For more information on the Darwin Core recommendations: http://rs.tdwg.org/dwc/terms/index.htm#occurrenceID
- Format: Barcodes are often alphanumeric. The most common barcode standard used for herbarium specimens are Code 39. It’s a good idea to avoid using special characters (!@#$%&) and spaces when possible. The size of the barcode label will depend on the space available on the specimen. Smaller sized lichen and bryophyte specimens may make a barcode of the full GUID (ca 18 digits) impractical. In this case, the barcode might only represent the numeric portion of the identifier or have a collection code of only one to two digits.
- Set Number of Digits: Barcodes with a uniform number of digits aids in catching and avoiding errors within the database. For the numeric portion of the barcode, collections typically use 7-8 digits with left padded zeros. For example, ABC herbarium with 275,429 lichen specimens might have a barcode sequence from ABC:L:0000001 to ABC:L:0275429. If the collection chose to match barcode and accession number, a specimen with an accession number of 1234 would be something like ABC:L:0001234.
- Readable Identifier: Include the human readable digits with the barcode so that one can read the identifier without the need of a barcode reader.
- Ordering Barcodes: When ordering preprinted barcodes, it’s a good idea to order enough extra barcodes to cover incoming specimens for the next 10 years or more.
- Pre-printed -vs- Barcode Printing: This TCN project recommends using pre-printed barcodes.
- Pre-ordered barcodes bought in quantity are typically the cheaper way to go in the long run. This option avoids the need to buy and maintain printer, ink, blank barcodes, etc.
- Purchasing pre-printed barcodes as a batch ensures that each barcode is unique. If one prints their own, ensure that your database application restricts the entry of duplicate barcodes.
- Affixing pre-printing barcodes to specimen is generally fast since one doesn’t have to wait for printer to pop-out barcode.
- Barcode printers make it easier to print 3 barcodes of the same number for specimens with 3 sheets. However, a regular printer can be used for printing an occasional barcode.
- Using a barcode printer may be easier if a collection wishes to match barcodes with accession number, yet one needs to be very careful with typing the accession number in correctly. Errors like this can lead to more than one specimen having the same barcode.
- Sequential –vs– Matching Accession Number
- Matching barcodes is more work, time consuming, and typically the more expensive option. This is particularly true if barcodes are preprinted.
- If one matches barcodes with accession numbers, a method is needed to ensure that no two specimens receive the same barcode identifier. Multiple specimens accidentally being given the same accession number is a typical problem within herbaria (e.g. stamp failed to advance).
- Not matching produces one more identification number that can be assigned to a specimen. Institutions that decide not to match barcodes with accession numbers, often decide to do away with the old accession number in place of the new barcodes.
- Multiple Sheet Specimens: There is disagreement on how to handle specimens that consist of multiple sheets. Some prefer that each specimen gets its own barcode while others assign the same barcode to each sheet. From the database perspective, using a single barcode identifier for all sheets of a specimen is preferred. Ideally, a single specimen should be represented by a single record within the database, whether the specimen consists of one or ten sheets. When general users query a database, they typically want the return count to correctly identify the true number of unique specimens rather than the number of sheets. When they look at the details of specimen record, they typically prefer to view images of all the sheets at once rather than having to click on separate records representing a different sheet. Finally, in the event of an annotation, the data managers should not have to enter the same annotation for multiple records within the database. Multiple records representing a single specimen not only increase data maintenance workload, but it also creates an increased possibility of ambiguity if each records states something different because each record was updated differently.