NIST PubID scheme: Ambiguity in representation of Revision

[ronaldtse]

From @mico

This is a Revision 1 of AMS 300-8. The parsed PubID now puts the "1" as a "part" which is wrong. After looking at these examples in https://www.nist.gov/system/files/documents/2021/08/26/Publication-ID-Proposal_26Aug21.pdf I made it work this way.

I got to conclusion: If there are no other parts (volume, part, edition) then PubID with revision should look like: NIST AMS 300-8-1. But if there are other part, then revision with letter will be added, example: NIST AMS 300-8v1r1

Interesting. Did not know sometimes the edition/revision is represented as "-N". I think it will be very confusing to use "-N" as edition vs "rN" (e.g revision 4 as "r4"). We will have to report back to NIST that this is an issue.

Originally posted by @mico in https://github.com/metanorma/nist-pubid/issues/8#issuecomment-1000121010

[ronaldtse]

The current PubID scheme supports the current practice where Revision/Edition is expressed in two ways:

1. By suffixing the report number with -N (where N is the Revision/Edition number), e.g. NIST SP 800-73-4, NIST HB 135-2020
2. By suffixing the report number with rN or eN (where N is the Revision/Edition number), e.g. NIST AMS 300-8r1, NBS CIRC 11e2

——

The problem is that these are ambiguous for someone not familiar with the PubID scheme:

a) The pattern “nnn-mmm-ooo” exists in legacy identifiers.

In the existing data set, we know the last part “ooo” is the revision/edition number. e.g. NIST HB 101-1-1990, NIST FIPS 11-3-1991. The only exception to the rule "nnn-mmm-ooo” is “NBS CRPL 4-m-5”, where “4-m-5” is the full report number, a special case of 1 is acceptable.

In the 2021-08 NIST PubID spec, there is the example of "NIST SP(IPD) 800-85B-4”, which fits this pattern.

b) The pattern “nnn-mmm” is ambiguous in legacy identifiers. A reader would not know whether “mmm” is the revision/edition number or the part number.

Here are some examples that can be confusing:

• "NIST HB 135-2020” means Edition 2020? (this is Edition 2020)
• "NBS IR 80-2030” means Year 2030? (this is Part 2030, not Revision/Edition 2030… [IR 80 goes through many of the 20XX numbers])

c) Moreover, the “update pattern” of PubID also places “-yyyy” at the end of the PubID, which causes further confusion amongst users.

e.g. "NIST SP 800-73-4/Upd1-2016” gets the reader confused.

——

I can think of the following ways to mitigate this:

• Use an unambiguous representation of Revision/Edition/Year so that it is separated from Part, i.e., do not use “-YYYY”. Such as “eYYYY”/“rYYYY" or the ISO-style “:YYYY” all work.

• Support parsing both patterns 1 and 2, but after migration to PubID, only generate pattern 2 (rN or eN). For the “update pattern”, also use an unambiguous separator for Revision/Edition e.g. instead of "NIST SP 800-73-4/Upd1-2016”, use something like "NIST SP 800-73-4/Upd1r2016” or "NIST SP 800-73-4/Upd1:2016”