Should we include filename code prefix as part of the value of the corresponding metadata? #2
|
Protocol version 1.2: There is confusion about whether filename codes used as prefixes in filenames (e.g., “PI” for ProjectID in Table 4.1) should also be included in the corresponding metadata values, or whether they should appear only in the filename itself. |
Replies: 1 comment
|
No. The prefixes are only needed in the filenames and would add complication if also included in the metadata files. Right now, some of the data science people are clear that. Here is some text from prior email messages: "I would advise we not include the field prefixes in the metadata files; these should be used exclusively for the filename. Including the field prefixes would be less compatible with DarwinCore and inconsistent with how most metadata formats are structured. It would also require additional steps to parse the prefix from the value. The prefix is included in the filename because there is no other structure to link the values with the intended field, but the columnar structure of the metadata file creates that context." |
No. The prefixes are only needed in the filenames and would add complication if also included in the metadata files.
Right now, some of the data science people are clear that. Here is some text from prior email messages:
"I would advise we not include the field prefixes in the metadata files; these should be used exclusively for the filename. Including the field prefixes would be less compatible with DarwinCore and inconsistent with how most metadata formats are structured. It would also require additional steps to parse the prefix from the value. The prefix is included in the filename because there is no other structure to link the values with the intended field, but the columnar structu…