[Esip-preserve] an interesting method of citing a geographic subset

[Greg Janée]

The UCR-STAR (https://star.cs.ucr.edu) system, developed at UC Riverside, has developed an interesting method of citation.  This sytem supports URLs of the form

https://star.cs.ucr.edu/?OSM2015/buildings&d#mbr=9q4ysqb,9q7er8z

in which the fragment identifier (the "#mbr=9q4ysqb,9q7er8z") identifies (using some kind of encoded representation) a minimum bounding rectangle reflecting a region of interest.  (When the above URL is visited it gets redirected to "...#center=35.3323,-119.0918&zoom=10" whose interpretation is a little clearer.)

What is interesting is that this system supports, and advertises, DOIs (well, DOI URLs anyway) of the form

https://doi.org/10.6086/N1JW8BWH#mbr=9q4ysqb,9q7er8z

Notice the same fragment identifier!  I believe this is intended to be a means of citing a geographic subset of this larger resource.

I find this to be an intriguing and interesting method of citation.  If you're not familiar with fragment identifiers, they're actually processed by the client, not the server.  That is, if a client such as a browser visits

https://doi.org/10.6086/N1JW8BWH#mbr=9q4ysqb,9q7er8z

it actually visits

https://doi.org/10.6086/N1JW8BWH

This is the "actual" DOI that has been registered and that redirects to

https://star.cs.ucr.edu/?OSM2015/buildings&d

Only after the resource has been downloaded, does the browser then "look" for the fragment named "mbr=9q4ysqb,9q7er8z" within the resource.  For HTML pages, that means looking for an anchor having that name and scrolling to that position.  But in this case, there's some swishy JavaScript going on that interprets the fragment and issues a redirect to the "...#center=35.3323,-119.0918&zoom=10" form, which in turn requires additional JavaScript to interpret.

There are two things going on here:

1. The PID proper does not include the fragment; the DOI is just 10.6086/N1JW8BWH.  So, the URL form of the PID is being used as a means of adding information to the PID, of qualifying it.  This practice, I would argue, is creating a new kind of identifier.  Persisting the citation means persisting the whole URL, and the interpretation of that URL.

2. A fragment must refer to something within the resource.  In this case, the fragment is processed by some code; without the code, it is uninterpretable.  Ergo, for such a citation to persist, the associated JavaScript must be present and invoked.

Does anybody know of other examples of this kind of subset citation mechanism?

-Greg