Longevity of Records

[nschorgh]

How do cave data records have to be designed so they remain fully valid or can be seamlessly updated as changes occur?

Country names: Over the past half century, numerous changes in country names have occurred, mostly because nations were divided into smaller nations (e.g., the dissolution of the Soviet Union, Yugoslavia, or East Timor's split from Indonesia). Even over the span of mere decades country names can change. The UIS data exchange commission proposed a field <previous-country-name> (CA231) to deal with such changes. It cannot be expected that every CaveXML database will be updated on the day a new country name becomes official. For that reason, it is practically necessary that old country names remain valid entries in the list of allowed country names defined in cavexml.xsd, at least for a transitional period. CaveXML already allows more than one country name in a record, because a single cave could have entrances in two countries. What I propose is that, for the time being, new country names are placed before/above the old country names. Once too many old country names pile up, it might be good to introduce a <previous-country-name> element, but the transitional solution should be good enough for the next few decades.

Link rot: The solution to link rot are persistent object identifiers, such as doi numbers for bibliographical references. The doi 10.1016/j.icarus.2020.114271 forwards to https://doi.org/10.1016/j.icarus.2020.114271, although some years earlier it would have forwarded to a URL beginning with http://dx.doi.org/ instead. Either way, DOIs are permanent. Huge volumes of specific information about caves are not published in academic journals, but often in caving magazines and other media that do not have doi numbers or any other type of persistent object identifiers. If a website hosting this material goes down, a new URL has to be found, or, worse, the information may have disappeared forever. There is an archiving task to be undertaken to make sure these publications remain available in the future and have persistent identifiers assigned to them.

Caves and their content can change over time. Volcanic eruptions can erase older lava tubes. Caves that once had perennial ice might now only have seasonal ice. Guano mined from caves may be depleted. And caves in ice (glacial caves) can change in length from year to year. Currently, CaveXML provides no mechanism to deal with such changes in an automated way, although notes can be left in the comments and curation fields. Adding year numbers as time stamps would address these issues, but only a tiny fraction of caves change significantly over decades, so this might not be worth the additional complexity it introduces.